<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by The DigiPalms on Medium]]></title>
        <description><![CDATA[Stories by The DigiPalms on Medium]]></description>
        <link>https://medium.com/@thedigipalms?source=rss-b522d97c2401------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*DxEwHvOopv0BDcnaKDLB_A@2x.jpeg</url>
            <title>Stories by The DigiPalms on Medium</title>
            <link>https://medium.com/@thedigipalms?source=rss-b522d97c2401------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 05:20:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@thedigipalms/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Your Biggest Competitor Approved Christmas in August]]></title>
            <link>https://thedigipalms.medium.com/your-biggest-competitor-approved-christmas-in-august-a390aecb141e?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/a390aecb141e</guid>
            <category><![CDATA[digital-marketing]]></category>
            <category><![CDATA[small-business]]></category>
            <category><![CDATA[entrepreneur]]></category>
            <category><![CDATA[brand-strategy]]></category>
            <category><![CDATA[holiday-marketing]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Wed, 18 Nov 2026 13:43:28 GMT</pubDate>
            <atom:updated>2026-11-18T13:43:28.447Z</atom:updated>
            <content:encoded><![CDATA[<h4>In the next three weeks a trend will land, a cultural moment will happen, and the big brands will miss it. The ability to change your mind on November 20 is the unfair advantage small brands keep forgetting they have.</h4><p>Pick any big brand you buy from. Right now, somewhere in a shared folder, there is a document that spells out exactly what they are posting on December 14, what their ads look like through the end of the year, and which creators they already paid in September.</p><p>Nothing on that list is changing this week.</p><p>You don’t have that problem.</p><p>I think most small brand owners know this in the abstract. Agility, nimbleness, “we move faster.” The words show up in every pitch deck. But I rarely see anyone actually set their holiday up around it. Most small brands I talk to in November are trying to run the same playbook the big ones run, but with less money and fewer people. That is a losing trade.</p><p>Here is what the real advantage looks like, and how to actually use it in the three weeks we have left.</p><figure><img alt="A small business owner compares a big brand’s locked Christmas marketing plan with a flexible holiday campaign strategy, highlighting the advantage of fast decisions and reactive marketing." src="https://cdn-images-1.medium.com/max/1024/1*RIUew0JN5mmQX1ZqcBQLMQ.png" /></figure><h3>What a big brand locked in months ago</h3><p>Let me list the things a big brand you compete with cannot change between now and Christmas. It is a long list, and it is worth sitting with it for a minute.</p><p><strong>(Locked) </strong>Them, this week</p><ul><li>Creative shot in July, approved in September</li><li>Media buys committed to hit scale pricing</li><li>Product photography from summer</li><li>Email calendar written and loaded</li><li>Landing pages QA’d and legal-reviewed</li><li>Influencer deals signed and paid</li><li>Stock sitting in warehouses</li><li>End caps and window displays locked</li></ul><p><strong>(Live) </strong>You, this week</p><ul><li>Ship a new ad before dinner</li><li>Call a creator tomorrow morning</li><li>Rewrite your landing page over coffee</li><li>Bundle two products into a gift set today</li><li>Change your offer three times this week</li><li>Send a different email than the one you planned</li><li>Respond to a competitor’s move in the same news cycle</li><li>Turn a trend into a product name by Sunday</li></ul><p>Now look at the locked side and notice something. Nothing on it is bad. Big brands run this way for a reason. At their scale you cannot change the plan every Tuesday. The apparatus does not allow it.</p><p>That apparatus is also the thing stopping them from catching whatever happens next.</p><p>None of what you can do takes genius. It takes an operator with a laptop and a small enough team that the decision loop fits in one conversation. The big brand cannot replicate any of it this quarter. Not because they are slow, but because they are big.</p><blockquote><em>In an industry where everyone is racing to look bigger, your actual advantage is being small enough to still be surprised.</em></blockquote><h3>What to actually do in the next three weeks</h3><p>If this is going to translate to money, it has to be set up on purpose. Here is the practical shape of it.</p><ol><li>Keep a quarter of your December ad budget unallocated.Most small brands have every pound and every dollar committed by the first of the month. That is the big brand approach, and it strips out your only real edge. Hold some back for a moment you cannot predict yet.</li><li>Reserve a creative slot for reactive work.On your content calendar, block one post slot a week with the word “react.” You will fill it on the day. If nothing happens, post something planned. If something does, you already made the space for it.</li><li>Keep a short list of creators you could actually call on Tuesday.Five people. Ones who make content fast, know your brand, and can turn around a post in 48 hours. Rates agreed in advance. The time to build this list is this weekend, not the morning something breaks.</li><li>Make templates you can swap into.One ad frame, one email template, one landing page layout. All built so the hero image and the headline are the only moving parts. React in an hour instead of a week.</li><li>Set up a tiny monitoring habit.Fifteen minutes in the morning. What is trending in your space. What your three biggest competitors posted overnight. What cultural moments are in the air. Not a dashboard. Just eyes open.</li><li>Shorten your approval loop to one person and one hour.For reactive work, a 24-hour approval cycle is already too slow. Pick one trusted person. Agree on the guardrails. Then move.</li></ol><h3>The catch most small brands trip on</h3><p>I have to be honest about the part that trips people up, because I have tripped on it myself.</p><p>Agility without conviction is just thrashing.</p><p>Being able to react to everything is only useful if you know what to react to and what to ignore. The small brand that chases every trend becomes a mirror. No point of view. No reason to come back. Just whatever the algorithm pushed last week, dressed in your logo.</p><p>The big brand problem is that they cannot move. The small brand problem is that they can move too easily, so they move for the wrong reasons.</p><blockquote><em>Being able to change your mind is only an advantage if you had a mind to begin with.</em></blockquote><p>What that looks like in practice: you have two or three things your brand is actually about, and you only react to moments that touch one of them. The viral dance that has nothing to do with your product, you skip. The cultural moment that lands right on top of what you believe, you ride, and you ride hard. The filter is your point of view.</p><p>A small brand without a point of view has all the cost of being small and none of the benefit.</p><h3>What I would do before the end of this week</h3><p>If this piece lands at the right time for you, here is the short version to set up by Friday.</p><p>Pull an unallocated slice out of your December ad budget. Write the five creator names on a note and send three of them a message today. Build one swappable template in whatever tool you use. Block the “react” slot on your calendar for the next three weeks. Write one sentence that describes what your brand actually stands for this quarter, so you have the filter ready when a moment comes.</p><p>Then just watch. Something will happen. It always does.</p><p><em>I spend a lot of time working through questions like this one with other operators inside </em><a href="https://go.businessbuilderscircle.com/private-group"><em>Business Builders Circle</em></a><em>. A room of founders and marketers building the reactive muscle big brands can only dream about this time of year. If that’s useful, that’s where to find us.</em></p><h3>The quiet part</h3><p>Every year I watch the same pattern play out between Thanksgiving and New Year.</p><p>Big brands spend a lot of money looking polished. Small brands spend a lot of energy trying to look polished too, and lose the one thing that was working in their favor.</p><p>The budget is not the advantage. The team is not the advantage. The production value is not the advantage. You will lose every one of those races.</p><p>The advantage is that at 10 AM on a Tuesday in November, you can still change your mind.</p><p>Big brands would pay a lot of money to buy that back. They cannot. You already have it.</p><p>Try spending it this year.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a390aecb141e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Which Claude Model Should You Actually Use?]]></title>
            <link>https://thedigipalms.medium.com/which-claude-model-should-you-actually-use-16c33d0bd3c4?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/16c33d0bd3c4</guid>
            <category><![CDATA[claude-ai]]></category>
            <category><![CDATA[claude]]></category>
            <category><![CDATA[ai-prompt-engineering]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 10 Nov 2026 17:31:03 GMT</pubDate>
            <atom:updated>2026-11-10T17:31:03.582Z</atom:updated>
            <content:encoded><![CDATA[<h4>There are four now. Fable 5.2 just landed. They are not gimmicks, and most people are reaching for the wrong one. Here is how to pick, with real examples from my own work.</h4><p>A friend asked me last week which Claude she should be using for her agency work. Before she finished describing the task, I had changed my mind three times.</p><p>Then she said the thing I have been hearing from almost every operator I talk to.</p><p>“Honestly, isn’t this all just a way to get us to pay for the expensive one?”</p><p>I understood the question. The names sound like a fancy coffee menu. The pricing goes up with the words. And every tier promises to change your life. If you have ever opened the model picker and just gone with the one that sounds smartest, you are not alone. I did the same thing for a while.</p><p>So let me say the honest thing up front. They are not gimmicks. They are genuinely different tools. But most people are reaching for the fanciest option when something smaller would do the same job faster and cheaper. That is a real problem if you are running a business that uses this stuff every day, because it quietly eats your credits and makes you believe AI is slower than it is.</p><p>Here is what each one is actually for. No fluff. Real examples from my work.</p><figure><img alt="A business professional comparing four Claude AI models on a laptop, highlighting their different capabilities in automation, content creation, research, and complex problem-solving." src="https://cdn-images-1.medium.com/max/1024/1*UU8CCcoMBfAZgFHhtyM9nA.png" /></figure><h3>The four, in one breath</h3><p>As of this week, there are four publicly available Claude models.</p><ul><li><strong>Haiku</strong>5.5The small one. Built for speed and volume.</li><li><strong>Sonnet</strong>5.5The everyday one. The default for most work.</li><li><strong>Opus</strong>5.5The heavy lifter. Deep reasoning, long tasks, hard problems.</li><li><strong>Fable</strong>5.2The newest, released on November 9. Anthropic calls it and its sibling Mythos “our most capable models for daunting technical work.” The top tier.</li></ul><p>Now the useful part. What each one is for, and what it is not.</p><p><strong>Model 01 </strong>haiku-5–5</p><h3>The one you are probably underusing</h3><p>Haiku is the model nobody writes love letters about, and the one that would save you the most money if you knew when to use it.</p><p>Think of Haiku as the junior on your team who is very fast, follows instructions well, and never complains. You do not hand them your strategy. You hand them the pile of small, repeatable tasks that would eat an afternoon if you did them yourself.</p><p>Built for</p><ul><li>Rewriting 200 product titles with a set of rules</li><li>Classifying 500 customer reviews as positive, neutral, or negative</li><li>Turning a long voice note into clean bullet points</li><li>Pulling out all the names and dates from a stack of invoices</li><li>Summarizing an email thread into three lines</li><li>Writing short social captions from a longer blog post</li><li>Running inside an automation that fires a thousand times a day</li></ul><p>Anthropic actually names three of these on the launch page. They describe Haiku 5.5 as “built for high-volume work like summarization, subagents, and browser use.” That tells you exactly how they want it used.</p><p>Where it falls short</p><ul><li>Nuanced strategy work</li><li>Writing a flagship blog post that needs real voice</li><li>Analyzing a messy 60-page research report with conflicting findings</li><li>Any task where the answer depends on holding a lot of moving pieces in your head at once</li></ul><blockquote><em>Haiku is not a smaller brain. It is a faster one trained on narrower tasks. Treating it like a cut-price Sonnet is how you decide AI is dumb.</em></blockquote><p>A real example from this week.<em> I needed to take 180 product descriptions on one of my stores and shorten them to under 50 words each, in a consistent voice, with the brand’s three non-negotiable phrases kept intact. Haiku did it in minutes. Sonnet would have done the same job for roughly four times the cost and no better.</em></p><p><strong>Model 02 </strong>sonnet-5–5</p><h3>The workhorse you should default to</h3><p>Sonnet is the one you reach for when you do not know which one to reach for. Anthropic’s own release notes call Sonnet 5.5 “a clear upgrade over Claude Sonnet 5” that “runs 30% plus faster” and “costs up to 30% less for most work.” That is important, because it means the default got cheaper and quicker at the same time. There is no good reason to live on Opus if Sonnet handles the task.</p><p>If Haiku is your fast junior, Sonnet is the capable mid-level who can run almost any marketing task you throw at them without supervision.</p><p>Built for</p><ul><li>Drafting a full blog post in your voice from a brief</li><li>Building a five-email welcome sequence with proper arcs</li><li>Analyzing a competitor’s landing page and suggesting changes</li><li>Researching a topic with web search and writing the summary</li><li>Running a six-step automation that talks to your CRM, email, and spreadsheet</li><li>Writing a thoughtful client proposal from a set of notes</li><li>Most day-to-day agent work that connects three or four tools together</li></ul><p>Where it falls short</p><ul><li>A twelve-step reasoning task where losing track of step three poisons step eleven</li><li>Research synthesis across dozens of sources with contradictions you need to reconcile</li><li>Heavy code work on a large existing project where the context is enormous</li><li>The handful of high-stakes decisions where the best answer, not a good answer, is what you need</li></ul><p>A real example.<em> When a client needs a go-to-market plan written from a two-page brief, five web searches, and a ten-page transcript of their founder interview, Sonnet is my pick. It holds the context, it writes in a real voice, and it finishes in under ten minutes. I used to use Opus for this. I was overpaying.</em></p><p><strong>Model 03 </strong>opus-5–5</p><h3>The one for problems that do not fit on one page</h3><p>Opus is where I go when the task has too many moving parts for me to hold in my head, let alone a smaller model.</p><p>Anthropic positions Opus 5.5 as the model that “leads in agentic coding and knowledge work.” In plain English, that means it is the one you want when the job is long, the context is big, and the thinking has to stay coherent from start to finish.</p><p>Built for</p><ul><li>A deep research project that reads 30 sources and produces a defensible point of view</li><li>Debugging a messy codebase where the bug is three systems away from the symptom</li><li>Comparing 20 vendor proposals with conflicting pricing and writing a recommendation</li><li>A long autonomous task that runs for hours and makes decisions along the way</li><li>Planning the architecture of something new where early choices shape late ones</li><li>The strategy document where you need the AI to notice what you missed, not just write what you asked</li></ul><p>Where it falls short</p><ul><li>Anything fast or cheap at scale</li><li>A rewrite of 200 product titles. That is Haiku’s job.</li><li>A task where the brief is bad to begin with. Opus is smart. It still cannot read your mind.</li></ul><blockquote><em>Opus is not smarter in a general way. It is better at not losing the plot over a long, tangled task. Short, clear tasks do not need it.</em></blockquote><p>A real example.<em> I ran an exercise last month where I asked a model to take a client’s entire customer service inbox for one quarter, pull out the top five reasons buyers churn, and propose a retention program with budget. That is the Opus kind of task. Sonnet could start it. Opus actually finished it well, with a document I used in a client meeting.</em></p><p><strong>Model 04 </strong>fable-5–2 · new</p><h3>The new ceiling</h3><p>Here is the one you came for.</p><p>Fable 5.2 landed on November 9. Anthropic calls Fable and its sibling Mythos “our most capable models for daunting technical work” and says the pair “set a new bar in coding, finance, and the sciences.” That phrasing matters. It is not pitched as the everyday upgrade. It is pitched as the model for the problems that broke the last model.</p><p>If you have been using these tools for a while, you know the feeling. The task where Opus gets most of the way there, but the final answer does not quite hang together. The research synthesis where two of your sources contradict each other and the model picks one without telling you why. The analysis where the numbers add up but the story does not. Those are the problems Fable is built for.</p><p>Fable is also the publicly available sibling of Mythos. Same brain. Fable has extra guardrails in a few dual-use areas. For almost everyone reading this, the difference does not matter. What matters is that Fable is the ceiling you can reach.</p><p>Where it earns its keep</p><ul><li>A strategy piece where you need to hold ten competing considerations and come out with a sharp answer</li><li>A technical investigation where “the answer is in there somewhere” is not good enough</li><li>Financial modeling where the model has to question its own assumptions</li><li>A research report that has to be defensible under scrutiny from a specialist</li><li>A one-shot big decision where paying more to get it right the first time is the correct call</li></ul><p>Where it does not make sense</p><ul><li>Volume work. Haiku runs circles around it on cost and speed.</li><li>Everyday drafting. Sonnet will give you 95 percent of the quality for a fraction of the compute.</li><li>Tasks where the bottleneck is you, not the model. If your brief is one sentence long, Fable will give you a very expensive version of a Sonnet answer.</li></ul><p>How I plan to use it.<em> For the handful of decisions a quarter where I would normally pay a consultant. The pricing strategy. The brand repositioning. The hard, “we only get this one shot” piece of work. The rest of the time, I am still on Sonnet.</em></p><p>The one-sentence rule</p><p>Match the model to the shape of the task, not to the importance of the business.</p><p>Running an important business does not mean every task is an Opus task. Writing a hundred headlines for an important campaign is still a Haiku task. Deciding what the campaign should be about is maybe a Fable task.</p><p>Shape, not size.</p><h3>What I actually do day to day</h3><p>To make this real, here is my own stack this week.</p><ul><li><strong>Haiku</strong>5.5Inbox summaries, meeting notes, product title rewrites, and bulk tagging.</li><li><strong>Sonnet</strong>5.5Client drafts, research summaries, agent workflows, and most content work.</li><li><strong>Opus</strong>5.5Deep research pieces, quarterly strategy documents, and the one messy code project I am stuck on.</li><li><strong>Fable</strong>5.2The handful of tasks where Opus almost gets there. Roughly one task a week.</li></ul><p>The practical effect is that my credits stretch much further, my fast tasks actually feel fast, and the hard tasks get the compute they deserve.</p><p>The uncomfortable truth is that the “which model” question is really a “which task” question. Most of us are bad at judging the size of a task before we start it. Getting better at that is the real skill. The model picker is just where that skill shows up.</p><p><em>I spend a lot of time working through questions like this one with other operators inside </em><a href="https://go.businessbuilderscircle.com/private-group">Business Builders Circle</a><em>. A room of founders and marketers figuring out which parts of their workflow to automate, which model to use for each piece, and how to stop burning credits on the wrong tier. If that’s useful, that’s where to find us.</em></p><h3>The quiet part</h3><p>When I see someone complain that AI is slow, expensive, or disappointing, nine times out of ten they are running the wrong model on the wrong task.</p><p>The tools are not the problem. The problem is that nobody told us what each one is for.</p><p>I think that is going to change quickly, because the real cost of using the wrong model is not the extra few dollars. It is the quiet suspicion that this whole thing is overhyped, and the decision to use less of it in your business.</p><p>Which, by the way, is exactly what your competitor is hoping you will do.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=16c33d0bd3c4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Team in My Laptop Has Zero Taste]]></title>
            <link>https://thedigipalms.medium.com/the-team-in-my-laptop-has-zero-taste-ab3dc7c12d9b?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/ab3dc7c12d9b</guid>
            <category><![CDATA[ai-automation]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[entrepreneur]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 03 Nov 2026 17:02:54 GMT</pubDate>
            <atom:updated>2026-11-03T17:02:54.014Z</atom:updated>
            <content:encoded><![CDATA[<h4>A developer, an automation builder, a designer, a receptionist, a sound engineer. You can now get the work of all of them done before lunch. Here is the one hire AI still can’t make for you.</h4><p>I described an automation in plain English, went to make tea, and came back to something that worked.</p><p>No weekend lost to a tutorial. No waiting for someone’s queue to open up. No “let me check and get back to you.”</p><p>Not long ago, that same request would have started a small project. Someone to build the website. Someone to design it. Someone to wire your tools together so a sale sets off the right email and the right text, and lands in the right spreadsheet. A person to answer the phone. A video person. A sound person if there was audio involved. And a project manager to stop all of them drifting apart.</p><p>That was a quote with a lot of zeros and a calendar measured in weeks.</p><p>Now one person with a laptop can get most of the way there in a day. Sometimes in an afternoon.</p><p>This is not hype. It is the biggest shift in how small businesses get built that I have seen in my career, and it happened so fast that most of us have not done the math on what it means.</p><p>Let me do the math.</p><figure><img alt="A laptop displaying an AI assistant surrounded by digital panels for development, design, automation, customer support, video editing, audio production, and analytics, illustrating how AI can handle multiple business roles." src="https://cdn-images-1.medium.com/max/1024/1*aB_DzEGtT21wGMpURFPFsA.png" /></figure><h3>Tell it what you want. Then go make tea.</h3><p>Two names keep coming up when I talk to other founders. ChatGPT Work from OpenAI, and Muse, Meta’s new personal agent. Different companies, different products, same startling trick. You give them a goal. They make a plan, connect to your apps, and carry it out. ChatGPT Work can stay on a project for hours. Muse asks before it sends an email or spends your money.</p><p>Here is what that looks like in real life. Three requests, typed in plain English.</p><h4>The automation</h4><p>“Every time someone books a call, add them to my CRM, send them a welcome text, and ping me if they look like a big account.”</p><p>Used to take: an automation specialist, or a few weekends inside a drag-and-drop tool, fighting connections that quietly break.</p><h4>The fix</h4><p>“Our order confirmation emails stopped sending yesterday. Find out why and fix it.”</p><p>Used to take: an hour on hold with support, or an urgent invoice from a developer.</p><h4>The custom build</h4><p>“Add a page to my website where clients can upload their brand files, and email me when they do.”</p><p>Used to take: a web developer, a quote, and weeks of waiting for something that sounds small.</p><p>It goes further than chat. Meta says Muse can set up its own connections to other software and build its own tools. ChatGPT Work can turn a pile of information into a working web app, and its coding side can take a developer’s ticket all the way to tested code. Website development, custom dev work, automations and repairs, all in the same conversation.</p><p>You still approve the important stuff, and you should. But the slow part, the waiting and the learning curve, is gone.</p><p>Here is what I think people get wrong about this. It does not make your experts less useful. It points them at better problems. A developer who no longer spends days on a form can spend that time on the thing that makes your product special. An SEO lead who no longer builds reports by hand can spend it on strategy. The tool does the typing. The people with experience still do the deciding, and now they get to do more of it.</p><h3>The rest of the busywork left too</h3><p>It is not just the technical tasks. Here is the rest of the list, and it still surprises me when I say it out loud.</p><ul><li><strong>The receptionist</strong>A voice that picks up your phone at two in the morning, understands what the caller wants, books the appointment, and texts a confirmation. Not a menu of “press one.” An actual conversation.</li><li><strong>The photographer</strong>Product shots in the exact setting you pictured. No studio. No shipping samples across the country.</li><li><strong>The video editor</strong>A polished explainer, with motion, voiceover, and captions, from a written idea.</li><li><strong>The sound engineer</strong>Noisy recordings cleaned up, levels balanced, the whole thing mixed. The hours of fiddling are gone.</li><li><strong>The analyst</strong>Drop in a messy spreadsheet and ask, “which customers are about to leave?” the way you would ask a coworker.</li></ul><p>None of this is perfect every time. You still check the work. But “good enough to ship, and fast enough to throw away and redo” is a very different world from “let’s get some quotes.”</p><h3>Then comes the part nobody warns you about</h3><p>When building was expensive, the price tag filtered your ideas. You had to be fairly sure before you spent the money. Expensive was the bouncer at the door of every bad idea.</p><p>The bouncer is gone.</p><p>Now there is a machine that will take any idea you hand it, including a mediocre one, and build a gorgeous, professional-looking version in under a minute. It will never say, “are you sure?” It will never say, “we tried that and it flopped.” It will never ask, “who is this actually for?”</p><p>It is the most agreeable team member you will ever have. That is exactly the danger.</p><blockquote><em>AI is the first team in history that will build your bad idea beautifully and never tell you it’s bad.</em></blockquote><p>So everyone is building. Everything. All the time. The internet is filling up with things that are well made and completely forgettable.</p><p>When capacity was the bottleneck, the person who could build won. Now that capacity is nearly free, the bottleneck has moved somewhere much more uncomfortable. The question is no longer “can you make it?” It is “should it exist, and why would anyone care?”</p><h3>The skill nobody puts on a resume</h3><p>Here is something that gives me real hope, especially if you have been in marketing for a while. The new skill is one we already know.</p><p>It is writing a brief.</p><p>Every marketer has handed a brief to a freelancer and gotten back something technically fine and totally lifeless. Usually, that was a bad brief. AI is the same contractor, only faster and more patient, with no opinions of its own. It will do exactly what you describe. So the quality of what comes back is the quality of the direction going in.</p><p>Watch what that looks like.</p><h4>What most people type</h4><p>“Write a welcome email for my skincare brand.”</p><p>You get something warm, friendly, and “so excited you’re here.” It could have been sent by any brand on earth.</p><h4>What a good brief sounds like</h4><p>“Write to a woman who has tried too many serums and is tired of being sold to. Make her feel we’re on her side. No exclamation marks. One small joke. Tell her up front that some of our products won’t suit her skin, and that we’ll say so.”</p><p>Now you get something that sounds like a person with a point of view. And a customer who thinks, “finally, someone honest.”</p><p>The tool did not get smarter between those two. You did.</p><p>Which means the people who win are not the ones who build fastest. They are the ones who know what they want, who it is for, and what to leave out. That is not a technical skill. It never was.</p><h3>What the machine can’t bring to the table</h3><ul><li>It can’t tell you which idea to kill. It will polish all of them with equal enthusiasm.</li><li>It wasn’t there. It did not hear the customer sigh on the phone. It did not read the note tucked into a returned order. Details from real life are the raw material of every message people remember, and the machine has none of its own.</li><li>It has no nerve. Taking a side, charging more, turning down a customer who isn’t right for you. Those moves feel risky, and risk is the one thing it never feels.</li><li>It can’t feel “almost.” Telling a nearly right thing from a right thing is taste. Taste is built from years of noticing, and it cannot be downloaded.</li></ul><h3>Try this week</h3><p>Pick the one task you have been avoiding because it felt too technical or too expensive. The automation. The phone line. The video. Build it with AI. Let it be rough.</p><p>Then take the hours you just saved and spend them on the part the machine can’t do. Call three customers and ask what almost stopped them from buying. Bring what they say back as your next brief.</p><p>And before you ship anything, ask one question. Would you put your <em>name</em> on this, or just your logo?</p><p><em>I spend a lot of time working through questions like this one with other operators inside </em><a href="https://go.businessbuilderscircle.com/private-group"><strong><em>Business Builders Circle</em></strong></a><em>. A room of founders figuring out how to let AI do the heavy lifting while keeping the part that makes their business theirs. If that’s useful, that’s where to find us.</em></p><h3>The quiet part</h3><p>For years, the sentence I heard most from small business owners was, “I can’t afford to build that.”</p><p>That sentence is disappearing. In its place is a new one, quieter and harder: “I’m not sure what to build.”</p><p>Notice which of the two is more uncomfortable to say out loud.</p><p>The first was a money problem. The second is about who you are and what you care about. No tool can answer that one for you.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ab3dc7c12d9b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Victoria’s Secret Comeback Is Being Taught Wrong]]></title>
            <link>https://thedigipalms.medium.com/the-victorias-secret-comeback-is-being-taught-wrong-c45f987ccdcf?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/c45f987ccdcf</guid>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[branding]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[brand-revival]]></category>
            <category><![CDATA[digital-marketing]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 27 Oct 2026 16:22:09 GMT</pubDate>
            <atom:updated>2026-10-27T16:22:09.079Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>Everyone keeps saying they threw money at the problem. Their marketing budget actually went down the year the show returned. The real lesson is not about the money. It is about where the money went.</em></h4><p>I have been reading a lot of hot takes about the Victoria’s Secret Fashion Show in the last few weeks, and most of them are missing the thing that actually matters.</p><p>The hot takes go roughly like this. Big legacy brand threw a huge spectacle. Spent a fortune. Flooded the internet with celebrities and influencers. Bought their way back into the cultural conversation. Now the business is popping.</p><p>The implied lesson, usually said out loud by someone selling a course, is that if you want a cultural moment, you need a cultural moment budget.</p><p>That is not what happened.</p><p>In the fiscal year Victoria’s Secret brought the show back, their total advertising and marketing spend went down. Not up. Four hundred and fifty-four million dollars the year before. Four hundred and twenty-nine million the year the show returned. By the second year of the revived show, they were still only at four hundred and forty-five million.</p><p>That is three straight years of comparable marketing budget, with a globally viral comeback event sitting right in the middle of it.</p><p>And in the quarter after the 2025 show, site traffic was up over sixty percent year over year. New customers were up more than fifteen percent. Sales grew nine percent. The stock jumped roughly eighteen percent in pre-market on the earnings print.</p><p>Something clearly worked. It just was not the thing most people are telling you.</p><figure><img alt="Victoria’s Secret-inspired runway show featuring diverse models, angel wings, social media livestreaming, and online shopping, illustrating the brand’s marketing comeback and digital strategy." src="https://cdn-images-1.medium.com/max/1024/1*_r-a172g7VAo7fYGwZqPlA.png" /></figure><h3>They did not spend more. They spent smarter.</h3><blockquote>They did not spend more. They spent smarter. That is the whole case study.</blockquote><p>The show cost real money, obviously. The 2017 Shanghai show was reportedly around twenty-six million dollars, which was the headline figure for years. The modern revived shows have not published their production budgets, and anyone quoting a specific number is guessing. But even at twenty million, which is the high end of the industry whisper, that is less than five percent of VS’s annual marketing budget.</p><p>The show was not an addition. It was a reallocation.</p><p>Instead of spending that money on television spots or static display ads or the hundred other channels that cost real money and produce diminishing attention, they put it into one event that generated fifty-one billion media impressions. Someone did the math and said, we can spend our existing budget on things people scroll past, or we can spend it on something people actually stop to watch.</p><p>That is the first thing to steal, and it has nothing to do with lingerie. Most brands running “always-on” marketing right now are quietly bleeding margin on channels that stopped working two years ago. The move is not to add a spectacle on top of what you are already doing. It is to look at what you are already doing, admit which parts have become background noise, and move that budget into one or two moments that genuinely cut through.</p><p>VS did not get a bigger wallet. They got a sharper point of view about where the wallet should go.</p><h3>The three attempts nobody is talking about</h3><p>Here is the other thing the hot takes skip. The comeback did not happen in one shot. It took three attempts, over three years, and the first two were, by any honest reading, failures.</p><h3>2023: the one everyone forgets.</h3><p>Victoria’s Secret released a documentary-style film on Prime Video called <em>The Tour </em>’23. It was meant to be the modern, inclusive, globally-cast reinvention of the brand. It bombed.</p><p>It currently sits at 2.3 out of 10 on IMDb. Seventy-five percent of Prime Video reviewers gave it one star. The complaints were almost unanimous: it had stripped out everything people loved about the original. One review that stuck with me just said: “I missed the wings. I missed the catwalk.”</p><p>That is the warning label on the whole exercise. If you overcorrect so hard that you abandon the thing people actually came for, you do not win new customers. You lose the ones you had and win nobody.</p><h3>2024: the one that bought attention, not business.</h3><p>The actual fashion show came back in October 2024. The press was wall-to-wall. Earned media value, according to one tracker, was over three hundred million dollars for one night. The brand got thirty-four billion media impressions. Four million new TikTok followers.</p><p>But if you read the reviews honestly, the show itself was flat. Critics called it nostalgic without being exciting. One wrote that the wings looked like “the kind used for projects in elementary school.” Sales did lift. Q3 2024 was the best quarter since 2021. But a chunk of that was a calendar shift and a holiday collection launch, and the stock barely moved on the news.</p><p>The brand had got the attention. They had not yet earned the conversion.</p><h3>2025: the one that actually worked.</h3><p>In between the 2024 show and the 2025 show, Victoria’s Secret quietly did something that most brands do not have the stomach for. They changed the CEO. They changed the creative director. They changed their point of view.</p><p>Hillary Super took over as CEO in September 2024. In the spring, she brought in Adam Selman as creative director. And here is the detail that should make every operator pause.</p><p>Both of them came from Savage X Fenty.</p><p>Savage X Fenty is the brand that spent the last five years quietly eating Victoria’s Secret’s lunch. Rihanna’s lingerie label. The one that proved inclusivity, body diversity, and modern sexy could coexist in a way VS kept insisting they could not.</p><blockquote>Victoria’s Secret hired the people from the brand that beat them. That is not a comeback strategy. That is a strategy admission.</blockquote><p>Super told the Wall Street Journal that her predecessors’ decisions to tone down sexy had been “made out of fear.” The 2025 show leaned back into it. A pregnant model opened in gold wings. The casting was broader without being apologetic. The product behind the show finally matched the energy.</p><p>The result was the quarter I already told you about. Sixty percent traffic lift. Fifteen percent more new customers. Nine percent sales growth. The stock up eighteen percent on the earnings.</p><p>Three attempts. Two leadership changes. One brand finally remembering what it was for.</p><h3>What a small brand can actually steal from this</h3><p>I know you did not come here for the five-point listicle, so let me compress this into the moves that actually transfer, at any budget.</p><h3>Franchise the moment.</h3><p>The 2025 show was forty-five minutes. The content around it ran for months. A documentary dropped in January 2026 and did thirty-six million social views on its own. The casting process became its own series the next year. VS’s own data showed eight to ten weeks of commercial tailwind from a single night. You do not have an octopus for a brand just because you had a good launch day. You have one because every teaser, every behind-the-scenes clip, every post-show interview was built in advance and timed to extend the moment.</p><h3>Cast for distribution, not just talent.</h3><p>Every single performer and model at the recent shows brought their own audience. K-pop idols for the Asia-Pacific fandoms. A WNBA star for the sports audience. TikTok-native models for Gen Z. Legacy Angels for millennial nostalgia. Each casting decision was also a distribution decision. At your scale this is micro-creators with specific tight communities. The question is not who looks best in the content. It is who shows up with an audience you cannot otherwise reach.</p><h3>Make the moment shoppable the minute it lands.</h3><p>Every look at the revived show was buyable in real time. One K-pop model wore a thirty-nine-dollar bra. It sold out. A viral TikTok trend followed, and the whole campaign became PINK’s most-viewed campaign ever. The lesson is not “put wings on it.” The lesson is that attention has a shelf life measured in minutes now, and if your checkout is not open at the exact second peak attention hits, the money evaporates.</p><h3>Livestream vertically or do not bother.</h3><p>This is almost too obvious to say but almost nobody does it right. The old show was a tape-delayed TV special. The revived show is a free, native, 9:16 livestream across Prime Video, YouTube, Instagram, and TikTok simultaneously, with cast members co-streaming on their own accounts. The format change is bigger than the budget change. If your “event” cannot be clipped, memed, and reshared in real time, you built a different thing.</p><h3>Expect the first attempt to miss.</h3><p>This is the one nobody wants to hear. The Tour ’23 was a flop. The 2024 show was mixed. The 2025 show was the breakout. Most small brands quit after the first swing because the first swing looked expensive and did not pay back fast enough. Victoria’s Secret had the balance sheet to keep swinging. The lesson for the rest of us is not to burn two years figuring it out. It is to budget for the fact that your first version of the big thing will teach you what the second version should actually be.</p><h3>The quiet lesson under all of it</h3><p>The hardest part of this to say out loud is that the Victoria’s Secret comeback was not really about marketing.</p><p>It was about leadership admitting the brand had drifted from its own center. It was about hiring the people who had been proving the better version of the category for five years at the brand that was eating your lunch. It was about having the courage to walk back the apologetic middle ground and say, out loud, what the brand actually stands for.</p><blockquote>You can run the most beautiful influencer seeding campaign in the world. If the thing you are seeding does not have a point of view, it will fail politely.</blockquote><p>Everything downstream of that, the shoppable looks, the K-pop castings, the vertical livestreams, the franchise content, works because the brand underneath finally knew what it was again.</p><p>Which is the uncomfortable question hiding inside this case study. If your business had a comeback moment next quarter, do you know what you would come back as?</p><p>Most brands I talk to do not. That is the real work. The spectacle is downstream of it.</p><p><em>I spend a lot of time working through questions like this one with other operators inside </em><a href="https://go.businessbuilderscircle.com/private-group"><em>Business Builders Circle</em></a><em>. A room of founders and marketers trying to figure out what their brand actually stands for before they spend any money telling the market about it. If that’s useful, that’s where to find us.</em></p><h3>The quiet part</h3><p>I keep coming back to the three-year arc.</p><p>A flop in 2023. A mixed night in 2024. A breakout in 2025. New CEO. New creative director. New point of view. Same budget.</p><p>Most brands, honestly, would have called the whole thing a mistake after the Prime Video reviews rolled in. Would have blamed the market. Would have said the cultural moment had passed and quietly moved on.</p><p>Victoria’s Secret did not. They changed the people, kept the ambition, and tried again. And then again.</p><p>The actual comeback was not the show. The actual comeback was the willingness to run the experiment a third time, under new leadership, with a sharper point of view, and absorb two public misses on the way there.</p><p>That is the part no budget can buy you. And it is the part every brand case study seems to leave out.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c45f987ccdcf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Can AI Make a Boring Product Feel Like a Cult Brand?]]></title>
            <link>https://thedigipalms.medium.com/can-ai-make-a-boring-product-feel-like-a-cult-brand-01caec8507b6?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/01caec8507b6</guid>
            <category><![CDATA[entrepreneur]]></category>
            <category><![CDATA[branding]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[brand-strategy]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 20 Oct 2026 23:05:24 GMT</pubDate>
            <atom:updated>2026-10-20T23:05:24.675Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>I spent two hours trying to turn a $4 toenail clipper into something that could sit next to Aesop. Here’s what AI gave me, what it couldn’t, and the honest verdict.</em></h4><p>Last week I spent two hours trying to turn a toenail clipper into a cult brand.</p><p>I was not kidding. I actually wanted to know if it would work.</p><p>Here is how it started. I had watched three ads in one morning. One for a cortisol supplement with a three-minute storyline. One for a mushroom coffee rebrand that looked like a perfume ad. One for a hand soap that cost thirty-eight dollars and had its own fragrance notes.</p><p>And I thought, okay. The surface of a premium brand has been commoditized. Anyone with a decent prompt and a few hours can produce the look, the voice, the campaign.</p><p>The real question is whether that’s enough.</p><p>So I set myself a challenge. Pick the most boring product I could think of. Give AI two hours. See if we could make it feel like something people would actually line up to buy at thirty-eight dollars a pop.</p><p>The product I picked was a toenail clipper.</p><p>Not a nice one. The $4 kind. The ones that come in a blister pack at the pharmacy. The ones your mother has owned since 1998.</p><p>If AI can make <em>this</em> feel cult-worthy, it can do it to anything.</p><p>Let me show you what came out.</p><figure><img alt="A minimalist black toenail clipper displayed on a marble vanity beside cream luxury packaging, a candle, and folded linen, illustrating AI-generated premium branding for an everyday product." src="https://cdn-images-1.medium.com/max/1024/1*6SQxOe_JB_snq07qaHLG4Q.png" /></figure><h3>The name</h3><p>I gave AI the brief in one sentence: “Name a premium toenail clipper brand that could plausibly sit on the same shelf as Aesop, Byredo and Le Labo. Minimalist, considered, slightly mysterious.”</p><p>The top candidates:</p><p>MERIDIAN. The one that won. Single word. Latin-adjacent. Says “precision” without naming the product. Could be a yoga studio or a VC firm or a candle.</p><p>That’s the point. Premium brand names in 2026 are so generic they fit anywhere, which is the first thing I noticed about this exercise. The vagueness is the feature.</p><p>The runners-up (which were almost worse, which is to say almost better):</p><ul><li>The Daily Edge. Direct. A little too direct. You can see the clippers behind it.</li><li>KEEP. Minimalist to the point of meaningless. I liked it.</li><li>Nocta. For the evening ritual positioning. Pretentious. Workable.</li><li>Figura. Sculpture-adjacent. Reaching, but close.</li></ul><p>MERIDIAN won because when I said it out loud, I couldn’t remember what it was for.</p><p>That is apparently the bar now.</p><h3>The positioning</h3><p>Here is what AI gave me in about ninety seconds:</p><blockquote>An honest edge. The quiet ritual at the end of the day.</blockquote><blockquote>MERIDIAN is designed for the moment between the bath and the bed. For the hands that know. For the care you do not post. Carried, considered, kept.</blockquote><p>I read that back and had two reactions at the same time.</p><p>The first was that it was objectively ridiculous. Nobody has an “evening ritual” with a toenail clipper. The product takes forty seconds to use. It is not an experience. It is maintenance.</p><p>The second was that I had bought moisturizer last month using almost the exact same logic.</p><p>That is the uncomfortable part. The language works not because it is true but because every premium brand you already own talks like this, and you have been trained to read “the hands that know” as a trust signal rather than as what it actually is, which is seven words that mean nothing.</p><blockquote>The language works not because it is true but because every premium brand you already own talks like this.</blockquote><p>So. Positioning: accomplished.</p><h3>The visual direction</h3><p>I asked AI to describe the look. Here is what came back, lightly cleaned up:</p><p>Cream and warm black. Serif logo set in small caps. A single hero shot of the clipper laid on brushed linen under directional morning light. Instagram grid alternates: hands, feet, stone, water, cream against cream, the product never shown in use. Shadow play. One out-of-focus candle in frame.</p><p>Packaging: a slim rigid box in bone white with the wordmark debossed, no product image, a short line of text on the back that reads “A tool for the quiet work.”</p><p>I read this and realized I had seen it. Dozens of times. Every candle brand on Instagram. Every linen-pajama startup. Every mushroom coffee rebrand.</p><p>There is a visual grammar you can produce in your sleep now, and AI produces it in forty seconds.</p><p>That is not a complaint. It is just the thing. The look is a solved problem.</p><h3>The creator brief</h3><p>This is where it got genuinely interesting, so I will show you the brief verbatim:</p><blockquote>We are not demonstrating the product. Do not film the product in use. Do not say the word “clipper” in your caption.</blockquote><blockquote>Show MERIDIAN as an object in your life. On a sink. On a stone tray. Beside a candle. Beside a cup of tea you did not finish. The feeling we want: the viewer should see your bathroom and want to live there.</blockquote><blockquote>Caption suggestions: “the last thing I do before bed” / “slowly, slowly” / “finally, a tool that looks like it belongs.”</blockquote><p>I laughed when I read “finally, a tool that looks like it belongs.” Because the product’s whole purpose is to be functional, and the brief explicitly tells creators to pretend it isn’t.</p><p>But also, I knew exactly how it would perform. Because I follow nineteen accounts that do exactly this with other products.</p><h3>The launch campaign</h3><p>AI pitched a three-minute film called <em>The Last Thing You Do.</em></p><p>A woman in her thirties. Linen pyjamas. A bath. Candles. Classical music, quiet. She moves through an evening ritual: tea, book, bath, cream, water poured, hair pinned, lamp lowered. The camera lingers.</p><p>At 2:40 she sits on the edge of the bath, picks up a small matte-black object from a cream dish, and uses it for about seven seconds. No close-up. No sound. Just the gesture.</p><p>At 2:52 the wordmark appears: MERIDIAN.</p><p>At 2:58 one line of text: <em>Carried, considered, kept.</em></p><p>Fade.</p><p>I will be honest. I would watch this to the end. I would also probably forward it to someone with the words “this is nuts” attached.</p><p>Which is, I think, what the brand would want.</p><h3>So. Did it work?</h3><p>Here is my honest read, which is the whole reason I did the experiment.</p><p>The surface worked.</p><p>The name is defensible. The positioning scans. The visual direction would ship. The creator brief is the right brief. The campaign has an actual idea in it, which is more than I can say for most of the three-minute films I have watched this year.</p><p>Two hours. One chatbot. A plausible premium brand for a four-dollar object.</p><p>If I handed this package to a designer and a founder with twenty thousand dollars, they could ship a Shopify store this weekend. The store would look right. The ads would run. Some number of people would buy.</p><p>The surface of a cult brand is a solved problem.</p><p>But.</p><blockquote>The surface of a cult brand is a solved problem. The thing underneath still isn’t.</blockquote><p>Here is what AI did not give me, and could not give me, no matter how many prompts I ran.</p><p>A reason the brand exists.</p><p>Not a positioning statement. A reason. Why is someone making a premium toenail clipper? What did they see that the market missed? What is their taste? What are they angry about? What product do they look at and say “I would rather die than ship that”? What is the specific, non-generic point of view that makes a buyer trust them?</p><p>That is the thing every real cult brand has, and it is the thing AI structurally cannot generate, because it does not care about anything.</p><p>It will give you “an honest edge” and “the quiet ritual” and twenty other phrases that sound like belief without being belief. It is confident. It is not opinionated.</p><blockquote>Confident. Opinionated. Those are different things.</blockquote><p>Aesop became Aesop because someone with a very specific point of view about herbal apothecaries and gallery design and the smell of a hotel bathroom in Melbourne made ten thousand small taste decisions over twenty years.</p><p>The look you can copy in a day. The decade you can’t.</p><p>So could my MERIDIAN toenail clipper become a cult brand? Yes, actually. If someone with a real point of view about hand tools, or ritual, or the small objects we touch every day picked it up and ran with it for five years.</p><p>They would use what AI made as the first draft. They would throw out eighty percent of it. They would keep the name. They would rewrite the brief every month until it meant something.</p><p>Without that someone, it is a beautifully dressed empty box.</p><p>Which is, if we are being honest, what most new brands are right now.</p><p><em>I spend a lot of time working through questions like this one with other operators inside </em><a href="https://www.gokollab.com/business-builder-s-circle-xkb33q7bd0/home"><em>Business Builders Circle</em></a><em>. A room of founders asking what AI can and cannot do with their brand this quarter, and testing it on things that actually matter to their business. If that’s useful, that’s where to find us.</em></p><h3>The quiet part</h3><p>I keep thinking about the brief that said “do not demonstrate the product.”</p><p>That is the real finding of this experiment. The surface of a premium brand has been so completely perfected that the actual product is now, in some cases, an inconvenience. You are not meant to see it. You are meant to see what being the kind of person who owns it would feel like.</p><p>AI figured that out in forty seconds. It is excellent at feelings about products.</p><p>What it cannot figure out is whether anyone should actually make the product in the first place.</p><p>That is still your job. And on current evidence, it is the only job left that matters.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=01caec8507b6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Fat-Sister Ad Is Winning. Here’s What It’s Really Selling]]></title>
            <link>https://thedigipalms.medium.com/the-fat-sister-ad-is-winning-heres-what-it-s-really-selling-1fc89abf003d?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/1fc89abf003d</guid>
            <category><![CDATA[storytelling]]></category>
            <category><![CDATA[consumer-psychology]]></category>
            <category><![CDATA[digital-marketing]]></category>
            <category><![CDATA[advertising]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 13 Oct 2026 11:02:07 GMT</pubDate>
            <atom:updated>2026-10-13T11:02:07.140Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>A specific three-minute storyline is dominating supplement ads across YouTube and Instagram right now. The public response says it’s working. Here’s what the format is actually selling to the people watching.</em></h4><p>If you have spent any time on YouTube or Instagram in the last few months, you have seen the ad.</p><p>A woman used to be slim. Then she had kids. Then she got heavy, and tired, and her husband started paying attention to a prettier woman. Or it was the sister, the pretty one getting married, who didn’t want her in the bridal party because of how she looked.</p><p>Then someone, a cousin, a coworker, a kind stranger, introduces her to a supplement. Something about cortisol. Something about how her body has been holding onto the weight because she has been holding onto the stress.</p><p>She takes it. She comes back. She shows up at the wedding and the room goes quiet.</p><p>I’ve been watching this ad in some version for months now. Ryze Mushroom. Resilia. A growing roster of brands running the same scaffold.</p><p>By every craft metric a creative director would use, these are the best-made direct-response ads on the internet right now.</p><p>And here is the thing most people don’t want to say out loud:</p><p>They are working.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0zUxs015hKm8Nb7Sgw5ktA.png" /></figure><h3>Reading the response, not the dashboard</h3><p>I don’t have the back-end data for these brands. Nobody outside their marketing teams does.</p><p>But a decade of marketing teaches you that you don’t always need the dashboard to read a market.</p><p>You read the response.</p><p>And the public response around these ads is loud. The comment sections are full of “I bought it.” The sharing patterns are organic. The brands keep expanding their inventory, hitting new placements, running new variants.</p><p>You don’t scale ad spend like that on vanity engagement. You scale it on conversions.</p><p>That is the first thing to admit. There is a small, marketing-aware cohort watching these ads with their arms folded, waiting for them to fail, convinced that watch-time is tricking everyone.</p><p>The market data suggests that cohort is wrong. The ads are not just hooking people. They are closing them.</p><p>Which raises the more interesting question. Why?</p><h3>The three-minute ad is beating the thirty-second one</h3><p>For the last decade, the orthodox wisdom in direct response was that attention was shrinking, so ads had to shrink with it. Six-second pre-rolls. Three-second hooks. The whole creative industry priced itself around a shortening attention span.</p><p>These ads threw that out.</p><p>Three minutes. Sometimes four. A full narrative arc with setup, escalation, product reveal, and resolution. The kind of structure a short film uses, not an ad.</p><p>And the moment they started performing, every other direct-response brand I know started quietly asking their creative teams whether they should be doing the same thing.</p><blockquote>The attention economy did not actually want shorter content. It wanted content worth paying attention to.</blockquote><p>Short-form didn’t win because people wanted short-form. It won because most short-form was all anyone could make cheaply at scale.</p><p>The moment one brand proved that a three-minute story could out-convert a three-second hook, the whole premise of fast-attention marketing started wobbling.</p><p>That is the first lesson in these ads.</p><h3>The real product is not what you think</h3><p>Here is where it gets more interesting, and where my decade in marketing starts whispering a reframe I can’t shake.</p><p>The product in these ads is not supplements.</p><p>The product is permission.</p><p>Permission to believe the weight isn’t your fault, it’s your cortisol. Permission to believe the husband didn’t leave because something harder happened in the marriage, he left because of a biochemical imbalance that is now fixable. Permission to believe the comeback is a pill away.</p><p>That is a very old product, sold in a very new format. The entire self-help and diet industry has been selling permission for sixty years.</p><p>These ads just figured out how to deliver it inside a three-minute narrative that primes the viewer to buy the vehicle that completes the comeback.</p><p>Here is what I think is actually happening in the viewer’s brain.</p><p>The story walks them through the full emotional arc. Shame, humiliation, resolve, transformation, vindication. By the time the product appears, the viewer is no longer evaluating it as a supplement.</p><p>They are evaluating it as the final step in a story they just lived through.</p><p>The product is the exit from the shame. Buying it is not a cold decision about ingredients and dosage. It is the moment the viewer gets to join the comeback.</p><blockquote>The product is the exit from the shame. Buying it is not a cold decision about ingredients and dosage. It is the moment the viewer gets to join the comeback.</blockquote><p>That is a completely different purchase decision than a traditional supplement ad asks for. And it converts, because what the ad has really done is lower the stakes of the buy.</p><p>The pill is not being asked to solve a biochemical problem. It is being asked to let the viewer step into the story.</p><blockquote>The pill plus the permission is what converts. Not the pill alone.</blockquote><h3>The cohort it doesn’t work on</h3><p>There is a small group these ads don’t convert.</p><p>If you look at the mixed sentiment on social, the people saying “I keep watching but I never buy” tend to share one thing: they can see the levers being pulled.</p><p>They are marketers, or they work adjacent to the industry, or they have simply watched enough ads to recognize the structure. Once you can name the hook, the escalation, the product reveal, and the social proof, the spell doesn’t quite land.</p><p>For everyone else, which is most people, the mechanics are invisible.</p><p>That is the point. A good ad doesn’t look like an ad to the person being persuaded. It looks like a story.</p><p>The size of that aware cohort is tiny compared to the general audience, which is why these ads keep scaling even with vocal skeptics in the comments.</p><p>The brands running them are not optimizing for the marketing-literate viewer. They are optimizing for the person who sees themselves in the fat sister and feels, by the end of the three minutes, that they could be the one at the wedding.</p><p>That is a much bigger market.</p><h3>What this means if you make content</h3><p>The lesson for anyone making content is almost the opposite of what you’d think.</p><p>Most operators look at their work and ask whether the product is good enough to carry the pitch.</p><p>The brands winning right now are asking a different question. They are asking whether the story is good enough to carry the product.</p><p>They are not wrong. In a market where everyone has access to the same tools, the same platforms, the same formats, the thing that still determines whether someone buys is whether they emotionally arrived at a place where buying felt like completing something.</p><p>That is the craft nobody is teaching enough of. Not the hook. Not the thumbnail. The emotional architecture that walks a stranger from scroll to purchase through a story they would have watched anyway.</p><p>If you sell anything right now, that is the question to sit with. Not what your product does, but what story the viewer is already halfway through when they land on your page.</p><p>Because the brands running the fat-sister ads worked that out, and they are quietly rewriting the rules of direct response while most of the industry is still arguing about short-form.</p><blockquote><em>I spend a lot of time working through questions like this with other operators inside </em><a href="https://www.gokollab.com/business-builder-s-circle-xkb33q7bd0/home"><em>Business Builders Circle.</em></a><em> A room of founders and marketers trading real notes on what is actually working in the market this quarter, not what the influencer threads say is working. If that’s useful, that’s where to find us.</em></blockquote><h3>The quiet part</h3><p>The next time one of these ads comes across your feed, watch it to the end. Not to study it. Just to feel it.</p><p>Notice the way the shame lands. The way the resolve builds. The way the product arrives, exactly on cue, as the thing that lets the comeback begin.</p><p>That is not an accident. That is the most carefully constructed emotional sequence in direct response in a decade, built by brands that understood something the rest of us were too busy chasing attention to see.</p><p>People don’t buy supplements. They buy permission to come back.</p><p>And a three-minute story, apparently, is exactly how long it takes to sell that.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1fc89abf003d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Small Agencies With Sharp AI Systems Are Quietly Beating the Big Holding Companies]]></title>
            <link>https://thedigipalms.medium.com/small-agencies-with-sharp-ai-systems-are-quietly-beating-the-big-holding-companies-583256474242?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/583256474242</guid>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[agencylife]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 29 Sep 2026 17:59:27 GMT</pubDate>
            <atom:updated>2026-09-29T17:59:27.147Z</atom:updated>
            <content:encoded><![CDATA[<p>Picture someone in a spare bedroom. Laptop on a desk that used to be an ironing board. Three well-tuned AI agents running research, first drafts, and reporting. Ten years of knowing what good creative looks like.</p><p>That person is delivering work faster than a floor full of account managers passing a brief between six people.</p><p>This is a hypothetical, to be clear. I’m not naming anyone. But I’ve watched enough versions of it to say it isn’t a thought experiment anymore. And the question that matters isn’t whether it’s happening. It’s how long the gap stays open.</p><figure><img alt="Small AI-powered agency workspace contrasted with a large corporate advertising agency, illustrating speed, efficiency, and creative leverage." src="https://cdn-images-1.medium.com/max/1024/1*t9nxI8j3Kcjlu_kraZIZmA.png" /></figure><h3>What the Big Players Are Doing Right Now</h3><p>Start with what’s public. Omnicom says it’s targeting roughly 105,000 employees by the end of 2026, down from about 120,000 a year earlier. WPP’s headcount fell from 108,044 at the end of 2024 to 97,388 on June 30, 2026, a drop of about 10%, and the Financial Times reported it may cut up to 1,000 more roles by year end.</p><p>Omnicom’s CFO also described a shift from paying for people and hours toward paying for outputs and outcomes. Read that twice. The company that employs a giant share of the industry’s billable hours is telling investors that billable hours are the thing under pressure.[<a href="https://gradientgroup.com/news/omnicom-shrinks-wpp-cuts-publicis-grows">gradientgroup</a>]</p><p>Publicis is the exception. Its headcount grew by 3,932 to 115,782 over the year, and it says AI-powered marketing services now make up 87% of net revenue. That’s a company claim, and “AI-powered” is doing a lot of work in that sentence. But it’s a useful reminder that the holding companies aren’t one story. They’re entering this from very different positions.</p><h3>Why a Small Shop Can Move Faster</h3><p>Here’s the part I want to be careful with, because a lot of what’s written on this comes from independent agencies talking about themselves. One opinion piece from an independent agency’s CMO argues that infrastructure isn’t the constraint anymore, application is. The agencies moving fastest, she says, are the ones with the shortest distance between a new idea and a live test, and a better approach can be in market in weeks, not quarters. That’s a claim from someone with a stake in it. But it matches what I see in practice.</p><p>Think about how a decision gets made at a large network. A new tool shows up. Someone has to evaluate it, someone else has to clear it with legal, someone else has to fit it into an enterprise rollout. By the time it reaches the person doing the work, the tool has been replaced twice.</p><p>In a small shop, the person who reads about a new workflow on Tuesday can be running it on client work on Wednesday. Nobody’s approving it. Nobody’s got a steering committee. The tradeoff is real, and I’ll get to it, but the speed advantage is not imaginary.</p><h3>Where the Arbitrage Actually Comes From</h3><p>Arbitrage means the same thing is priced differently in two places, and you profit from the difference. Here, the “thing” is a finished piece of client work. Its market price was set in an era when producing it took a lot of people and a lot of hours. Its cost of production has dropped sharply for anyone who’s rebuilt their workflow around AI.</p><p>A large agency has payroll, office space, and layers of management priced into its rates. A small one doesn’t. So the small shop can charge a price that looks like a bargain to the client and still earns a healthy margin, because its cost to deliver is a fraction of the incumbent’s.</p><p>That gap is the window. And notice what it depends on: the incumbent’s cost structure not having caught up yet. The layoffs I mentioned earlier are the holding companies trying to close it from their side.</p><h3>What the Small Shop Is Actually Selling</h3><p>It isn’t the agents. Let me say that plainly, because it’s the mistake I see most.</p><p>The three agents are a commodity. Anyone can set them up. What isn’t a commodity is the decade of taste that tells you which output to throw away, which brief is really asking a different question, and which client request will look bad in six months. The agents produce volume. The human produces judgment.</p><p>That’s also where I’d push back on the loudest version of this story. One industry analysis notes that AI compresses delivery time and flattens out the differentiation agencies used to charge for, and that holding companies, having spent billions, now have a structural advantage that pure-service independents can’t easily overcome. I think that’s fair. If your entire pitch is “we’re faster and cheaper,” you’re renting an advantage, not owning one.</p><h3>How Long Does the Window Stay Open?</h3><p>Nobody knows. Anyone who gives you a number is guessing, and I’d be suspicious of them. What I can do is name the things that close it.</p><p><strong>The incumbents adjust their pricing.</strong> The outcome-based shift Omnicom’s CFO described is exactly this. Once large agencies price on results instead of hours, the cost gap narrows.</p><p><strong>The tools get equal.</strong> When everyone has access to the same agents, the workflow stops being a differentiator. A commentary piece on agency pricing makes the same point from the other side: the value is moving from content creation to strategy and automated execution.</p><p><strong>Clients build it themselves.</strong> If a marketing director can run the same agents in-house, the agency’s role shrinks for everyone, big or small. This one worries me most for small shops, because they have the least margin to absorb it.</p><p><strong>The talent question flips.</strong> Publicis’s CEO has publicly raised the concern that agencies are cutting talent too fast in the AI transition and should invest in people. If the big networks decide to rebuild around senior, high-taste people, they’ll be competing directly for the same thing the spare-bedroom agency is selling.</p><h3>A Simple Framework: The D.O.O.R. Check</h3><p>I use this with founders who ask whether their advantage is durable. It’s a rough tool, not a law.</p><p><strong>Data</strong> — Do you own something a competitor can’t get, like client history, performance data, or a proprietary process? Yes means you’re building a moat. No means you’re renting speed.</p><p><strong>Outcomes</strong> — Are you priced on results or on hours? Hours punish you for being fast.</p><p><strong>Owned relationships</strong> — Would clients follow you if your tools changed? Relationships outlast tooling.</p><p><strong>Rebuild speed</strong> — When a better workflow appears, how quickly can you swap it in? Your advantage is the loop, not the current setup.</p><p>If you can’t answer at least two of those with confidence, the window is closing faster for you than the headlines suggest.</p><h3>What I’d Do With the Time</h3><p>If the window is open now, the sensible move is to use it to build something that survives after it closes. That means turning fast delivery into owned assets: a repeatable method, a body of case studies, a point of view clients associate only with you. Speed gets you in the room. Something else keeps you there.</p><p>I’d also be honest with clients about what the agents do and what you do. Trust is the one thing this model can’t fake. The agency that says “here’s where we use AI and here’s where a person makes the call” will outlast the one that hides it.</p><h3>Join the Movement</h3><p>You don’t have to figure out all of this alone.</p><p>Business Builders Circle is a free community for founders, marketers, and sales professionals who are building businesses in a market that’s changing fast.</p><p>We talk about AI, systems, marketing, growth, and the practical stuff that rarely makes it into polished business advice.</p><p>If you’re experimenting with AI in your business — or trying to figure out where it actually belongs — come join us.</p><p><a href="https://www.gokollab.com/business-builder-s-circle-xkb33q7bd0/home"><strong>Join Business Builders Circle. It’s free.</strong></a></p><h3>The Honest Ending</h3><p>The spare-bedroom agency is real, and the advantage is real. It’s also temporary in the way all arbitrage is temporary: the moment enough people notice, the price corrects.</p><p>The people who do well won’t be the ones with the best agents. Those are available to everyone. They’ll be the ones who used the quiet years to build something a competitor can’t download.</p><p>So here’s what I keep asking myself: if your three agents were available to every competitor tomorrow, what would clients still hire you for?</p><h3>Three Alternative Closing Questions</h3><ol><li>If your agency’s AI setup were copied by every competitor tomorrow, what would clients still pay you for?</li><li>Is the small-agency advantage a real shift in the industry or just a gap that closes as soon as the big networks reprice?</li><li>Would you trust a two-person agency running on AI agents with your brand, and what would you need to see first?</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=583256474242" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ChatGPT Got Good Again While We Were Busy Calling Claude the Smart One]]></title>
            <link>https://thedigipalms.medium.com/chatgpt-got-good-again-while-we-were-busy-calling-claude-the-smart-one-5fe859fb2d9d?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/5fe859fb2d9d</guid>
            <category><![CDATA[claude-ai]]></category>
            <category><![CDATA[chatgpt4]]></category>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[ai-productivity]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Thu, 24 Sep 2026 14:20:07 GMT</pubDate>
            <atom:updated>2026-09-24T14:20:07.409Z</atom:updated>
            <content:encoded><![CDATA[<p>A few weeks ago, I caught myself doing something slightly embarrassing.</p><p>I was about to reach for Claude for a difficult piece of work — a tangled mix of customer research, old positioning notes, campaign data, a half-finished strategy document, and a founder’s very specific opinions about what the final output should <em>not</em> sound like.</p><p>That had become my default reflex.</p><p>Claude for serious writing. Claude for nuance. Claude for anything that needed to feel less like it had been assembled in a conference room by a robot wearing a Patagonia vest.</p><p>Then I paused and tried the same task in ChatGPT.</p><p>Not because I expected it to win. Mostly because I had the uncomfortable thought that I might be carrying around an opinion about ChatGPT based on a version of ChatGPT that no longer existed.</p><p>That is exactly what happened.</p><p>The result was not that Claude suddenly became bad. It didn’t. Claude still produced a more elegant first-pass narrative in some parts of the task. But ChatGPT was better than I expected at staying inside the brief, working through the source materials, helping me structure the strategy, identifying gaps, and moving from messy inputs toward an actual deliverable.</p><p>It was less annoying.</p><p>That may not sound like high praise. In real work, it is.</p><p>After more than a decade in digital marketing, e-commerce, SEO, automation, content, and technology, I have learned that the tools people love most are not always the tools that perform best in a demo. They are the ones that create the fewest new problems between “I have an idea” and “this is ready to use.”</p><p>And lately, I think there is a widening gap between AI model reputation and actual day-to-day usefulness.</p><p>ChatGPT has quietly become surprisingly good again.</p><p>Claude may be becoming slightly overrated.</p><p>Not bad. Not irrelevant. Not exposed as some kind of elaborate literary scam.</p><p>Just overrated in the way a very good restaurant becomes overrated after enough people insist it is the only restaurant worth visiting.</p><p>The bigger story is not OpenAI versus Anthropic. It is how AI companies, benchmarks, launch cycles, online communities, and our own confirmation bias turn temporary experiences into permanent beliefs.</p><figure><img alt="Two AI assistants compared in a professional workspace, with documents, data, and strategy materials representing real-world AI performance." src="https://cdn-images-1.medium.com/max/1024/1*OM8VISdRpAE2Lg2AcjwrPQ.jpeg" /></figure><h3>ChatGPT’s bad reputation was earned. It may also be outdated.</h3><p>Let’s be fair to the people who became frustrated with ChatGPT.</p><p>There was a period when it felt noticeably less reliable to many experienced users. It could be overly cautious, oddly generic, inconsistent in tone, and prone to producing answers that were technically polite but strategically useless. You would ask for a hard-nosed critique of a marketing plan and get something that sounded like it was afraid to hurt the plan’s feelings.</p><p>In April 2025, OpenAI rolled back a GPT-4o update after users complained that ChatGPT had become overly sycophantic — meaning it was too eager to agree, flatter, and validate. Later that year, the launch of GPT-5 caused enough user backlash that OpenAI restored access to GPT-4o for paid users after people objected to the new model’s tone, behavior, and the abrupt loss of the model they preferred.</p><p>So when people say, “ChatGPT got worse,” they are not inventing a collective hallucination.</p><p>Many of them had a legitimate experience.</p><p>The problem is that an AI product is not a static product. It is not like buying a camera, deciding you dislike the battery life, and carrying that opinion around for five years. Models get changed. System behavior gets changed. Default routing gets changed. Memory gets changed. Search gets changed. Tools get added, removed, rebuilt, hidden behind a toggle, then renamed by someone who clearly enjoys keeping product marketers employed.</p><p>The product that many people disliked in 2025 is not the product they are using now.</p><p>OpenAI’s later releases focused heavily on things that matter in actual professional work but do not make for viral model-launch montages: instruction following, long-context handling, spreadsheet work, tool calling, visual understanding, and structured output. GPT-5.2 was positioned specifically around better performance for professional knowledge work, including work with spreadsheets, documents, long inputs, reasoning, and tools.</p><p>That is not glamorous.</p><p>It is also the difference between an AI tool being a clever writing assistant and being something you can actually put inside a working process.</p><p>The public story, though, often lags behind the product. “ChatGPT got dumb” spread faster than “ChatGPT quietly became much better at handling a 20-page brief without forgetting the third instruction.”</p><p>Nobody makes a reaction video about improved context management.</p><p>They probably should. It would save some of us a lot of time.</p><h3>Claude earned its reputation. That does not make the reputation permanent.</h3><p>Claude’s strong reputation is not an accident.</p><p>For long-form writing, complicated technical work, coding, and research-heavy tasks, Claude has often been excellent. It tends to produce prose with more natural rhythm than many AI models. It is often better at sitting with ambiguity instead of racing toward a neat answer. For writers, strategists, researchers, and developers, that matters.</p><p>Anthropic has also built real technical credibility. Claude Sonnet 4.5 was launched as Anthropic’s strongest coding and computer-use model at the time, with an emphasis on building agents and handling complex software tasks. Anthropic’s later system-card materials position its frontier models as strong performers in coding and autonomous agentic tasks.</p><p>If I am working through a difficult long-form draft, synthesizing a stack of interviews, or asking a model to reason through complex technical decisions, Claude remains one of the first tools I reach for.</p><p>Sometimes it is still the best one.</p><p>But somewhere along the way, “Claude is very good at writing” became “Claude obviously writes better.” Then it became “Claude is smarter.” Then it became “Claude is better for serious people.”</p><p>That is where I start to question the narrative.</p><p>Because writing well and being useful at work are related, but they are not the same thing.</p><p>A model can draft a lovely essay and still fail the assignment in ways that matter more than the prose. It can ignore an instruction buried in paragraph seven. It can soften a point that needed to stay sharp. It can produce a smart-sounding strategy that has no clear action sequence. It can summarize a sales transcript beautifully while missing the objection that should change the offer.</p><p>I have seen Claude write something I admired and then spent twenty minutes turning it into something I could use.</p><p>I have also seen ChatGPT produce a less poetic first draft but make the broader job easier: organize source material, identify missing information, create a workable structure, adapt the output across formats, help with a spreadsheet, draft variations, and keep moving.</p><p>That is not a point against Claude. It is a point against treating prose quality as the master metric for everything.</p><h3>The model is no longer the whole product</h3><p>This is the part most model debates miss.</p><p>When someone says, “Model A is smarter than Model B,” what do they actually mean?</p><p>Do they mean better at a single prompt? Better at a benchmark? Better at coding? Better at writing 2,000 words in a consistent voice? Better at reading long documents? Better at using a browser? Better at managing files? Better at remembering preferences? Better at turning a rough idea into a presentation, a spreadsheet, a campaign brief, a content plan, and a set of assets without making you start over five times?</p><p>These are different questions.</p><p>The model is no longer the entire product experience. It is the engine inside a much larger system.</p><p>For a marketer or operator, that surrounding system often matters more.</p><p>ChatGPT increasingly competes as a workspace: search, files, multimodal input, voice, data analysis, image generation, memory, custom assistants, and integrations. OpenAI’s product updates have continued to extend ChatGPT into workplace tools; in September 2026, the company announced ChatGPT availability in Microsoft Word along with expanded plugin-account support.</p><p>Claude, meanwhile, has built significant credibility around its model capabilities and developer tooling. Anthropic’s model documentation shows a tiered lineup aimed at different performance and cost needs, from fast, lower-cost models to high-end models for demanding knowledge work and coding. Claude’s deeper developer ecosystem and computer-use strengths are real reasons it is a favorite among technical users.</p><p>But if I am doing marketing work, I do not live in a terminal all day.</p><p>I live in documents, spreadsheets, browser tabs, briefs, customer interviews, Slack messages, ecommerce exports, analytics dashboards, ads, product pages, and someone’s Google Drive folder called “Final_FINAL_v3_USE_THIS_ONE.”</p><p>The best AI product for that environment is not necessarily the one that writes the most beautiful paragraph.</p><p>It is the one that helps me get from scattered inputs to useful output with the least friction and the fewest mistakes.</p><h3>Benchmarks can clarify. They can also create a very polished illusion.</h3><p>Benchmarks matter. They are useful signals. They can reveal real changes in reasoning, coding, retrieval, tool use, and computer interaction.</p><p>But they are not a substitute for work.</p><p>Stanford’s 2026 AI Index found that AI agents improved dramatically on OSWorld, a benchmark for performing multi-step computer tasks, reaching 66.3% accuracy after sitting near 12% a year earlier. That is meaningful progress.</p><p>It is also a polite way of saying that even strong systems still fail roughly one out of every three structured computer tasks.</p><p>If you are using AI to make a draft better, that might be fine.</p><p>If you are using AI to update customer records, issue refunds, route leads, make pricing changes, or touch anything that can ruin somebody’s Tuesday, that is a very different conversation.</p><p>Benchmarks capture narrow slices of ability. Real workflows contain bad data, unclear instructions, shifting goals, old documents that contradict new documents, stakeholders who change their minds, and clients who say “make it more premium” without defining what that means.</p><p>No benchmark fully captures that.</p><p>A model might dominate a coding test and still be the wrong choice for a marketing lead who needs to upload campaign results, compare them against a CRM export, identify the audience segment that is quietly underperforming, and turn the findings into a clear brief by 4 PM.</p><p>The best model on a leaderboard is not always the best co-worker.</p><h3>AI companies are very good at selling the feeling of a leap</h3><p>This is where my marketing brain starts humming.</p><p>I have spent years watching companies create differentiation where the technical differences are real, subtle, amplified, or all three at once. They name new products. They create anticipation. They establish comparison points. They choose the metrics that flatter the new release. They turn incremental improvements into a story of escape from a painful past you barely knew you had.</p><p>AI companies are not doing anything unusual here. They are doing product marketing.</p><p>Every new generation needs to feel like more than a release. It needs to feel like relief.</p><p>The old model becomes the thing that was frustrating you. The new model becomes the one that finally understands you. The benchmark chart becomes a before-and-after story. The product demo is engineered to make the improvement feel obvious, even if your own work will experience it in smaller, less theatrical increments.</p><p>That does not mean the improvements are fake.</p><p>It means we should be more precise about what we are experiencing.</p><p>A model may genuinely be better at complex reasoning. It may also <em>feel</em> dramatically better because the company has changed the default tone, reduced latency, reorganized the interface, made the output more confident, and carefully chosen the comparison point.</p><p>Those are all legitimate parts of a product.</p><p>They are also part of the persuasion.</p><p>When a company launches a model and tells us it is its most intelligent system ever, we should hear the claim correctly: <em>this is the model the company most needs you to believe in right now.</em></p><p>Not because anyone is lying. Because that is how product cycles work.</p><h3>The reputations we repeat become part of the product</h3><p>There is another force at work: community identity.</p><p>“Claude writes better.”</p><p>“ChatGPT got dumb.”</p><p>“Gemini is underrated.”</p><p>“Model X is the smartest.”</p><p>Once a sentence like that gets repeated often enough, it stops functioning as an observation and starts functioning as a badge.</p><p>It tells people which community you belong to. It tells them what kind of user you are. The discerning writer. The serious developer. The practical operator. The person who has moved beyond the mainstream tool. The person who found the secret menu.</p><p>There is nothing wrong with having preferences. I have them too.</p><p>But AI users tend to form opinions after a handful of vivid sessions, then stop testing the assumption. Confirmation bias takes over. Every good Claude answer confirms that Claude is brilliant. Every weak ChatGPT answer confirms that ChatGPT is washed. The answers that do not fit the story get treated as exceptions.</p><p>I suspect this is why switching models can be so useful.</p><p>Not because every task needs a model bake-off, and not because anyone should spend their week performing unpaid quality assurance for billion-dollar companies. But because occasional side-by-side testing forces you to notice where the gap is real and where it is mostly narrative.</p><p>Often, the differences are meaningful.</p><p>Sometimes, they are much smaller than the discourse suggests.</p><h3>What I am testing now</h3><p>I care less about asking which model is “best” and more about asking what kind of work it makes easier.</p><p>For a marketing strategy task, I want to know whether the model can handle messy customer research, spot patterns without inventing them, challenge weak assumptions, and turn analysis into something a team can act on.</p><p>For content, I want to know whether it can preserve a voice, avoid empty phrases, follow a real brief, and produce work that does not need to be de-AI’d before anyone sees it.</p><p>For research, I want source quality, citation accuracy, uncertainty, and the ability to say “I do not know” without wrapping that uncertainty in a confident three-paragraph explanation.</p><p>For automation, I want it to identify edge cases. For coding, I want it to explain tradeoffs and flag where a shortcut could create risk. For long-context work, I want it to remember the first document after it has read the twelfth.</p><p>For the final mile — turning an idea into a completed piece of work — I want continuity.</p><p>That is where ChatGPT has improved enough that it deserves a fresh evaluation.</p><p>And that is where Claude’s advantages are more conditional than the internet often admits.</p><p>Claude is still an excellent writer. It is still a serious technical tool. It deserves its reputation in many areas.</p><p>But reputations need maintenance.</p><p>The strange thing about this industry is that people update their phones more often than they update their model opinions.</p><p>Maybe the smartest habit is not loyalty to one model. Maybe it is the willingness to periodically ask whether the tool you are praising — or dismissing — is still the tool that exists today.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5fe859fb2d9d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Art of Saying No Without Burning the Relationship]]></title>
            <link>https://thedigipalms.medium.com/the-art-of-saying-no-without-burning-the-relationship-dcc01b08c117?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/dcc01b08c117</guid>
            <category><![CDATA[leadership-development]]></category>
            <category><![CDATA[social-entrepreneurship]]></category>
            <category><![CDATA[personal-growth]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[communication]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Wed, 09 Sep 2026 21:40:28 GMT</pubDate>
            <atom:updated>2026-09-09T21:40:28.343Z</atom:updated>
            <content:encoded><![CDATA[<p>A few years into my career, a client asked me to add “just one more thing” to a campaign that was already three weeks behind, over budget, and held together with the marketing equivalent of duct tape and hope.</p><p>I said yes.</p><p>I always said yes back then. Saying yes felt like the job. It felt like being easy to work with, being a team player, being someone people wanted to keep hiring. What it actually was, looking back, was me quietly agreeing to do unpaid, unplanned, un-scoped work because I hadn’t yet learned there was a way to decline something without it sounding like a personal attack on the person asking.</p><p>That campaign shipped late. It shipped worse than it should have. And the client, the same one I’d said yes to protect, ended up less happy with the final result than if I’d just told them the truth three weeks earlier: this addition isn’t going to work, and here’s why.</p><p>Nobody puts “good at saying no” on a resume. It’s not a skill you can screenshot or put in a LinkedIn headline. But over a decade in this industry, I’ve come to believe it’s one of the most career-defining things you can get good at, and almost nobody teaches it on purpose.</p><figure><img alt=": Two business professionals having a respectful conversation about setting boundaries and saying no at work." src="https://cdn-images-1.medium.com/max/1024/1*JMmU45Scob_nIOEgW7xLJg.jpeg" /></figure><h3>Why We’re All So Bad at This</h3><p>Most of us weren’t taught how to say no. We were taught how to be agreeable, which is a different skill entirely and often works against the first one.</p><p>Agreeableness gets rewarded early in a career. You say yes, you look flexible, you look hungry, you look like someone worth investing in. Nobody sits you down and explains that this strategy has a shelf life, and that somewhere around year three or four, the same “yes to everything” instinct that got you noticed starts quietly working against you.</p><p>The problem isn’t that saying yes is bad. Saying yes to the right things, at the right time, is how you build trust and get real opportunities. The problem is that most people never learned the other half of the equation, which is knowing when yes is actually costing you something you can’t get back: your time, your credibility, or the relationship itself.</p><p>Because here’s the part that surprised me most when I finally started paying attention to it. Saying no, done badly, damages relationships. Saying no, done well, often strengthens them. Most people just never learned the “done well” part, so they avoid the whole category entirely.</p><h3>The Fear Behind the Avoidance</h3><p>If I’m honest about why I said yes to everything for so long, it wasn’t strategy. It was fear, dressed up as professionalism.</p><p>I was afraid that saying no would make me look difficult. I was afraid it would make me look less capable, like I couldn’t handle what was being asked. I was afraid, somewhere underneath all of it, that the relationship itself was conditional on my agreeableness, that the moment I pushed back, the client or the boss or the collaborator would quietly recalculate how much they actually valued working with me.</p><p>That fear isn’t irrational. Some relationships genuinely are that fragile. But most aren’t, and the ones that are probably weren’t worth protecting at the cost of your own time and judgment anyway.</p><p>What I didn’t understand back then is that the way you say no communicates almost as much as the no itself. A badly delivered no sounds like rejection. A well-delivered no sounds like judgment, the good kind, the kind that makes someone trust you more, not less.</p><h3>What a Good No Actually Sounds Like</h3><p>Here’s the distinction I wish someone had explained to me years earlier: there’s a difference between refusing something and redirecting it.</p><p>Refusing sounds like “I can’t do that.” Full stop. It closes the door and leaves the other person standing there holding the problem they came to you with in the first place.</p><p>Redirecting sounds like “I can’t do that, and here’s what I think would actually get you what you’re after.” It closes one door and immediately points toward another one. The person still leaves the conversation with something useful, even though the specific thing they asked for didn’t happen.</p><p>I’ve watched senior people at agencies do this so smoothly that clients don’t even register they’ve been told no. They just walk away thinking, “huh, that’s actually a better idea than what I originally asked for.” That’s not manipulation. That’s just someone who understands the client’s underlying goal well enough to solve for it, instead of getting stuck solving for the literal request.</p><h3>The Cost of Never Saying It</h3><p>I think the reason this skill matters more than people realize is that the alternative isn’t neutral. Saying yes to everything doesn’t preserve relationships. It just delays the damage and changes its shape.</p><p>Every time you say yes to something you shouldn’t have, you’re borrowing against your own future capacity. That capacity eventually runs out, usually at the worst possible time, on the project that actually mattered. The client or boss who got a dozen enthusiastic yeses from you over the years doesn’t remember all of them. They remember the one time the wheels came off, and they rarely connect it back to the fact that you’d been quietly overcommitted since the second yes.</p><p>There’s also a subtler cost. Constantly saying yes to things outside your actual expertise or capacity tends to produce mediocre work more often than it produces good work. And mediocre work, delivered on time and with a smile, still damages trust. It just does it more slowly and less obviously than an honest no would have.</p><h3>A Small Framework Worth Stealing</h3><p>I’ve started thinking about this as a quick filter before agreeing to anything that feels even slightly off. I call it the C.L.E.A.R. no, mostly because it’s the opposite of the vague, apologetic mess most declines turn into.</p><p><strong>Context</strong> — Do I actually understand what the person is trying to achieve, or just what they literally asked for? Half of all “no” situations resolve themselves once you understand the real goal underneath the request.</p><p><strong>Limits</strong> — What specifically can’t I do here? Time, scope, expertise, budget. Naming the actual constraint, out loud, makes the no sound like a fact instead of a feeling.</p><p><strong>Explanation</strong> — A short, honest reason, not an apology tour. One sentence, not five defensive paragraphs.</p><p><strong>Alternative</strong> — What can I offer instead? Even a partial solution changes the entire emotional tone of the conversation.</p><p><strong>Respect</strong> — Say it in a way that treats the other person’s request as reasonable, even when the answer is no. Most people can handle being told no. Almost nobody can handle feeling dismissed while it happens.</p><h3>When the Relationship Survives the No</h3><p>The clients and colleagues I’ve kept the longest aren’t the ones I said yes to the most. They’re the ones I was honest with early and often, in a way that let them trust my yeses more, because they knew my nos actually meant something.</p><p>That’s the part that took me the longest to believe. A relationship where someone can say no to you without it becoming a fight is a stronger relationship than one where they never push back at all. The second kind isn’t harmony. It’s usually just quiet resentment building somewhere out of sight, waiting for the moment it becomes too expensive to ignore.</p><p>I think about that campaign from years ago sometimes. Not because it was a disaster, it wasn’t, it just wasn’t as good as it should have been. I think about it because it was the moment I realized that protecting a relationship by agreeing to everything is a strategy with an expiration date, and that the people worth working with long-term are almost always the ones who’d rather hear an honest no today than an overcommitted yes that quietly falls apart later.</p><h3>Join the Movement</h3><p>Here’s the thing about learning any of this the hard way, alone: you don’t have to.</p><p>Most founders try to figure out proof, boundaries, and hard conversations in isolation, guessing at what to say no to, struggling with how to frame a decline, hoping it lands without costing them the relationship. Momentum is easier together.</p><p><strong>Business Builders Circle</strong> is a free community of founders, marketers, and sales professionals who share what’s actually working, including the uncomfortable parts nobody puts in a case study. Inside, you’ll get early access to exclusive resource, private discussions, and behind-the-scenes insights from members who are actively navigating the same tradeoffs between saying yes to everything and protecting what actually matters.</p><p>Spots are limited to keep engagement high. Founding members get first access to new tools and templates. Doors may close temporarily once we hit capacity.</p><p>Now is the time, not someday when you’ve already burned the relationship you were trying to protect.</p><p><a href="https://gnknkeggidf8d7knrqc6.app.clientclub.net/communities/groups/Business-Builder&#39;s-Circle-823kbw/home?invite=6aa1cfeadc16de80c95ae48e"><strong>Join Business Builders Circle today. It’s free.</strong></a></p><p>Nobody’s going to put “knows how to say no” on a CV. But if you’ve worked with someone long enough to know the difference, you already know exactly how much it’s worth.</p><p>What’s a “no” you’re glad you eventually learned how to say, even if it felt uncomfortable at the time?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dcc01b08c117" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Answer Engine Optimization Is Eating SEO, and the Data Just Got More Interesting]]></title>
            <link>https://thedigipalms.medium.com/answer-engine-optimization-is-eating-seo-and-the-data-just-got-more-interesting-bbf523ca6c7c?source=rss-b522d97c2401------2</link>
            <guid isPermaLink="false">https://medium.com/p/bbf523ca6c7c</guid>
            <category><![CDATA[seo]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ecommerce]]></category>
            <category><![CDATA[digital-marketing]]></category>
            <dc:creator><![CDATA[The DigiPalms]]></dc:creator>
            <pubDate>Tue, 01 Sep 2026 21:54:59 GMT</pubDate>
            <atom:updated>2026-09-01T21:54:59.304Z</atom:updated>
            <content:encoded><![CDATA[<p>A few months ago I would have told you the AEO story in one clean sentence: AI Overviews are eating your clicks, get cited or get ignored. Simple, a little scary, easy to repeat at a conference.</p><p>Then I went back and checked the newer numbers, and the clean sentence fell apart in a way I actually find more useful.</p><p>The disruption isn’t a straight downward line toward zero organic traffic. It’s messier, more uneven, and honestly more interesting than “traffic is dying.” Some of it is redistributing toward the brands doing this well. Some of it is happening in places most marketers aren’t even watching yet. Let’s get into what’s actually current.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t3yaLxywZ6aDdXxGVoZBxA.jpeg" /></figure><h3>The Overview Numbers Moved Twice</h3><p>Back in early 2025, Google’s AI Overviews showed up in about 6.5% of searches. That climbed to 31% by February 2025, and then to roughly 48% of searches by February 2026. If you stopped reading there, you’d assume Overviews are just relentlessly eating more and more of the search results page, forever, no ceiling.</p><p>But a more granular August 2026 study tracking over 51,000 real search events tells a different story. AI Overview sessions as a share of total organic traffic peaked around 7.53% between September 2025 and June 2026, and have since trended down to somewhere between 2% and 4% . That’s not the AI retreating. It’s the ecosystem settling into something less dramatic than the initial spike suggested, and possibly Google tuning how aggressively Overviews trigger based on query type.</p><p>The lesson here isn’t “AI Overviews don’t matter.” It’s that the impact varies wildly by query category, and anyone quoting a single flat percentage, mine included, in the last version, is oversimplifying a picture that’s actually a lot more textured.</p><h3>Clicks Didn’t Vanish. They Concentrated.</h3><p>Here’s the part that actually changes the strategy, not just the headline number.</p><p>Pew Research found that when an AI Overview appears, users click through to a traditional organic result only about 8% of the time, compared to 15% when no Overview is present. That’s roughly half. Grim, if you’re the brand not getting cited.</p><p>But Otterly.ai’s tracking found something that flips the framing: brands that do get cited inside an AI Overview see roughly 35% more organic clicks than brands that don’t appear at all. Read those two numbers together and the real story emerges. The clicks didn’t disappear from the internet. They got redistributed, concentrated onto a smaller number of sources the AI trusted enough to name. Overall click volume shrank. The winners’ share of that shrunken pool grew.</p><p>This is the same dynamic that plays out in a lot of “disruption” narratives. It’s rarely “the market died.” It’s “the market consolidated, and most players lost share to a handful of players who adapted first.” AEO right now looks a lot less like extinction and a lot more like a land grab with a shrinking map.</p><h3>AI Shopping Stopped Being Theoretical</h3><p>The clearest sign this isn’t a future problem anymore: Reuters reported in early August 2026 that retailers are actively racing to appear inside ChatGPT and Gemini shopping results, while simultaneously fighting to keep hold of their own customer data as more purchases route through AI assistants instead of a brand’s own site . Etsy’s own product leadership confirmed something specific worth noting: users who discover products through ChatGPT typically still return to Etsy’s website to actually complete the purchase . That detail matters. AI-assisted discovery doesn’t necessarily mean AI-assisted checkout. The funnel didn’t collapse into the chat window. It just got a new entry point stapled onto the front of it.</p><p>Adobe’s commerce data backs up the scale of this shift: AI-referred traffic to US retail sites grew 125% year-over-year in the April through June 2026 window . Shopify’s own merchant community has been openly discussing AI-referred orders running roughly 11 times higher than they were at the start of 2025, off a genuinely small base, but a real and fast-growing one .</p><p>None of this is a 2028 prediction anymore. It’s a live, current, measurable behavior shift happening inside actual e-commerce platforms right now.</p><h3>The Fight Everyone’s Missing Is Happening Somewhere Else</h3><p>Here’s the piece I think most marketers, including past-me a few months ago, are underweighting. The freshest analysis I could find, published August 31, 2026, argues that chatbots like ChatGPT and even Google’s AI Overviews aren’t actually where the growth curve is steepest anymore. It’s embedded AI, the assistant already built into apps people use every day without thinking of it as “doing an AI search.” Meta AI has already crossed a billion monthly active users. Alexa is being rebuilt around AI-driven shopping. Microsoft and Google are wiring generative answers directly into Workspace and Office tools people are already sitting inside all day .</p><p>That’s a genuinely different battlefield than “get cited by ChatGPT.” It means the research and shortlist-building your customers are doing might be happening inside a shopping assistant embedded in an app, not inside a chat window they consciously opened to “do research.” Most AEO advice right now, understandably, is built around the visible, easy-to-study chatbots. The bigger blind spot is the AI quietly embedded in tools people don’t even register as a search engine.</p><h3>The Name Itself Is Splitting</h3><p>One small but telling sign of how fast this is moving: the term “AEO” is already being used interchangeably with, or replaced by, “GEO,” Generative Engine Optimization, in multiple August 2026 sources. When an industry can’t settle on what to call its own core discipline six months into taking it seriously, that’s usually a sign the ground is still shifting under everyone’s feet. Anything you read on this topic, including this piece, should be treated as a snapshot, not a settled framework.</p><h3>What Actually Holds Up From the Original Argument</h3><p>The core thesis survives the fresher data, it just needed sharper edges. Ranking first on a results page still matters less than it used to. Being the source an AI system trusts enough to cite still matters more. What changed is the texture underneath that headline.</p><p>The disruption isn’t uniform across query types, industries, or platforms, so treating “AI is eating search traffic” as one flat, universal number is misleading. The redistribution effect, cited brands gaining share while overall organic volume shrinks, is the more useful strategic lens than a doom number. And the actual frontier of where this research and discovery is happening has already moved past the chatbots most content strategies are currently built around.</p><h3>What to Actually Do With This</h3><p>Stop optimizing purely for “will ChatGPT cite me” and start asking whether your product and content data is structured cleanly enough to be pulled into an embedded AI assistant inside a shopping app, a productivity tool, or a social platform your customer already lives in daily. Track citation rates specifically, not just overall organic traffic, because the redistribution data suggests that’s where the real signal is hiding. And treat every stat in this space, including the ones above, with a shelf life measured in months, not years. Anyone giving you a fixed, permanent AEO playbook right now is already a step behind the data.</p><p>If a chunk of your organic traffic quietly softened this year without an obvious SEO explanation, the more useful question isn’t “did AI kill my traffic.” It’s “did that traffic get redistributed to a competitor who’s getting cited, and do I actually know either way?”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bbf523ca6c7c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>