<?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[CTO Insights - Medium]]></title>
        <description><![CDATA[We step in when others step out. We enhance legacy code and elevate tech teams to set you up for long-term success. - Medium]]></description>
        <link>https://medium.com/we-are-madewithlove?source=rss----39bf59ff8129---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>CTO Insights - Medium</title>
            <link>https://medium.com/we-are-madewithlove?source=rss----39bf59ff8129---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 07:31:14 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/we-are-madewithlove" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Revolutionising diagnosis: How large language models can drive a 10x change]]></title>
            <link>https://medium.com/we-are-madewithlove/revolutionising-diagnosis-how-large-language-models-can-drive-a-10x-change-b3482897aaab?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/b3482897aaab</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[madewithlove]]></dc:creator>
            <pubDate>Tue, 01 Jul 2025 13:21:29 GMT</pubDate>
            <atom:updated>2025-07-01T13:21:29.511Z</atom:updated>
            <content:encoded><![CDATA[<h3>Using large language models (LLMs) as creative collaborators</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*puGLLzMp6v2t7UDN80pF9w.jpeg" /></figure><p>Albert Einstein is (incorrectly) credited with saying that <strong><em>Creativity is intelligence having fun</em></strong><em> </em>— the source of which is unknown. Being the most intelligent species on our planet has meant that our creativity is the highest form of intelligence having fun. That fun is <strong>expressed</strong> in many ways — art, music, painting and more. We’ve had tools to help with the many forms of creative expression, but a <strong>human</strong> has always been the “intelligence”.</p><p>Large language models (LLMs) are the latest <strong>tool</strong> to help us in our creative endeavours. However, this is no mere <strong>incremental</strong> improvement. LLMs could be what Andy Groove would call <em>a 10x change</em>. He described it as:</p><blockquote>What such a change does to a business is profound and how the business manages this transition determines its future.</blockquote><p>The concept of the last interface is that humans interact with a single AI system that enables them to achieve all their needs with technology. This article explains the possibility of LLMs being a <strong>10x technology</strong> and explores why they will become one of the last interfaces we use to interact with computers.</p><h3>LLMs — The (second) last interface</h3><p>LLMs challenge a long-held idea that humans have cornered the market on creativity. For the first time in history, machines can, on certain dimensions, rival human creativity. Specifically, LLMs based in the realm of problem diagnosis and solution.</p><p>There are several fields where problems require diagnosis by a human. These fields include <strong>healthcare</strong>, <strong>customer support</strong>, <strong>legal</strong>, and <strong>urban planning</strong>. Creating high-impact software solutions for these spaces is challenging, and a lot of <a href="https://madewithlove.com/blog/a-companys-first-steps-in-ai/">companies are struggling taking their first steps in AI</a>. This difficulty comes from the fact that conventional software solutions are rigid in their design. Typical software solutions attempt to guide the user through a linear process. However, exploring problems in the above fields requires flexibility, particularly across numerous contexts that are both difficult to design in advance and are ever-expanding. Their dynamic nature compounds this.</p><blockquote><em>A computer’s ability to participate in a contextually relevant conversation is a level of creativity never seen before in a computer. A computer becomes a true </em><strong><em>conversational partner</em></strong><em>.</em></blockquote><p>LLMs enable the versatility that these <strong>contexts</strong> require, allowing a user to interact in natural language. Natural language results in a conversation with a computer, i.e. a non-linear, iterative, and explorative experience. As a diagnosis is naturally an <strong>iterative</strong> and <strong>explorative</strong> experience, it can be stated that conversations are one of the best ways to diagnose a complex problem. A <strong>conversation</strong> with a computer instantly elevates it above any conventional software experience. A computer’s ability to participate in a contextually relevant conversation is a level of creativity never seen before in a computer. A computer becomes a true <strong>conversational partner</strong>.</p><h3>An experience inflection point</h3><p>Some of society’s most valuable services today are ones where experts diagnose complex problems and <strong>recommend</strong> courses of action. Access to doctors, lawyers and accountants is time-constrained and limited. This article will refer to the ability to diagnose complex problems and recommend actions as a form of creativity.</p><p>Utilising LLMs to design solutions around this time-constrained expertise will <strong>democratise</strong> access. LLMs’ inflection point is in bringing this particular human creativity to more people faster and at a lower cost.</p><h3>Healthcare — scaling your doctor’s capabilities</h3><p>An application of this concept is in the realm of <strong>medical services</strong>. Ireland, for example, is facing significant challenges in delivering healthcare. Many towns have fewer doctors than the WHO (World Health Organisation) recommended density of 100 doctors/100k citizens. The shortage has particularly impacted those over the age of 75 who on average visit their doctor 6.5 times/year. This is in comparison to the middle-aged who visit 3.8 times/year and younger people who visit 3–3.7 times/year ().</p><p>A conversation-based digital assistant that could access the patients medical records would empower the individual to have a greater level of independence to manage their health. With appropriate levels of training and autonomy, the assistant could handle the essential tasks of prioritising and booking appointments, suggest medication refills, or offer tailored advice. Patients get faster and cheaper access to tailor-made assistance. Overall, this would lower the <strong>barrier</strong> to entry across all demographics and provide faster and more cost-effective support.</p><h3>Time, cost, and uncertainty — the 3 obstacles</h3><p>Paul Grice introduced a set of rules ( <strong>Maxims</strong>) for conversations: <strong>Quantity</strong> (how much you say), <strong>Quality</strong> (how truthful you are), <strong>Relevance</strong> (staying on topic), and <strong>Clarity</strong> (being clear and not confusing). It describes how people converse. Here’s how these rules transfer to LLMs.</p><h3>Quantity: Say Just Enough</h3><p>For LLMs, quantity means they need to provide just the right amount of information-not too much, and not too little. LLMs know a lot because they&#39;re trained on vast amounts of <strong>text</strong>, but they don&#39;t know everything, especially private or niche information. To make up for this, LLMs need a way to get extra info without constantly retraining.</p><h3>Quality: Be Truthful</h3><p>Quality means the LLM should provide accurate, <strong>reliable</strong> information. But there&#39;s a catch: LLMs sometimes make up things when they don&#39;t know the answer. They <strong>hallucinate</strong>. This is one of the bigger challenges in making LLMs reliable.</p><h3>Relevance: Stay on Topic</h3><p>An LLM needs to understand the <strong>context</strong> of a conversation and keep its responses on track. For example, if you&#39;re talking about kings, queens, bishops, and rooks, the LLM needs to know you&#39;re discussing chess, not royal families or birdwatching. Being able to keep up with the flow of the conversation is crucial.</p><h3>Clarity: Be Clear</h3><p>Clarity is all about making sure the LLM&#39;s responses are easy to understand. Long, complicated, or vague answers can make conversations hard to follow.</p><p>Keeping these rules in mind, there are still a few hurdles in making LLMs work perfectly:</p><ul><li><strong>Time</strong>: Training LLMs takes a long time and involves <strong>massive</strong> amounts of data. Even making small updates can take months, which doesn&#39;t work well for fast-paced software development.</li><li><strong>Cost</strong>: Running LLMs is expensive. They need powerful computers to function, and using them <strong>regularly</strong> can rack up significant costs.</li><li><strong>Uncertainty</strong>: Even with everything in place-good prompts, access to more data, and <strong>cost management</strong> there&#39;s still no guarantee that the conversation will always go smoothly. Things like tone, style, and emotional nuance are still tough for LLMs to get exactly right.</li></ul><h3>To conclude about large language models</h3><p>In summary, LLMs are set to <strong>transform</strong> how we interact with technology, moving beyond traditional software to become creative collaborators in problem-solving. This shift offers technical leaders an opportunity to <strong>scale</strong> human-like decision-making and expertise across industries.</p><p>However, challenges remain. Ensuring LLMs are reliable, adaptable, and transparent will be essential. As we integrate these systems, the focus should be on enhancing-not replacing-human creativity.</p><h3>Want to learn more about AI in software engineering?</h3><p><em>Originally published at </em><a href="https://madewithlove.com/blog/revolutionising-diagnosis-how-large-language-models-can-drive-a-10x-change/"><em>https://madewithlove.com</em></a><em> on November 18, 2024.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b3482897aaab" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/revolutionising-diagnosis-how-large-language-models-can-drive-a-10x-change-b3482897aaab">Revolutionising diagnosis: How large language models can drive a 10x change</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Integrating AI into your engineering team]]></title>
            <link>https://medium.com/we-are-madewithlove/integrating-ai-into-your-engineering-team-75d37f492153?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/75d37f492153</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[scaleup]]></category>
            <dc:creator><![CDATA[madewithlove]]></dc:creator>
            <pubDate>Tue, 10 Jun 2025 09:15:39 GMT</pubDate>
            <atom:updated>2025-06-03T16:20:19.458Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*JY4sYXM7YEHg7k5g.png" /></figure><h3>You need to start thinking about AI</h3><p><strong>Artificial intelligence</strong> is becoming more significant in our everyday lives. A study showed developers can complete coding-related work <strong>twice as fast</strong> when assisted with generative AI. For Copilot alone, 70% of users found themselves more <a href="https://www.microsoft.com/en-us/worklab/work-trend-index/copilots-earliest-users-teach-us-about-generative-ai-at-work?ref=madewithlove.com">productive</a>. If you multiply this across your entire engineering team, it’s a number you cannot ignore.</p><p>With the potential to handle tedious tasks efficiently, free up time and make your team more <strong>productive</strong>, AI tools, as they have become more mature, now span a wide range of applications. From assisting with code creation to enhancing customer support, ignoring these changes means risking decreased efficiency by undertaking manual tasks needlessly. In this article, I will explore some benefits of <strong>integrating</strong> this into your team and how to go about it. First, let’s look at the two different ways of working.</p><h3>Comparing AI used with non-AI approach</h3><p>We have two developers, Bob and Annie. Bob works for a travel-related company, but the company has banned using any AI tooling for now. Annie, an ex-colleague of Bob now works for a company that is building a transportation app, her company encourages the team to experiment with AI. Let’s explore how the two workflows differ.</p><h3>Version control</h3><p>After three draining hours tracking down a bug, Bob has just completed some work on a new feature. He is relieved to have completed the work, but it took him two days longer than estimated. He commits it to GitHub via the command line. The last pull request he created was called <em>another styling fix</em>; he received some criticism for this being too vague. Taking that on board, he creates his pull request. starts a CI process, but Bob knows this will take a while. He will also have to wait for another developer to review the work, so he decides to go for lunch. Before figuring out what to eat, his mind drifts to doubting his implementation. His only feedback so far was from his own manual testing, so he goes for coffee instead.</p><p>Annie is also wrapping up some work on a feature. just helped her catch a bug in the last function she wrote; it even suggested linting changes helping her code become more readable. Assisted by Copilot, she has worked faster while feeling like she has increased the quality of the code. The extra time available has enabled her to take some courses in Python, something she has really enjoyed. She commits the code to GitHub with the help of pre-commit hooks. When she worked with Bob, she once created a commit message, which was <em>a snazzy update</em>. Remembering this makes her shudder, this would probably not have happened with Copilot. It now helps her generate commit messages in <a href="https://devblogs.microsoft.com/visualstudio/write-your-git-commits-with-github-copilot/?ref=madewithlove.com">Visual Studio</a>. Like Bob, she creates a pull request. Her team uses , which helps detect any code smells. This is configured through GitHub actions. She makes some immediate changes, relieved she didn’t get a colleague to check the code before the fixes had gone in. She then requests a review and realises she’s starving, so she goes to lunch.</p><h3>Merging to staging</h3><p>When Bob returns from lunch, he sees another developer approved his pull request. He finds the process of merging to staging stressful, so before merging, he gets another coffee. handles the deployment. No user interactions are simulated, but the team uses for end-to-end testing. If something breaks during the process, any change will have to occur manually. Bob checks his emails while keeping an eye on the build logs.</p><p>Annie returns, having eaten Pizza in the park, and she sees that her colleague has approved her code. She is enthusiastic about merging her work to staging. This is done through GitHub actions and . Once the code is on staging, and perform testing. They handle covering many edge cases and simulate user behaviour. If issues were to occur, the code could be automatically rolled back. While this is going on, she is able to explore the next ticket she could pick up, using the time to plan.</p><h3>Merge to production</h3><p>Once everything on staging is good, Annie will deploy her work to production. The team uses alongside for monitoring. If there are any issues, a rollback will happen automatically. Once the deployment has finished, she heads home feeling satisfied.</p><p>Bob is ready to deploy to production. The team uses for monitoring, but the deployment is primarily manual. He almost deployed the wrong branch, so he took a breather and got his fifth coffee of the day before making the deployment. If anything were to go wrong, a rollback would have to occur manually. He knows he has been in the office for a long night. Knowing he won’t get paid overtime for his efforts, Bob wonders if he should be working elsewhere.</p><p>We see two approaches undertaken by Bob and Annie. The approach taken by Bob has a much higher dependency on human input and requires far more manual work. Consequently, it is more prone to errors and has a higher cognitive burden for one person.</p><h3>Areas that integrating AI can help your team</h3><p>If we explore the main areas that AI can assist, highlighted by the example above, what we really see are:</p><p><strong>Code generation</strong> — This includes the generation of tests. With the help of boilerplate code using tools such as GitHub CoPilot and Tabnine developers can <strong>work faster</strong>. For you can create better test coverage with more edge cases covered. In recently performed 96% of developers found themselves to be faster in repetitive tasks. 88% perceived themselves as more productive.</p><p><strong>Code Review: </strong>Tooling can help analyze the code for code smells and bugs, inconsistencies and security issues. This will help the human reviewers ensure the code is thoroughly checked. By complementing the <strong>human element,</strong> more time can be freed up for the intricate details of the review process. Copilot can assist with the creation of the pull request, DeepCode for analysis and suggestions and SonarQube for code smells.</p><p><strong>Debugging and error reduction</strong> — By utilising machine learning algorithms AI can assist with detecting errors and debugging. Tools such as SonarQube and Code Climate will scan for potential issues.</p><p><strong>Prototyping: </strong>Either through assisting with creating the prototype itself or by analysing data to create a better<strong> user experience</strong>,AI can help. Tools such as Figma and Uizard can create wireframes and mockups. The latter will even turn images into a design. Both of these can boost UX and innovation.</p><p><em>Reading tip</em>: <a href="https://madewithlove.com/blog/spiraling-out-the-art-of-ai-optical-illusions/">Spiraling Out: the art of AI optical illusions</a></p><p><strong>Documentation: </strong>By assisting with automating, generating, updating and correlating documentation, AI can reduce the amount of manual human effort required. Another benefit is <strong>consistency</strong> around all your documents. Additional tools are also available to help with code comments and commit messages. Grammarly can help with the grammar, whilst tools such as Mintlify Writer and Document 360 can assist with generation.</p><p><strong>Testing and detecting bugs: </strong>By improving test coverage through running tests and predicting bugs AI can build productivity and <strong>quality</strong> here. Humans may miss considering all the different edge cases, which is another place AI can step in. Aside from Copilot to help with the authoring of the tests, other tools to explore include Saucelabs, a platform to help automated testing across browsers, Testim which uses AI to help create and execute tests and Mabl, which uses machine learning to create and update tests.</p><p><strong>Communication:</strong> Through note-taking, meeting summaries and condensing long discussions, AI can streamline communication. Algorithms can work to prevent common misunderstandings, and chatbots can act as virtual assistants to answer common questions. Otter.ai, Grammarly, and Slack are just some tools you can explore.</p><blockquote><em>It may seem that you are losing human touch, but really we are freeing up human time</em></blockquote><h3>How to integrate AI into your current processes</h3><p>When you look at integrating AI into your processes, it may seem that you are losing human touch, but really we are freeing up human time. By doing less repetitive tasks, humans can boost their energy for creativity. For a company looking to scale, the ability to optimise resources and <strong>automate</strong> tasks also allows it to grow rapidly. On the other hand, it also helps if you need to descale.</p><p>There are a variety of approaches you can take to integrate AI into your business. Ultimately, these will require a <strong>strategy</strong> to be led from the top and an overall shift in the <strong>mindset</strong> of an entire company. It is a good idea to start small. Look at an area where you have a lot of manual work or where you think there is potential to cut costs. This can open the door to where AI can most assist you.</p><p>With the ability to enhance the complete <a href="https://madewithlove.com/blog/developer-experience-and-why-a-business-should-prioritize-it/"><strong>developer experience</strong></a>, it is best to integrate tooling across the entire team. If you have one developer who uses AI and saves time but then depends on another team member who does not, gains will be lost. Equally, when <a href="https://madewithlove.com/blog/tag/collaboration/"><strong>collaborating</strong></a>, if colleagues are not familiar with each other’s tooling, the process can become slower.</p><p>Through the use of metrics and insights, you can make <strong>better decisions</strong>. Key performance indicators will assist with choosing which metrics to monitor. Examples may include code quality, development speed and error rates. Engaging in conversation with your team about the impact of the tooling and performing surveys will provide you with more information. You do not want to be paying for something that you cannot realise the benefit of, or worse, that is slowing down your team</p><h3>Can your startup benefit from AI?</h3><p>For early-stage companies, integrating AI may be dismissed because it requires more resources. This does not need to be the case. With a lack of legacy code and overhead, early-stage companies can often be the best candidates to implement this cultural shift. There are benefits to be gained from using tooling to boost the smaller teams you may have. Seek advice from those who have already taken the journey, external consultancies can help here.</p><p>Define the strategy for <strong>integrating</strong> into the business, align this with your goals, and show the team you are excited by the change. You may need to invest in upskilling your team. Creating a shared literacy around the tooling and an awareness of ethics will help with the cultural shift. You will need to do more than establish the direction. Not everyone is as excited about AI; many still fear it will take their jobs away, so be empathetic.</p><p>Don’t feel overwhelmed by making sweeping changes all at once; AI offers an opportunity to be implemented in a phased manner. You are taking part in a change that will affect tomorrow, <strong>staying ahead</strong> in today’s markets and opening doors to unlock new opportunities.</p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/integrating-ai-into-your-engineering-team/"><em>https://madewithlove.com</em></a><em> on September 26, 2024.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=75d37f492153" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/integrating-ai-into-your-engineering-team-75d37f492153">Integrating AI into your engineering team</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Due diligence for AI startups: what actually matters]]></title>
            <link>https://medium.com/we-are-madewithlove/due-diligence-for-ai-startups-what-actually-matters-a715f1040f95?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/a715f1040f95</guid>
            <category><![CDATA[startupş]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[cto]]></category>
            <category><![CDATA[scaleup]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[madewithlove]]></dc:creator>
            <pubDate>Tue, 10 Jun 2025 09:14:57 GMT</pubDate>
            <atom:updated>2025-06-10T09:10:45.183Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*a9q5hhseaJiTKAgf.jpg" /></figure><p>Let’s be clear upfront: we don’t just audit <strong>AI</strong>. We audit entire companies.</p><p>Engineering. Product. Process. Team structure. The actual shape of the codebase. All the usual suspects. You can read more about that <a href="https://madewithlove.com/services/audits/?ref=madewithlove.com">on our audit services page</a>. But lately, something’s <strong>shifted</strong>.</p><p>More and more companies show up with a pitch deck that says <strong>“AI-powered”</strong> in size 72 font. And when we look under the <strong>hood</strong>, we find classification models, pipelines, , vector stores, agents, and a whole bunch of magic wires no one can really explain. So we’ve started digging deeper.</p><p>This post isn’t for AI researchers or prompt whisperers. It’s for founders, product leaders, and investors who want to understand what makes an AI system <strong>production-ready</strong> or dangerously <strong>duct-taped</strong>.</p><h3>Where the AI lives</h3><p>First question: What kind of AI are we dealing with? Classical machine learning? Prompting an LLM? Retrieval-augmented generation (RAG)? Agentic tool use? Each option brings a different <strong>flavour</strong> of complexity, and most of them will bite you somewhere eventually.</p><p>Then there’s the <strong>data foundation</strong>. Where does it come from? Is it cleaned? Is it labelled? Is anyone watching what goes in, or does the model just eat whatever it’s fed?</p><p>I’ve seen jobs that try to process thousands of documents through a single-threaded script, keeping a single HTTP connection open for 24 hours straight with no retries, no logging, no fallbacks. If it <strong>fails</strong> halfway, nobody knows which parts ran or what broke. If your product has pipelines or jobs, someone needs to own them. And if your system ever <strong>scales up</strong>, we want to know what <a href="https://madewithlove.com/blog/infrastructure-is-a-business-decision/">infrastructure</a> handles that load and how confident the team is that it won’t fall over.</p><p>One last tip: beware <strong>vendor</strong> lock-in. If every AI call is wired directly to , switching to another provider later becomes a full . It’s not exciting work, but having an abstraction layer is what makes that kind of pivot even possible. And remember, you never really know where a vendor will go next in <strong>pricing</strong>, policy, or product direction.</p><h3>Who actually owns the AI part?</h3><p>This is where it often gets messy. Is there a data scientist? Great. Where does their work live? In a <a href="https://jupyter.org/?ref=madewithlove.com">Jupyter notebook</a>? In production? Somewhere in between?</p><p>I often see notebooks <strong>floating</strong> in separate repos, not versioned, not reviewed, and definitely not built to run in <strong>production</strong>. What starts as experimentation becomes part of the product by accident. Code gets copied over. Scripts get triggered <strong>manually</strong>. No one really knows who owns what. And once things break, the team scrambles to debug a notebook that was never meant to run outside of someone’s laptop.</p><p>If developers write prompts, are those prompts <strong>reviewed</strong>? Versioned? Tested? If agents make decisions, is anyone watching what they actually do?</p><p>And then there’s the side. If you’re building an AI product with zero AI experience <strong>in-house,</strong> we’ll ask who’s going to maintain all this because it won’t <strong>maintain</strong> itself.</p><h3>Can it evolve without breaking?</h3><p>Prompts change. Embedding models get replaced. Models get retrained. We look for versioning, rollback paths, and migration plans. If someone <strong>updates</strong> a prompt and suddenly accuracy drops 20%, does anyone notice? And can they <strong>revert</strong> it?</p><p>At one company, everything looked solid: 70% test coverage and well-documented <strong>pipelines</strong>. But none of the new AI features were <strong>tested</strong>. The AI code lived in a <strong>separate</strong> setup and didn’t fit their usual testing strategy. A small prompt change made responses sound friendlier but also less precise. The tone improved, but key details disappeared. The team only noticed when <strong>users</strong> started flagging vague or incomplete answers.</p><p>We don’t expect <strong>perfect</strong> processes. We expect signs that someone’s thinking ahead.</p><h3>Will you know when it breaks?</h3><p>Traditional systems break <strong>loudly</strong>. AI systems degrade <strong>quietly</strong>. <strong>Hallucinations</strong> increase. Token usage spikes. Output quality dips. Costs sneak up.</p><p>You can’t write a unit test that says “response must be exactly X”, but you can still check for <strong>quality</strong>. We look for validation datasets, prompt evaluations, latency checks, and human feedback loops. We ask what’s being monitored. Are there ? Feedback loops? Billing alerts? If the first time you notice a problem is when customers <strong>complain</strong>, it’s already too late.</p><p>And don’t forget: if your ingestion pipeline is letting garbage through, you don’t need to be an AI model to predict what comes out the other side. It’s going to be <strong>garbage</strong>. You can polish your prompts until they shine, but if you’re feeding the model 💩, the results are going to smell the same.</p><h3>Are you leaking anything?</h3><p>Sensitive data can appear in prompts, be inadvertently stored in embeddings, or end up recorded in logs. We ask whether anything private is being indexed, stored, or shipped off to a third party. We also ask whether your provider choices match your customers’ compliance needs. <strong>GDPR</strong> isn’t optional and relying on a big brand name alone is just the due diligence you didn’t do.</p><h3>AI isn’t magic. But it can break in magical ways.</h3><p>We don’t audit to play gotcha, I’m also not Sherlock Holmes. We audit to see whether the team is in <strong>control</strong>, or just putting twenty half-blind oracles in a room and hoping they agree on something.</p><p>If you’re building something <strong>AI-heavy</strong>, make sure someone’s asking the unglamorous <strong>questions</strong>. And if you want help doing that, well… you know where to find us. Because if the AI <strong>fails</strong> silently, your product fails with it and by the time you notice, your customers already have.</p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/due-diligence-for-ai-startups-what-actually-matters/"><em>https://madewithlove.com</em></a><em> on June 4, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a715f1040f95" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/due-diligence-for-ai-startups-what-actually-matters-a715f1040f95">Due diligence for AI startups: what actually matters</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Fractional CTO explained]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/we-are-madewithlove/the-fractional-cto-explained-68166dfb4686?source=rss----39bf59ff8129---4"><img src="https://cdn-images-1.medium.com/max/2600/0*cok_5LCk4rzWkaaI" width="3626"></a></p><p class="medium-feed-snippet">This article was originally published on January 31, 2023 on the madewithlove blog.</p><p class="medium-feed-link"><a href="https://medium.com/we-are-madewithlove/the-fractional-cto-explained-68166dfb4686?source=rss----39bf59ff8129---4">Continue reading on CTO Insights »</a></p></div>]]></description>
            <link>https://medium.com/we-are-madewithlove/the-fractional-cto-explained-68166dfb4686?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/68166dfb4686</guid>
            <category><![CDATA[cto-as-a-service]]></category>
            <category><![CDATA[series-a]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[saas]]></category>
            <dc:creator><![CDATA[Steve 'vudu' Tauber]]></dc:creator>
            <pubDate>Wed, 04 Jun 2025 07:17:31 GMT</pubDate>
            <atom:updated>2025-06-04T07:17:31.127Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Finding your North Star Metric]]></title>
            <link>https://medium.com/we-are-madewithlove/finding-your-north-star-metric-9280754b6e79?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/9280754b6e79</guid>
            <category><![CDATA[cto]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[scaleup]]></category>
            <dc:creator><![CDATA[Emmaw]]></dc:creator>
            <pubDate>Wed, 04 Jun 2025 07:14:44 GMT</pubDate>
            <atom:updated>2025-06-04T07:14:44.789Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zUdNoLKUfyE1Iehk.jpg" /></figure><p>Precise Maths, an educational mobile app, recently completed a Series A funding round. They are now looking to <a href="https://madewithlove.com/blog/scaling-the-a-team/"><strong>scale</strong></a> and expand into the Spanish market. The <strong>development</strong> team is working on a feature enabling users to book lessons with tutors. Meanwhile, the <strong>customer success</strong> team is improving documentation for a feature released a few months ago. This allows users to share their scores on social media. The weekly leadership sync has just ended. Tim, the CTO, is distracted. He can’t shake off the feeling that something feels off. At such an exciting time for the company, it’s almost like the teams are not working in harmony. Are they at risk of <strong>losing focus</strong>?</p><p>Although both teams contribute to the product, their goals are not fully . They are working towards their own set of priorities in <strong>silos</strong>, and they run the risk of moving in totally different directions. When multiple metrics are focused on, this can be further compounded, Tim knows from experience this could be catastrophic. The teams need <strong>aligning</strong> and <strong>refocusing</strong>. The company needs a <strong><em>North Star Metric</em></strong> to help them stay on track.</p><h3>What is a North Star Metric?</h3><p>The North Star Metric (NSM) is a single <strong>measure</strong> of a product’s <strong>value</strong> to its customers. It can act as an anchor aligning teams around a shared and common goal. The metric should be <strong>simple</strong> and focus on what the most loyal customers find the most <strong>valuable</strong> when interacting with the product. Overall, it should help a company measure its progress and provide a refined view of how well the company is doing. As such, the risk of <strong>misalignment</strong> is reduced, decision-making is improved, and <strong>sustainable growth</strong> is promoted.</p><h3>Why is a North Star Metric important?</h3><p>A North Star Metric is important for a few reasons. Primarily, it will assist the entire company in remaining <strong>aligned</strong> around a single goal. The goal itself represents the core value that the product is attempting to deliver to its customers. When startups move quickly, teams can start moving to their own rhythms, each pursuing different interpretations of <strong>success</strong>. By knowing your North Star Metric, you can be sure that every employee is working towards the same point, promoting <strong>focus</strong> and <strong>collaboration</strong> and reducing silos.</p><h3>How do I find the North Star Metric?</h3><p>There are certain <strong>rules</strong> to put in place when you start to decide on what to use. The North Star Metric needs to be a <strong>single metric</strong>. It can change over time, but you need to keep it in place for a long time to perform an adequate evaluation. If you collate all your metrics, the North Star should be at the top-a bit like the star at the top of a Christmas tree!</p><p>So how do you decide what to use? You will need to explore what is most <strong>critical</strong> to your business. Dig deeper than revenue-focused metrics and look for the <strong>why</strong> and <strong>what</strong> behind them. What is it that customers value the most about your product? What is the stand-out factor that makes them want to invest their time and money using it? Looking at things from this stance will enable you to explore what you need to achieve in order not just to bring in the revenue but to sustain it in the long term. You will definitely want to ensure you have the NSM clearly <strong>documented</strong> and easily accessible to everyone.</p><h3>Examples of a North Star Metric</h3><p>One approach you could take when looking to find your North Star Metric is to explore your top five or ten metrics. Slowly start to reduce these down to what you consider to be the most important two or three. At this point, you should begin to see some <strong>commonality</strong>, and you will find the most important one. Here are some examples:</p><ul><li>Holiday Accommodation: The number of nights a customer books</li><li>Transport application: How many journeys a user undertakes</li><li>Online beauty store: How many purchases a user makes every month</li><li>Fitness application: How many users complete a workout every week</li></ul><h3>Precise Maths: discovering their North Star</h3><p>For Precise Maths, if the discussed with the development team what they felt would be the most <strong>important</strong> metric they may state it needs to be how long the user is <strong>engaged</strong> with the application. The customer success team may wish to focus on the number of positive ratings the application has received. While these both address surface-level problems, they do not truly represent what keeps the customers returning to use the application or the heart of the product itself. As such,for Precise Maths, an example of the North Star Metric could be the number of lessons completed by the application users every month. If the company’s main goal is to help users improve their maths the number of lessons completed should align with how much a user’s maths skills are growing.</p><p>Why choose the number of lessons completed rather than how many lessons a user books? It is a much more <strong>valuable</strong> metric because it allows us to see a user’s engagement with the product over time. This will encourage both teams to work together to maximise user engagement and <strong>retention</strong>.</p><p>With the teams now focused on the same goal, they have a new filter for all the new work that arises. Meetings will be far more valuable, and teams will be much more motivated. The development team can focus on <strong>prioritising</strong> features that make it easier for users to take and complete lessons. The customer success team can prioritise creating <a href="https://madewithlove.com/blog/five-documents-every-startup-should-have/">documentation</a> and supporting users to take the lessons frequently. They can also explore how to help users understand the value of doing so. When the company starts to expand in Spain, growth and user satisfaction can be quickly <strong>measured</strong>.</p><h3>Conclusion</h3><p>As you start to break down your growth into how you deliver value to customers, it becomes easier for your team to see how they are or could contribute. This makes it much simpler to motivate them. Without a metric in place, things can become ambiguous and hazy. What makes your product <strong>unique</strong>, and what value do you bring to your customers? What actions do users take to gain the full <strong>value</strong> of your product?</p><p>For Precision Maths, defining the North Star Metric could be a turning point in their <strong>growth</strong> strategy. By quieting the noise in the hectic startup or scaleup environments, they now have a metric that can be worked on by all the teams. So, what is your North Star? Discovering it can help you steer through the complexities of growth and scaling and ensure all your teams are working towards <strong>success</strong>. If you want help to find your North Star Metric, <a href="https://smoothsailing.be/?ref=madewithlove.com">Smooth Sailing</a> helps start-tups with design, product management and research.</p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/finding-your-north-star-metric/"><em>https://madewithlove.com</em></a><em> on January 28, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9280754b6e79" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/finding-your-north-star-metric-9280754b6e79">Finding your North Star Metric</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building a customer support machine: a pragmatic approach for startups]]></title>
            <link>https://medium.com/we-are-madewithlove/building-a-customer-support-machine-a-pragmatic-approach-for-startups-4c77cff11887?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/4c77cff11887</guid>
            <category><![CDATA[cto]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[customer-success]]></category>
            <category><![CDATA[cto-as-a-service]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[madewithlove]]></dc:creator>
            <pubDate>Wed, 28 May 2025 12:30:17 GMT</pubDate>
            <atom:updated>2025-05-28T12:27:25.618Z</atom:updated>
            <content:encoded><![CDATA[<h4>By <a href="https://madewithlove.com/team/mike-veerman/">Mike Veerman</a></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*MBQGTpfnCNCnyGZZ.jpg" /></figure><p>Engineering teams don’t just build software- they keep the product running.</p><p>But that wasn’t always the case. DevOps only started to take shape around the late 2000s. The expression “ <em>You build it, you run it</em>” was coined in 2016! What is a pretty straightforward no-brainer today, was revolutionary at the time.</p><p>Software developers used to build their product and hand it over to an operations team. Those were other engineers who had little knowledge of the product’s inner workings but were tasked with keeping it up-and-running. They were in charge of monitoring, maintenance and support.</p><p>The big idea was that <strong>splitting up engineering capacity between project work and operations</strong> guaranteed enough bandwidth for both. As long as the product specifications were complete, these “Ops teams” could, in theory, field most support questions.</p><p>But that’s not what reality taught us. Customers would always find questions with undocumented answers. They would ask about an edge case no engineer even thought about! How on earth could the Ops team answer that? They had no choice but to loop in the original developers…</p><p>As software development became more agile, specification and documentation levels went down, and the number of questions that reached the engineers skyrocketed. Customers would report bugs but also request data exports. Customer support questions to disable a certain user account would end up on the developers’ board. The days of just building features were over. That leads to planning issues as soon as a product finds a bit of success.</p><p><strong>Answering customer questions takes time away from feature development</strong>. While not every customer request is an emergency, we can’t just throw them on the backlog and hope to tackle them a few weeks from now. And by definition, it’s impossible to know how many customer support requests you’ll get in a day. Customer support questions become a source of unpredictable distraction for the team.</p><p>So, after an early honeymoon phase, startups quickly began considering professionalising their support organisation to prevent chaos in the engineering team.</p><p>But where to start?</p><p>Large companies have call centres and a legion of customer support agents at their disposal. That’s not something a startup can afford.</p><p><strong>We need a way of handling customer requests that scales with the company</strong>. Here’s how to evolve such a system.</p><h3>A three-tiered system</h3><p>The traditional approach is a system with three levels of support.</p><ul><li><strong>1st line support.</strong> This is the <strong>contact surface with customers</strong>. It’s where they will look for answers, report issues and ask questions. In large organisations, this is typically a call-center.</li><li><strong>2nd line support</strong>. When 1st line can’t help, the question reaches the second line. This consists of <strong>non-technical product specialists</strong>. They can answer most functional questions. In the scenario of the large call-center, these are the specialised agents you get forwarded to after the initial agent can’t help.</li><li><strong>3rd line support</strong>. Questions that can’t be answered by 2nd line end up with the last line- <strong>the engineers who build the product</strong>. In the typical enterprise example, this is where you will receive an invitation to sit together with some of their developers to debug your issue.</li></ul><p>This classic setup becomes an inverted pyramid. We want as many of the questions answered in line 1. We want to minimise the number of questions that reach tier 3.</p><p>This three-tiered approach works for startups and large enterprises alike, but the implementation will be different. Startups don’t have a call-center, but they still need to have tier 1.</p><h3>Automating tier 1</h3><p>The first step is to fully automate tier 1. Our first line of support should not be a call centre but a knowledge base with relevant articles to solve the customer’s problem. When experiencing an issue, they should be able to find the answer for 90% of the cases. This can be as fancy as an AI support chatbot or as simple as good-quality FAQ documentation.</p><p>When you’re just starting out, this documentation will be minimal. We’ll grow it over time.</p><h3>Automating tier 2</h3><p>For 2nd line support, we want a different kind of automation. We need a tool where users can send their requests when they can’t find the answers in the docs. This is where staples like Freshdesk or Jira Service Desk come into play. These service desk ticketing systems act as a buffer between the customers’ requests and the engineering team. We really need that buffer so the engineers can focus on strategic work. When customers can add Jira tickets to the current sprint, the team can no longer operate. That’s also why direct Slack access to engineers is such a bad idea. We don’t want customers to have direct access to the engineers or their Kanban board. That leads to chaos.</p><p>We want to replace telephone calls, emails and direct Slack messages. Every question becomes a ticket in a centralised database. This makes it easy to stay on top of open support issues.</p><p>As an added benefit, these tools often offer knowledge databases which can be used for tier 1.</p><p><a href="https://madewithlove.com/services/technical-leadership/">Technical leadership - madewithlove</a></p><h3>Starting the machine</h3><p>With these simple layers in place, we can create a professional, rightsized support machine.</p><p>Customers with questions are directed to the support page, where they can consult the knowledge base.</p><p>If they don’t find the right answer, the question goes to the second line and becomes a ticket in the service desk software. This is where your customer success people can pick it up. If they can answer it, it goes back to the customers. If they can’t, they push the question to the engineers.</p><h3>Building a support feedback loop</h3><p>Now, in the beginning, most questions will still end up on 2nd or 3rd line support as your documentation is lacking. But as the second line starts answering or delegating those questions, they add them to the knowledge base. Instead of sending the customer an answer to their question, <strong>they send them a link to a helpful article</strong> — even if they have to write it on the spot. This leads to ever more automated responses to relevant user questions.</p><p>Over time, the inverted pyramid starts to appear.</p><h3>Scaling the machine</h3><p>This approach allows us to scale the support organisation organically. We start with the team’s product manager answering a few customer requests per week. However, this same system can grow to support an entire Customer Success department that handles thousands of requests for multiple teams.</p><p>The goal of professionalising customer support is not to avoid having questions reaching your engineers. It’s to prevent the same question from landing on their desk twice.</p><p>The questions that do end up on third-line support will still take away time from feature development. It’s not a magic bullet. Managing that requires a fair bit of capacity planning. A lot of the early knowledge base work will add to the product manager’s workload. But by reducing the influx of questions, we can tame the chaos and free up engineering capacity for strategic work.</p><p>In those early days, you don’t need an army of call centre agents or an operations team. You can start with a simple knowledge base and a ticketing system.</p><p><a href="https://madewithlove.com/cto-as-a-service/">Read more about how we help startups with CTO as a Service</a></p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/building-a-customer-support-machine-a-pragmatic-approach-for-startups/"><em>https://madewithlove.com</em></a><em> on May 16, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4c77cff11887" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/building-a-customer-support-machine-a-pragmatic-approach-for-startups-4c77cff11887">Building a customer support machine: a pragmatic approach for startups</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Understanding and managing technical debt and legacy code: a guide for founders]]></title>
            <link>https://medium.com/we-are-madewithlove/understanding-and-managing-technical-debt-and-legacy-code-a-guide-for-founders-0dbe8139b97e?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/0dbe8139b97e</guid>
            <category><![CDATA[scaleup]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[founders]]></category>
            <dc:creator><![CDATA[Emmaw]]></dc:creator>
            <pubDate>Thu, 22 May 2025 12:08:42 GMT</pubDate>
            <atom:updated>2025-05-22T12:08:34.124Z</atom:updated>
            <content:encoded><![CDATA[<p>In the rapidly evolving software industry, achieving tasks quickly often becomes the priority. Repeatedly taking this approach to anything in life comes with a cost attached.</p><p>In software, this is referred to as <a href="https://madewithlove.com/blog/on-technical-debt-what-is-it-how-is-it-created-how-to-avoid-it-and-how-to-fix-it/">technical debt</a>. <strong>Understanding technical debt and legacy code is crucial for founders to maintain sustainable growth and software quality</strong>.</p><h3>Thinking about debt</h3><p>Technical debt is a term introduced by Warren Cunningham, a programmer and author in the 1990s. Cunningham <strong>compared technical debt to financial debt</strong>. In the financial domain, borrowing money can provide the necessary capital for investment in growth; however, it comes with associated costs such as interest payments. If interest payments are neglected, debt accumulates over time, which is counterproductive to growth.</p><h3>What is technical debt?</h3><p>Technical debt arises when businesses prioritize speed over quality during software development. This often involves <strong>rapidly releasing software to outpace competitors</strong>, resulting in <strong>shortcuts</strong> that accumulate as technical debt.</p><p>Similar to financial debt, technical debt can <strong>grow quickly if not managed properly</strong>. Common causes include unforeseen technical issues, poor-quality implementations, and deferred maintenance.</p><p>The key to mitigating technical debt lies in its effective management. Just like in financial scenarios, excessive technical debt can hinder growth and productivity, making it essential for businesses to address it promptly and strategically.</p><h3>What is legacy code?</h3><p>In software, we can break things down further: We have <strong>technical debt</strong> and <strong>legacy code</strong>. Both of these can lead to outages, performance, and security problems, bugs, and lowered productivity. A company’s failure to invest in maintenance and <a href="https://madewithlove.com/blog/how-to-do-a-software-rewrite/">refactoring work</a> can accelerate both. However, it is important to be aware that technical debt and legacy code differ due to their origins.</p><p>Whilst technical debt is often the fallout of a rapid release, according to the <strong>Cambridge Dictionary The term legacy</strong> describes <em>“Something that is part of your history, or that remains from an earlier time”. In software, legacy code is used to describe code that may be delivering value to a business but is hard to maintain due to becoming outdated. Often, it is poorly documented, or the documents are out of date. This can escalate the struggle which may occur to understand it. It is not unusual to find development teams fearful of changing legacy code. Often, the code features components or modules that highly depend on each other. In software, this is called tightly coupled.*</em></p><h3>The problems with tech debt and legacy code</h3><p>The overall problem with <a href="https://madewithlove.com/blog/on-technical-debt-what-is-it-how-is-it-created-how-to-avoid-it-and-how-to-fix-it/">tech debt and legacy code</a> is <strong>reduced productivity</strong>. In the long term, <strong>the code is less performant</strong>, the <strong>maintenance issues</strong> can become draining, and the <strong>process of releasing new features</strong> can become more challenging.</p><p>The primary goal of a software developer is to create good code that is <strong>readable</strong>, <strong>testable</strong>, and <strong>extensible</strong>. Code created in this manner is also inherently maintainable. Whilst problems can start as minor ones, they can rapidly mount up, and with them, so does the cost of rectifying the issue, placing the revenue stream and, consequently, the business at risk.</p><p>While temporary fixes can be applied, this is akin to using a plaster for a short-term solution, and as complications grow, more resources are spent. This expenditure can be in the form of customer support time, developers fixing bugs, or simply the longer development cycle to release. Over time, it can become harder to recruit n<a href="https://madewithlove.com/blog/seniority-level-in-software-engineering-and-how-to-assess-it/">ew developers as they may not have the skillset to work with the legacy code</a> or may not wish to spend vast amounts of time performing maintenance tasks. Another factor to consider is that onboarding new members to the team can become more lengthy than necessary. Added together, this is a costly hit to a company and the overall morale of the development team can become eroded.</p><h3>Mitigation strategies</h3><p>While most companies will inevitably accumulate some amount of tech debt or legacy code, it is what they do with it that matters. Breaking the problems down and applying a defined strategy is important. If this is achieved during a company’s early stages, it becomes ingrained within the culture, leading to <a href="https://madewithlove.com/blog/tag/collaboration/">collaboration</a>, not resistance.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*EgA0C_xmE65QMY3P.png" /></figure><h3>Assessment</h3><p>To create the strategy, it is necessary to assess the problem, and there are several approaches you can take. These will likely entail all or some of a deep code review and the state of the test overage to highlight the most problematic areas within the code. A review of the project requirements, along with data collected from code analysis, code coverage, and code smell tools such as can assist here. As with anything, it is the realization that there is a problem that will lead to a resolution.</p><h3>Solving the problem</h3><p>As work moves to address the issues, a great starting point is to focus on increasing code test coverage, particularly through <a href="https://madewithlove.com/blog/fake-e2e-tests-with-react-testing-library/">end-to-end tests</a>, and improving the continuous integration and continuous delivery processes. Focusing on these processes will enable the development team to make small and safe changes.</p><p>Automating the continuous integration and <a href="https://madewithlove.com/blog/tag/continuous-delivery/">continuous delivery processes</a> is valuable for any company to invest their time in. Automation will save time, increase stability, and reduce errors. It will also grow trust and create a faster feedback loop. If you don’t already have these processes, they will enable the team to make minor changes with greater confidence. Using tests will not only decrease risk from changes implemented <a href="https://madewithlove.com/blog/refactoring-an-old-front-end-codebase-a-field-manual/">during the refactoring work</a>, but they can also assist in providing metrics for measuring the quality of the code and the ongoing improvements.</p><p>Refactoring code means improving it without changing its user-facing behavior. In real-life terms, this can be compared to taking a car in for its annual service. Refactoring can improve the software’s overall performance. If the refactoring work is too widespread too quickly, it can risk errors being introduced within the code. <a href="https://madewithlove.com/blog/on-technical-debt-what-is-it-how-is-it-created-how-to-avoid-it-and-how-to-fix-it/">To avoid this, it is important to make gradual, incremental amendments, perhaps even breaking the system down into smaller modules</a>. These should be tested thoroughly.</p><h3>Documentation</h3><p><a href="https://madewithlove.com/blog/five-documents-every-startup-should-have/">The importance of documentation</a> cannot be understated. One of the major factors as to why legacy code becomes unmanageable is the lack of communication around the code itself. Many companies struggle with retaining developers and as they join and leave the company the knowledge around the codebase can become inconsistent. As a company matures the fact that no developer was present when the system was created becomes a possibility. Consequently, documentation not just at the birth of the project but continuously updated documentation becomes vital.</p><p><a href="https://madewithlove.com/blog/tag/operational-documentation/">documentation</a> is seldom the focus point of a startup, and understandably so. However, as a company begins to scale, investing in creating and maintaining consolidated documentation becomes critical. <a href="https://madewithlove.com/blog/why-is-operational-documentation-more-important-than-you-think/">Good documentation should explain the logic and behaviors of the code</a>. It should also cover installation instructions, domain concepts, architectural diagrams, and deployment information. Within the code, brief comments can be used to explain complex parts of the code.</p><h3>Roadmaps</h3><p>The team should work towards a roadmap where issues are based on outcomes or problems. Overall, the roadmap is vital to communicate with both the engineering team and the stakeholders.</p><p>Ensuring that within <a href="https://madewithlove.com/blog/how-to-communicate-a-roadmap-to-vcs/">the roadmap</a>, time is allocated to tackle the technical debt or update the legacy code will ensure that ongoing improvements are made and that the most at-risk areas get prioritized correctly. If the problems are more critical, a company can allocate specific sprints, perhaps occurring at set intervals, focusing on refactoring and decreasing the technical debt.</p><p>Another approach is factoring in additional time when a feature is worked on to decrease the technical debt around that code. The company should find the right balance that works for them here, considering the scale of their issues.</p><p><strong>Reading tip</strong>: <a href="https://madewithlove.com/blog/five-documents-every-startup-should-have/">Five documents every startup should have</a></p><h3>Continuous learning</h3><p>Fostering a team culture where the team feels confident in making changes without fear is critical. One angle to this is promoting <strong>continuous learning</strong>. If a company views technical debt and legacy code as an opportunity, not a burden, it will far more likely overcome the issues. A growth mindset where <strong>experimentation</strong> is welcomed will ensure the code delivered is high quality and stakeholders are satisfied. <strong>Knowledge-sharing techniques</strong> are another approach that can assist with this. Programmers can use techniques such as <a href="https://madewithlove.com/blog/pair-programming-as-a-newbie-and-the-fear-of-judgment/">pair programming</a> or mobbing. These techniques, where developers work together, enable collaboration and can also be a tool to use during the onboarding of new members to the team. Additionally, <a href="https://madewithlove.com/blog/tag/code-review/"><strong>code reviews</strong></a>, a process where developers review each other’s code can be used as an opportunity to create discussion around architectural improvements. Overall, for the whole company, it can be beneficial to <strong>gain support from mentors and experts</strong>. Of course, books and courses can be purchased to assist the team in staying up to date with modern and best practices.</p><p>Codebases of every size and age will have flaws to some degree, and the code’s age can increase issues. It is important to state that age alone <a href="https://madewithlove.com/blog/how-to-do-a-software-rewrite/">should not be the reason to rewrite the entire codebase</a>. While the team must stay on top of the modern techniques, they should avoid the trap of being drawn to using the popular new framework that could be trending at that moment. While there are times when entire rewrites may be appropriate, doing so can create extra systems requiring maintenance, which can consume resources and may not be the best solution for the business.</p><h3>Prevention</h3><p>There are many clues, such as a rise in <strong>bug reports</strong>, <strong>delays</strong>, and <strong>costs</strong>, that may mean technical debt can no longer be ignored. Customers may be reporting glitches, which can be red flags for bigger issues; customer feedback, the timeframe for onboarding new developers, and outages can all provide clues to problems arising and how to manage them.</p><p>Pay attention to your development team. They are the ones interacting with the code frequently. Additionally, ensuring custom support problems are reported to leadership will provide another communication channel.</p><h3>Conclusion</h3><p>In conclusion, <strong>addressing technical debt requires a holistic approach</strong>. While there is no single solution, establishing a culture of open communication and <a href="https://madewithlove.com/blog/tag/psychological-safety/">psychological safety</a> is vital to solving these issues. Without fostering such an environment technical debt can build up without the awareness of leadership. Just like with a credit card, this can become a problem that can spiral rapidly out of control. Integrating both the business and technical strategies and ensuring resources are allocated adequately will safeguard against further accumulation and will assist with the long-term success of a company.</p><p><em>*From the </em><a href="https://dictionary.cambridge.org/dictionary/english/?ref=madewithlove.com"><strong><em>Cambridge Advanced Learner’s Dictionary &amp; Thesaurus</em></strong></a><em> © Cambridge University Press) </em><a href="https://dictionary.cambridge.org/?ref=madewithlove.com"><em>https://dictionary.cambridge.org/</em></a></p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/understanding-and-managing-technical-debt-and-legacy-code-a-guide-for-founders/"><em>https://madewithlove.com</em></a><em> on June 6, 2024.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0dbe8139b97e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/understanding-and-managing-technical-debt-and-legacy-code-a-guide-for-founders-0dbe8139b97e">Understanding and managing technical debt and legacy code: a guide for founders</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Eliminating SPOFs: the importance of documentation]]></title>
            <link>https://medium.com/we-are-madewithlove/eliminating-spofs-the-importance-of-documentation-731df139ac9a?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/731df139ac9a</guid>
            <category><![CDATA[cto]]></category>
            <category><![CDATA[sass]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[documentation]]></category>
            <category><![CDATA[scaleup]]></category>
            <dc:creator><![CDATA[Emmaw]]></dc:creator>
            <pubDate>Wed, 21 May 2025 09:59:12 GMT</pubDate>
            <atom:updated>2025-05-21T09:59:12.095Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5VKyC8bLMxSF7ZWs.png" /></figure><p>After a highly intense year at work, Sarah has gone on a much-needed two-week holiday. Although Louise, her CTO, had assured her she should take the opportunity to recharge, Sarah had told her team she would be <strong>reachable</strong> by phone in an <strong>emergency</strong>.</p><p>Sarah lands in Vietnam on Monday evening; she is jetlagged. She goes straight to her hotel, and when she goes to bed, she realises she must have left her phone in the taxi. She knows she probably can’t get it back, so she decides to sleep and will deal with it tomorrow.</p><p>Sarah works for CalUp, an early-stage <strong>startup</strong> in the Netherlands. They plan to gear up for a seed round later in the year, and stability will be important for their investors.</p><p>In Amsterdam, Louise is making a coffee in the office when her phone buzzes. She sees several emails coming through to the company’s support email. At the same time, she notices monitoring alerts coming through on Slack. She <strong>panics</strong>. The payment integration solution used within the application is failing.</p><p>She speaks with the other developers who realise an environment variable needs updating after the payment provider released a new API version. Nobody knows which of the 25 variables are safe to modify or what the new value should be. They have got this incorrect in the past, resulting in <strong>data loss</strong>.</p><p>After a tense discussion with the team, they decided to place an alert notifying users that transactions may fail. Louise emails their customers to apologise for a temporary issue with payment processing. Eight hours later, after many complaints from customers, Louise receives an email from Sarah explaining about her lost phone. She is then able to walk the team through the update required.</p><p>This is the world of <strong>SPOFs</strong> ( <strong>Single points of failure</strong>), where crucial information for your company lives in the head of one individual.</p><h3>What is a SPOF?</h3><blockquote>“A SPOF is a component within a system which, if fails, causes the entire system to fail.” (Li, 2018).</blockquote><p>A SPOF (single point of failure) can manifest in many different ways, but it is when a crucial aspect of your business is dependent on either an <strong>individual</strong>, a specific process or a system. If it is an individual, they are often the ones who have been with the company since its inception; they have been on the whole journey. Consequently, they have a complete understanding of the what and the why.</p><p>SPOFs can be an issue for <a href="https://madewithlove.com/blog/startups-dont-need-corporate-culture/"><strong>startups</strong></a> because teams are usually small, and individuals often wear many hats. One thing that exponentially worsens an SPOF and the scale of the risk attached? A lack of good <a href="https://madewithlove.com/blog/why-is-operational-documentation-more-important-than-you-think/">documentation</a>.</p><h3>Add documentation to the equation</h3><blockquote>“We just have too much on our plates right now. Once the release is out next month, we will focus on documenting everything.”</blockquote><p>That’s what the CTO told Sarah a month before she left for Vietnam. SPOFs are not isolated to a product’s technical aspects, either. Let’s explore other ways they could manifest.</p><h3>The technical aspects</h3><p>When only <strong>one</strong> developer understands a critical system or process you are in a <strong>vulnerable</strong> position. These can manifest in the form of custom scripts, undocumented APIs, or <a href="https://madewithlove.com/blog/grandfathering-in-legacy/">legacy systems</a> that nobody on the team understands. Let’s fast-forward for a moment. If Sarah creates some custom functionality and then leaves the company, how is anyone on the team going to understand it fully? If <strong>documentation</strong> exists from the start, whoever replaces Sarah will be able to understand how decisions were made.</p><h3>The operational aspects</h3><p>As with code, important product knowledge, business processes, or customer relationships can depend heavily on an individual. This creates <strong>risk</strong> if these individuals become unavailable. CalUp has Jessica as its CFO. How will the company process the end-of-month financials if she has a family bereavement? What would happen if Fred, their Head of Product, had an accident? He is the only one who has the <a href="https://madewithlove.com/blog/how-to-build-roadmaps-that-can-be-used-by-engineering-and-product-teams/"><strong>roadmap</strong></a> stored. The problem is that it is stored in his head.</p><h3>What is the impact of a SPOF?</h3><p>For CalUp, the impact was a day’s worth of lost transactions, angry customer relationships and a lot of stress.</p><p>Some of those relationships could be unrecoverable. Other impacts could include an investor pulling out or reducing the company’s valuation. It could lead to delays in releases, overtime, resignations from your team members, and emergency consulting costs. It also means handing opportunities to the <strong>competition</strong>.</p><p>When SPOFs exist, critical moments for a company simply become more stressful by default. For the team, it is no fun being <a href="https://madewithlove.com/blog/how-to-stop-depending-on-that-one-developer/">dependent on an individua</a>l, and for the individual, it is unnecessarily stressful to be so relied on.</p><h3>Mitigating the issue with a documentation culture</h3><p>The key for the business is <strong>reducing risk</strong>. If vital information is documented, the overall risk of the issues becomes lower. This means that if a key component does fail, the organisation can quickly and elegantly recover.</p><p>You may be wondering where to start and what to document. Split up the documentation into technical, operational and product.</p><p><strong>Technical documentation</strong> should include more than simple setup instructions. Architectural diagrams, a glossary of domain terminology, dependencies, and recovery procedures should be included. Failure scenarios can be written up along with the details of escalation pathways and who is responsible. API specifications are also shared when engineers collaborate, so this should also be considered. can help here. <a href="https://madewithlove.com/blog/tag/product-documentation/"><strong>Internal product documentation</strong></a> will document decisions, troubleshooting guides, and workflows. The product vision should be clearly correlated for the team to refer to. Roadmaps and relevant matters around it and then also technical debt are additional aspects to include. <a href="https://madewithlove.com/blog/why-is-operational-documentation-more-important-than-you-think/"><strong>Operational documentation</strong></a> should have enough details for an individual to successfully onboard themselves without relying too much on others’ involvement. This should include step-by-step guides for key processes, role-specific responsibilities, access to necessary tools and systems and also troubleshooting procedures and company policies.</p><p>The best time to address an SPOF is before a <strong>critical</strong> issue arise such as the one with CalUp. Documenting early means knowledge is preserved. When the <a href="https://madewithlove.com/blog/pragmatic-ways-of-keeping-documentation-up-to-date/">documentation</a> is there, everyone, current or future, can fully understand something, positively impacting knowledge sharing and continuous learning. You also reduce the number of questions the team need to ask and particularly for juniors on the team <strong>speed</strong> up the process of them becoming more independent and valuable to the company.</p><p>When Sarah tells Dave something and Dave then tells Lauren-his junior colleague, the message probably changes in the retelling. As we work to prevent any <a href="https://www.linkedin.com/pulse/how-stop-depending-one-developer-madewithlove-5q3ee/?trackingId=TFtHlipBS2eF6jVRPG5NlQ%3D%3D&amp;ref=madewithlove.com">knowledge silos</a>, we safeguard essential parts of information from being lost or miscommunicated. When knowledge held by an individual is documented it becomes an <strong>asset</strong> to the company. When the information is centralised, it is available to everyone twenty-four hours a day-reducing lost time spent searching for information. That part becomes <strong>critical</strong> in a moment of crisis.</p><p>It is unrealistic to expect perfection from documentation, especially in an early-stage company but if the mindset is one of growth and <strong>evolution</strong> as opposed to in the moment or completion, the company gradually becomes more robust. Ensure that documentation itself does not become a SPOF by undertaking regular review cycles. Dont make it an after thought.</p><h3>How did CalUp learn from their experiences?</h3><p>The company referred to the incident as ‘The phone loss incident’. Although it was a bad experience for the company, they focused on turning it into something <strong>positive</strong>. They instigated documentation Fridays. This was an initiative where the team spent Friday afternoons creating documentation. Roadmaps and the product vision were shared throughout the company, and everyone knew where to find them. The developers added videos for critical aspects of the code. Jessica, the CFO, added step-by-step guides for processes, and Sarah added ADRs (architecture design records) to ensure the context behind decisions was documented.</p><p>Although the <strong>growth</strong> of documentation occurred gradually, it quickly impacted the company. Soon, templates were ready for different requirements, and with everyone on the team contributing, it boosted <strong>collaboration</strong>. The company was sure to review the documentation regularly, too. After all, out-of-date documentation is almost as bad as no documentation. You are essentially sending someone on a chase for false knowledge. Lastly, guidelines should be created on what should be documented and by whom. As it grows, <strong>analytics</strong> are helpful to see who uses what the most or least.</p><h3>The role of leadership</h3><p>When you are ready to create a documentation culture, you need to ensure the team doesn’t view it as a lower priority than other work. Explore who the team usually depends on for answers to questions and what could break if they left. This should expose what knowledge exists only in the heads of individuals. Create <strong>policies</strong> around the documentation. If you view it as creating the foundation, it can be built upon as you go. Prevent gatekeepers and push.</p><p>When we perform our <a href="https://madewithlove.com/technical-due-diligence/?ref=madewithlove.com">due diligence audits</a>, we see many companies with poor documentation. This can decrease a <strong>valuation</strong>.</p><p>When CalUp reached their investment round the investor explained that to invest in early-stage companies, they have to evaluate the operational efficiency, the strength of the team, and their potential for scaling. They went on to say that their experience with CalUp was not like that of the companies they often see. They felt comfortable that what they were creating was being built to scale, and the documentation culture and maturity de-risked their investment.</p><h3>Conclusion</h3><p>Documentation means the team can remain <strong>aligned</strong> yet work individually, making the company more robust. If the investment round for CalUp had occurred eight months prior to ‘The phone loss incident,’ the company’s valuation could have been reduced. These SPOFs are by no means isolated to startups, in all cases, the due diligence processes explore these risks. If it was discovered that the entire codebase was handled by such a small team with no documentation available, the startup that could have sounded promising quickly became a risk.</p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/eliminating-spofs-the-importance-of-documentation/"><em>https://madewithlove.com</em></a><em> on March 20, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=731df139ac9a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/eliminating-spofs-the-importance-of-documentation-731df139ac9a">Eliminating SPOFs: the importance of documentation</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Making an impact by combining our services]]></title>
            <link>https://medium.com/we-are-madewithlove/making-an-impact-by-combining-our-services-26c847ef414a?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/26c847ef414a</guid>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[cto]]></category>
            <category><![CDATA[technical-leadership]]></category>
            <category><![CDATA[cto-as-a-service]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Yannick De Pauw]]></dc:creator>
            <pubDate>Mon, 19 May 2025 09:47:15 GMT</pubDate>
            <atom:updated>2025-05-19T09:46:54.974Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Ajv1M_hyllzvEW8j.jpg" /></figure><p>We don’t believe in isolated solutions or lone heroes who single-handedly hold the fortress against every challenge.</p><p>While our <a href="https://madewithlove.com/interim-cto/?ref=madewithlove.com">CTO ad interim</a> service and staff engineering support are each powerful in their own right, the real magic happens when we blend them. Why? Technical strategy without execution is just theory, a compass doesn’t sail the boat on its own now does it? On the other hand, strong execution without a clever strategy can lead to chaos, like a team of rowers paddling in different directions. Real progress is being made when we can combine our services.</p><h3>The perfect mix: strategy and execution</h3><p>Our CTOs ad interim bring <strong>clarity</strong> and <strong>direction</strong>. They focus on high-level technical strategy, process improvements, and leadership development. But vision alone isn’t enough. That’s where our <a href="https://madewithlove.com/blog/supercharge-your-startup-with-staff-engineers/">staff engineers</a> come in. They <strong>don’t just code when joining our client’s tech teams; they observe, lead by example, refine processes,</strong> and <strong>ensure knowledge transfer</strong> so teams can operate at their best. And they do it in a fully collaborative approach.</p><p>This combination helps startups and scale-ups solve complex technical challenges while simultaneously growing their internal teams. Our approach isn’t about long-term dependency. Quite the opposite- <strong>we aim to build self-sufficient teams that can thrive without us</strong>. That’s why we discuss our exit-strategy with our clients from day one.</p><h3>Making informed decisions</h3><p>One key element of our approach is ensuring that technical decisions are made with clarity and collaboration. Inspired by effective <strong>CTO decision-making frameworks</strong>, we are able to zoom in on specific details that need attention, but also zoom out to keep the overview, guard the strategy, and involve the right stakeholders. By creating space for asynchronous discussions and gathering team insights, we make sure that both leadership and engineering teams align on critical decisions.</p><p>Moreover, combining these two services allows for working on greater and more complex teams and products. We prioritise knowledge-sharing methods like feature passports to document reasoning behind technical choices. This ensures that future teams can understand the context behind past decisions, preventing knowledge loss when key contributors move on. This approach allows us to dive into both the broader architectural vision as well as the business impact and effort assessment of code changes, ensuring feasibility and alignment with strategic goals.</p><p>Beyond technical expertise, we bridge the gap between business and tech, ensuring both sides understand each other. We speak both business and tech fluently, providing technical proficiency, architectural oversight, compliance guidance, and mentoring. Our role is to create a level playing field in communication between all stakeholders, preventing misunderstandings that often arise when teams feel they are “speaking different languages.”</p><h3>Avoiding Key Person Risk</h3><p>One common challenge in technical teams is <strong>Key Person Risk</strong> (KPR), where critical knowledge is concentrated in a single expert. This is often seen when one developer is the go-to person for an entire infrastructure, framework, or system. While this can seem efficient in the short term, it creates bottlenecks and long-term risks.</p><p>Instead of relying on a single hero, we encourage <a href="https://madewithlove.com/blog/why-is-operational-documentation-more-important-than-you-think/">a knowledge-sharing culture</a>. One effective approach is shifting these specialists into advisory roles, where they guide others rather than doing all the work themselves. Techniques like mob programming or structured mentoring sessions ensure that expertise is distributed across the team. While this may feel slower initially, it builds long-term resilience, increases confidence among team members, and improves documentation. Over time, this reduces reliance on a single individual and fosters a more capable and autonomous engineering team.</p><h3>Why this matters for startups and scale-ups</h3><p>Growing a technical team while maintaining momentum is tough. Many startups struggle with scaling their engineering teams, defining clear career paths, and avoiding knowledge silos. Instead of plugging gaps with temporary hires or consultants who create more dependencies, we embed ourselves within your team, improve technical decision-making, and uplift the entire engineering culture.</p><p>We don’t just fix things. We empower teams to sustain and scale the improvements we implement. Whether it’s improving technical processes, mentoring engineers, or implementing robust architectural decisions, our goal is always to leave teams stronger than we found them.</p><p>Building a culture of continuous improvement makes us happy. And so does seeing a succesfull onboarding and hiring approach result in growth and scalability.</p><h3>Stepping back, not staying forever</h3><p>A key principle of our work is <strong>stepping back as soon as possible</strong>. Unlike traditional consulting firms that create long-term reliance, we structure our engagements to ensure teams become self-sufficient. By the time we step away, your technical leadership and engineering teams are fully equipped to continue growing and scaling with confidence. Our approach to technical leadership is not just about implementing solutions but also about enabling teams and individuals to grow.</p><p><a href="https://madewithlove.com/cto-coaching-and-mentoring/?ref=madewithlove.com">Through mentoring and coaching</a>, we help CTOs, engineering managers, and staff engineers develop the leadership skills needed to guide a SaaS business effectively. We embed ourselves in your teams, working hands-on to refine processes, enhance documentation, and create a sustainable development culture. By focusing on self-awareness, strategic thinking, and process improvement, we ensure that your technical leadership evolves alongside your company’s growth.</p><p><em>Originally published at </em><a href="https://madewithlove.com/blog/making-an-impact-by-combining-our-services/"><em>https://madewithlove.com</em></a><em> on May 9, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=26c847ef414a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/we-are-madewithlove/making-an-impact-by-combining-our-services-26c847ef414a">Making an impact by combining our services</a> was originally published in <a href="https://medium.com/we-are-madewithlove">CTO Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Extreme Autonomy — the heart of Free Range Management]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/we-are-madewithlove/extreme-autonomy-the-heart-of-free-range-management-aef25a76ec53?source=rss----39bf59ff8129---4"><img src="https://cdn-images-1.medium.com/max/2600/0*d0FLed1EjgbD7oBW" width="3999"></a></p><p class="medium-feed-snippet">What&#x2019;s the best way to lead knowledge workers? What is Free Range Management really about? What is extreme autonomy?</p><p class="medium-feed-link"><a href="https://medium.com/we-are-madewithlove/extreme-autonomy-the-heart-of-free-range-management-aef25a76ec53?source=rss----39bf59ff8129---4">Continue reading on CTO Insights »</a></p></div>]]></description>
            <link>https://medium.com/we-are-madewithlove/extreme-autonomy-the-heart-of-free-range-management-aef25a76ec53?source=rss----39bf59ff8129---4</link>
            <guid isPermaLink="false">https://medium.com/p/aef25a76ec53</guid>
            <category><![CDATA[free-range-management]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[autonomy]]></category>
            <category><![CDATA[extreme-autonomy]]></category>
            <dc:creator><![CDATA[Steve 'vudu' Tauber]]></dc:creator>
            <pubDate>Thu, 09 Nov 2023 09:07:10 GMT</pubDate>
            <atom:updated>2023-11-09T09:07:10.917Z</atom:updated>
        </item>
    </channel>
</rss>