<?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 Syashi Gupta on Medium]]></title>
        <description><![CDATA[Stories by Syashi Gupta on Medium]]></description>
        <link>https://medium.com/@syashigupta?source=rss-7763a4c26639------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Okx8e8d76dATQKMKz6tloA.jpeg</url>
            <title>Stories by Syashi Gupta on Medium</title>
            <link>https://medium.com/@syashigupta?source=rss-7763a4c26639------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 11:35:55 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@syashigupta/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[Experiment 01: Building a Personal AI Travel Visualizer]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@syashigupta/experiment-01-building-a-personal-ai-travel-visualizer-276e0f1c3c6a?source=rss-7763a4c26639------2"><img src="https://cdn-images-1.medium.com/max/1332/1*VBVdgqb14pnWoVIOB3V9tg.png" width="1332"></a></p><p class="medium-feed-snippet">This project started as an experiment. Not a startup idea, not a finished product. Just a way to learn by building something real.</p><p class="medium-feed-link"><a href="https://medium.com/@syashigupta/experiment-01-building-a-personal-ai-travel-visualizer-276e0f1c3c6a?source=rss-7763a4c26639------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@syashigupta/experiment-01-building-a-personal-ai-travel-visualizer-276e0f1c3c6a?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/276e0f1c3c6a</guid>
            <category><![CDATA[experimentsincode]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Mon, 20 Apr 2026 14:01:03 GMT</pubDate>
            <atom:updated>2026-04-20T14:01:03.646Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Under the Pyramids]]></title>
            <link>https://medium.com/@syashigupta/under-the-pyramids-259ec1f651c4?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/259ec1f651c4</guid>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[be-the-difference]]></category>
            <category><![CDATA[uxpowerup]]></category>
            <category><![CDATA[designingai]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Mon, 30 Mar 2026 13:01:00 GMT</pubDate>
            <atom:updated>2026-04-02T03:10:21.892Z</atom:updated>
            <content:encoded><![CDATA[<p><em>What UX PowerUp is….</em></p><p>I never thought I would be transported back to my undergraduate days through a two‑week, in‑person experience, preceded by three weeks of virtual sessions. Calling UX PowerUp a “workshop” doesn’t come close to capturing what it actually was.</p><p>What gets left out when you just call it “2 weeks of workshop” is your journey through a myriad of emotions, knowing yourself a little better, finding your purpose, exchanging germs[lol], and seeing and accepting others.</p><p>Every year, I wrote a letter at the end of each year to read at the next year’s end. This year, I read last year’s letter but did not write one.</p><p>I felt tired even thinking about what to write, as some quiet joy had slipped out of my life. I had chosen design because it brought play into my life, but somewhere between 7 to 6 and beyond, something had shifted.</p><p>I did not know what to expect from this workshop. I was reluctant. A lot was going on at work. I was exhausted, almost at the edge of burning out.</p><p>I also kept returning to a thought that maybe the universe sends you things exactly when you need them. And I now know I definitely needed this.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*grBDAVHOmKXBVsSnmHOcEA.png" /></figure><p><em>This felt like standing at the site of an ancient pyramid.<br>From the outside, it looked still, complete, almost distant.<br>But I knew there was something buried underneath layers of time, dust, and silence.<br>The excavation began.<br>Slowly. Patiently. Layer by layer.</em></p><p>What followed was not a moment of change, but a slow uncovering that began on Sunday, 8th March 2026, and kept unfolding as the days passed. Each conversation, each session, each person felt like brushing away another layer of sand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Eyq0NS-hDTu93X4B3CPAnQ.jpeg" /><figcaption>Mind map of things that I learnt [refined my drawing with Gemini]</figcaption></figure><h3><em>The Excavation</em></h3><p><em>At first, it is just surface.<br>Loose sand. Scattered fragments. Nothing that feels complete.<br>But then you find something small. Something meaningful.<br>And as you sit with it, as you turn it in your hands, more patterns begin to reveal themselves.<br>More fragments connect.<br>More glimmers appear.<br>Until one day, you step back and realize<br>this structure was always there. You just could not see it before.</em></p><p>Some things stayed with me more than others. Not as separate lessons, but as pieces of the same structure slowly coming together. Some ideas felt immediately true. Some resisted. And the resistance itself became part of the excavation, forcing me to ask what did not sit right and why. Some key takeaways from UX PowerUp were-</p><h3>1. Being Kind</h3><p>On being kind, remembering that every person is going through something.<br>Being kind feels simple, but it asks for awareness. It asks for a <strong><em>pause</em></strong>. It asks for choosing understanding over reaction.</p><blockquote>“Be kind, not nice.”</blockquote><p>Kindness is not about pleasing. It is about seeing. It is about responding with intention, even when it is uncomfortable. In a world moving as fast as this one, that choice feels heavier and more necessary than ever.</p><h3>2. Curiosity Over Loudness</h3><p>You do not have to be loud to prove your point. <strong><em>You have to be curious enough to ask the right questions.</em></strong><br>This felt unfamiliar to me. Almost against instinct. There is a part of me that wants to assert, to prove, to hold ground. But curiosity softens that. It works when you are trying to prove a point. It works when you are trying to get a buy-in. Everyone loves talking about themselves. And somewhere in that, the right question opens doors that force never could.</p><p>It was striking how different speakers, across completely different topics, kept circling back to this same idea through their own paths.</p><h3>3. Giving</h3><p>The concept of <strong>Givers, Takers, and Matchers</strong> in the workplace by Adam Grant stayed with me.</p><p>Interestingly, Givers occupy both the highest and the lowest ends of the successful at work spectrum. Matchers help move Givers forward. And Takers do more damage to teams.</p><p>Then came the idea that stayed with me the longest.<br>Giving has to hurt.<br>I struggled with this.<br>If giving hurts, then someone is getting hurt.<br>It could either be me or someone else.<br>I questioned- can giving only be for joy and not cause any pain?</p><p>And yet, when I looked around, the biggest acts of giving did not come from abundance. They came from people who had very little, and still chose to give. I heard it in a podcast and thought about the “Giving Pledge”, where billionaires commit to giving away wealth. But the pledge is not binding. They can step away. And something about that quietly made sense, maybe giving does hurt. And maybe that is why it is not sustained.</p><h3>4. Small Actions</h3><p>The smallest actions lead to the biggest impacts- “The butterfly effect”. Nipun Mehta used the term “Murmuration”.</p><p>A single movement, repeated and echoed, becomes something much larger than itself.<br> — An act of kindness.<br> — An act of learning.<br> — An act of showing up.</p><p>Each one feels small in isolation, but together they begin to form patterns, like birds moving in the sky, creating something alive and collective.</p><p><em>In the excavation, these are the small artifacts you almost overlook until you realize they are part of something much bigger.</em></p><h3>5. Learning</h3><p>Keep learning. Learning is not natural. It is practiced. It is built. The more one returns to it, the stronger it becomes.</p><p><strong>Learning Quotient</strong> defines whether you remain relevant, whether you can adapt, and whether you can continue to evolve.</p><p>Learning without action is another layer of sand. It sits there, covering what could have been uncovered. <strong><em>To learn is to act. To try. To fail. To return again.</em></strong></p><blockquote>“Ideas are cheap, but execution is everything.”</blockquote><p>And most of the time, the deepest learning comes when you are uncomfortable, when something feels difficult, when you are forced to confront something you do not yet understand.</p><p>Somehow in my brain, all of these threads started to tie together further.</p><blockquote>“You are the average of the five people you surround yourself with.”</blockquote><p>If I am the average of the five people I surround myself with, then isn’t it my responsibility to grow them as I grow? So that when I rise, they rise. And when they grow, I grow. → That becomes giving.-&gt; That becomes learning.-&gt; That becomes growth.</p><p>And maybe that’s how small actions compound into something much larger. But execution of that part is still in my hands.</p><p><em>As the excavation continued, something else surfaced.</em></p><p>Parts of me that had been buried.<br>The part that plays.<br>The part that loves creating with their hands.<br>The part that questions everything.</p><blockquote>“Make things with your hands.”</blockquote><p>There is something grounding in that. Something real. In a world increasingly shaped by abstraction and AI, it brings you back to yourself.</p><blockquote>“Thinking in first principles.”</blockquote><p><strong>Stripping things down. Removing assumptions. Rebuilding from truth.</strong></p><p>The most interesting thing about UXPowerUp is that it grounded me as a designer. It reminded me that my superpower is not an AI tool but my heart and my brain. The part of me that had been buried in the corporate routine came back to life</p><ul><li>Introducing play</li><li>Making things with my hands</li><li>Starting with curiosity</li><li>Thinking in first principles</li></ul><p>There was space again<br> — to think optimistically,<br> — to speak my mind,<br> — to not be afraid.</p><p>It felt like returning to something I had lost.</p><p>And maybe that is what this entire experience was.</p><p>Not a workshop.<br>Not just learning.</p><blockquote><strong><em>But an excavation. Of ideas. Of beliefs. Of self.</em></strong></blockquote><p>Until somewhere along the way, you stop searching for something new and realize you are uncovering something that was always yours.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/943/1*1LOBfEYhk-jDSxvNmRDr_Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/788/1*O_6MMNbIkplJz0bmQuPuFg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/812/1*1W_HTHEvp7ur5e4yuvn8lQ.png" /><figcaption>Day 1, Day 2, Day 3</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/848/1*PUI3rJ53CAwirGI32DPRmQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/853/1*Y-2Iv8l2mbf9j0tMRBucqA.png" /><figcaption>Day 4, Day 5.. some snippets of Day 6 as well</figcaption></figure><h3>Well, it&#39;s not all that UX PowerUp was…</h3><p>Alongside all of this inner excavation, there was also the work we did together. In the second week, we worked in teams with an NGO called Dignify, focused on supporting returning citizens as they transition back into society. It was about helping people find housing, stability, and a sense of dignity, so they could rebuild their lives and not return to the streets.</p><p>This was not just a design exercise. It was a reminder that design sits in real systems, affecting real lives. It asked us to listen more deeply, to think beyond solutions, and to approach the problem with empathy, responsibility, and care.</p><p>Somewhere in that process, design stopped feeling like output and more like impact.</p><p>Thankful to everyone who curated and who made this workshop successful ❤ #BeTheDifference #SAPBeyondCode #SAPDesign #UXPowerUp</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=259ec1f651c4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Redefining AI UX: Beyond the One-Click Magic]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@syashigupta/redefining-ai-ux-beyond-the-one-click-magic-99f80beb828b?source=rss-7763a4c26639------2"><img src="https://cdn-images-1.medium.com/max/2304/1*DfDsjI0BDXZIrtk2gweCtA.jpeg" width="2304"></a></p><p class="medium-feed-snippet">As I explored these diverse applications of Gen AI, a consistent theme emerged, leading to my most significant learning relating to how&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/@syashigupta/redefining-ai-ux-beyond-the-one-click-magic-99f80beb828b?source=rss-7763a4c26639------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@syashigupta/redefining-ai-ux-beyond-the-one-click-magic-99f80beb828b?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/99f80beb828b</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[ai-design]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Mon, 21 Jul 2025 13:11:01 GMT</pubDate>
            <atom:updated>2025-07-21T13:11:01.264Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Workflows & Prompt Engineering Templates for UX Designers]]></title>
            <link>https://medium.com/@syashigupta/workflows-prompt-engineering-templates-for-ux-designers-323607e3b0e1?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/323607e3b0e1</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ai-applications]]></category>
            <category><![CDATA[ux-design-process]]></category>
            <category><![CDATA[prompt-engineering]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Tue, 15 Jul 2025 13:03:02 GMT</pubDate>
            <atom:updated>2025-07-15T13:03:02.158Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gbp6g4UJ6OBS7eqKh8-BTg.jpeg" /><figcaption>Generated using Firefly</figcaption></figure><p>Generative AI, when utilized effectively, can significantly enhance designers&#39; UX work. It effectively comes down to setting up thoughtful prompts, using the right tools, and keeping control. The real power comes when your workflows make AI your sparring partner, not your replacement. So, I present ways of building a prompt for some design workflows.</p><h3>Practical AI-Powered Design Workflows</h3><h4>Idea 1: Creating User Journeys Prompt</h4><p>AI can help structure and populate user journey maps, providing a solid foundation for understanding user experiences.</p><p><strong>Workflow(</strong>Image 1<strong>):</strong></p><ol><li><strong>Core Journey Generation:</strong> Use the prompt in a general-purpose AI (ChatGPT, Gemini). I attach images, add context, and roughly write the flow.<br><strong>Example Prompt:</strong> “As a UX designer, create a detailed user journey map for a [type of product/service, e.g., ‘mobile banking app’]. Focus on a [specific persona, e.g., ‘first-time user setting up an account’]. Outline the stages of their journey: [List key stages, e.g., ‘Awareness, Onboarding, Initial Use, Problem Solving, Long-Term Engagement’]. For each stage, include:<br><strong>Actions:</strong> What the user does.<br><strong>Thoughts:</strong> What the user is thinking.<br><strong>Emotions:</strong> How the user feels (on a scale of negative to positive).<br><strong>Touchpoints:</strong> Where the user interacts with the product/service.<br><strong>Pain Points:</strong> Any frustrations or challenges.<br><strong>Opportunities:</strong> Areas for improvement or new features.”</li><li><strong>Detailing &amp; Context:</strong> Provide additional context iteratively. Read through and ask AI to revise some sections or add details to a section based on the data sources/context that you provide.</li><li><strong>Visual Mapping:</strong> Transfer the AI-generated text outline into a dedicated journey mapping tool to create the visual map.</li><li><strong>Team Collaboration:</strong> Share the AI-assisted journey map with your team for validation and further refinement, ensuring it accurately reflects user research and insights.</li></ol><p>⚠️ <strong>Watch for:</strong> Over-trusting AI’s assumptions, AI needs real user insights and a lot of context to frame the user journey to your context.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9HQSqV760FCdbSaPTnHsvw.png" /><figcaption>Image 1</figcaption></figure><h4>Idea 2: Vision Explorations &amp; Prototypes</h4><p>This is where Generative AI outshines as a design partner, helping you shape and refine complex interactive experiences, especially those including AI features. You can use these early prototypes in real user research: show screens, test mental models, gather feedback, and then re-prompt based on what you learn. These prototypes also serve as powerful design strategy tools for aligning ideas with stakeholders and developers.</p><p><strong>Workflow:</strong></p><ol><li><strong>Iterative Prompt Refinement with General AI:</strong> Engage in a conversation with your chosen AI. Your initial prompt invites the AI to guide you. Respond to its questions and read through the initial prompts. <br><strong>Example Prompt (for a general-purpose AI like ChatGPT or Gemini):</strong> “I’m designing an AI-powered interface and want to generate a detailed design prompt for a prototyping tool[mention the name of the tool]. Please guide me <strong>step-by-step</strong> by asking focused questions to help me define the interface structure, AI interactions, and visual design. The goal is to create a high-fidelity mockup or prototype that includes[enter details]. I’ll provide answers iteratively. Help me refine the prompt as we go.”</li><li><strong>Initial Visual Exploration (e.g., Figma AI):</strong> Once you have a well-defined prompt from your general AI interaction, take it to a design-specific AI tool like Figma AI or UX pilot.<br><strong>Observation:</strong> My experience shows that the more precise you are with visual descriptions and interaction definitions, the closer the AI-generated visuals get to your intent. Defining interactions, even conceptually, forces you to think through design problems more deeply.</li><li><strong>Application-Specific Prototyping (e.g., Lovable.ai, UX Pilot):</strong> Utilize a specialized prototyping tool to generate prototypes by feeding it the humongous prompt and early visual explorations.</li><li><strong>Refining prompts and prototypes</strong>: Collect feedback on the prototype, then use AI to recreate the prompt. Finally, add the output back to the prototyping tool.</li></ol><p>Tip: I used the Lovable Prompt generator GPT on Chat GPT for generating prompts for lovable.ai. Try using these prompt generators specific to the prototyping tool that you are using. <a href="https://chatgpt.com/g/g-6851c4c164ec8191931b0343ed4d2385-lovable-promptn">Link</a></p><p>⚠️ <strong>Watch for:</strong> Fidelity trap! Don’t confuse a slick AI prototype with real validation. Another one- even for the same prompt, results can vary vastly in different prototyping tools( Image 3).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f9FX-YTlPrCNi1Xn1DHm9w.png" /><figcaption>Image 2: Prototyping</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7DeKOWSTJBnO2YyBqO8itw.png" /><figcaption>Image 3: Different tools’ outputs for the same prompt. Depending on the model and data fed, a word can have different meanings in different tools</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NaJ0WGQpSaVIozK1HNEz8A.png" /><figcaption>Image 4: Made using Gemini and Figma</figcaption></figure><h4>Idea 3: The Concept of Promptframes</h4><p>Building on the idea of structured prompts and reusable components, “Promptframes” offer a powerful new design deliverable.</p><p><strong>Concept:</strong> A Promptframe is a design deliverable that documents content goals and requirements for generative-AI prompts based on a wireframe’s layout and functionality. I would take this a step further by enabling design teams to design their Promptframes, which developers can then integrate, much like design system components.</p><p><strong>How to use:</strong></p><ul><li><strong>Team Use:</strong> Designers build Promptframes for repeatable features (e.g., smart search, dynamic carousels, brand images).</li><li><strong>Plug &amp; Play:</strong> Developers and design teams can both use these prompt frames</li><li><strong>Vision Prototype:</strong> Test Promptframes multiple times, and alongside Promptframes, define the tool to be used to generate these. As mentioned in Image 3, changing the tool changes the outputs drastically.</li></ul><p><strong>Read more:</strong> NN/g:<a href="https://www.nngroup.com/articles/promptframes/"> https://www.nngroup.com/articles/promptframes/</a></p><h4>Idea 4: Creating Moodboards Workflow</h4><p>AI helps you explore styles fast, but your eye makes it meaningful.</p><p><strong>Workflow:</strong></p><ol><li><strong>Start with a spark</strong><br>Begin with a manually created moodboard, or generate key images with detailed prompts to set the vibe, or find an ideal reference. Use these images only as inspiration.<br><strong>Example Prompt (for a general-purpose AI like ChatGPT or Gemini): </strong>“You are a creative concept artist. Generate a moodboard with [main theme or vibe]. Include [specific elements: color palette, materials, textures, patterns, or style references]. Prioritize originality and avoid generic stock looks. Make it feel [adjectives: authentic, warm, minimal, edgy, etc.].”</li><li><strong>Define the mood</strong><br>Use AI to analyze the style of your curated assets. Let it help you sharpen your visual direction and discover more aligned references.<br><strong>Example Prompt (for a general-purpose AI like ChatGPT or Gemini): </strong>“You are an experienced art director. Analyze and describe the visual style of this collage in detail. Be technically specific. Use short, descriptive keywords and phrases only, no full sentences. Focus on color palette, composition, materials, textures, mood, and any distinctive style cues.”</li><li><strong>Craft &amp; refine</strong><br>Turn your visual analysis (visual style output from step 2) into generation prompts. Use these to create new moodboards, single assets, or remix styles.</li><li><strong>Generate, iterate, refine</strong><br>Play with AI outputs. Test, tweak, and evolve the prompt until the moodboard feels just right.</li></ol><p>I recommend using Firefly boards for making the process simpler — <a href="https://firefly.adobe.com/boards">https://firefly.adobe.com/boards</a><br>⚠️ <strong>Watch for:</strong> Generic or stock-like results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NzRoqAmOik5yh_pW8WWpfQ.png" /><figcaption>Image 5: Moodboard generation</figcaption></figure><h3>Final Thoughts</h3><p>For complex design workflows, generative AI should serve as a collaborative partner rather than a vending machine. Great AI UX goes beyond the model’s capabilities; it involves control, editing, and reuse patterns that enable your team to improve prompts over time. When executed well, these patterns save time, reduce hidden costs of poor automation, and foster genuine trust in your AI tools.</p><h3>Next Steps</h3><ul><li>Try these workflows on real projects.</li><li>Share your best prompts with your team.</li><li>Prototype with multiple models.</li><li>Experiment with Promptframes.</li><li>Highly recommend checking some <a href="https://www.linkedin.com/posts/adplist_design-activity-7345829036318695424-h_Dj?utm_source=share&amp;utm_medium=member_desktop&amp;rcm=ACoAAA3CMQMBcVTOCIk4Hi6tB26CPSQZCDAxK4c">design prompt cheatsheets</a></li><li>Stay curious! ;) Keep trying AI tools!</li></ul><p><strong>References:</strong> NN/g Promptframes | OpenAI | Gemini | Perplexity | Firefly Boards.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=323607e3b0e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Generative AI in Design Workflows]]></title>
            <link>https://medium.com/@syashigupta/generative-ai-in-design-workflows-b455d44fe22c?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/b455d44fe22c</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[ai-assistant]]></category>
            <category><![CDATA[notebooklm]]></category>
            <category><![CDATA[multimodal-ai]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Wed, 09 Jul 2025 15:02:59 GMT</pubDate>
            <atom:updated>2025-07-09T15:02:59.830Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*27XU06WtmBMY7MQQUbcd7A.jpeg" /><figcaption>Generated using Firefly</figcaption></figure><p>Listen to the podcast instead? <a href="https://notebooklm.google.com/notebook/fcfb9299-76c2-4414-af08-f3842fc053be/audio">Podcast</a></p><p>Generative AI is revolutionizing content creation, offering exciting possibilities for text, image, and video. However, truly integrating it into daily design workflows requires a strategic approach. I see Generative AI not as a magic bullet, but as a powerful, multi-tool assistant. No single model will take you from a blank canvas to a final mockup. Real creativity comes from skillfully combining tools and templating prompts for both teams and end-users.</p><p>We all use general-purpose AI tools for everyday design tasks, but <strong>what more could you strategically integrate and templatize</strong>? This article shares how I embed Generative AI in my design workflows, openly sharing practical ways to turn AI into a more powerful, reusable design partner.</p><h3>Types of AI Usage in Design Practice</h3><p>Generative AI can be categorized by its application in design:</p><ul><li><strong>Exploratory &amp; Strategic:</strong> Using AI for broad understanding and future planning.</li><li><strong>Task-Based Assistance:</strong> Employing AI for specific, often repetitive, design tasks.</li><li><strong>End-to-End Automation:</strong> Automating entire design processes, though this is less common for complex creative work.</li></ul><h3>My Generative AI Design Workflows</h3><p>Here are some practical ways I incorporate Generative AI into my design process:</p><h4>1) Deep Research &amp; Strategic Planning</h4><p>Generative AI excels as an assistant during the Research &amp; Planning phase. It helps me quickly map competitive landscapes and analyze trends. I use tools like ChatGPT, Gemini, and Perplexity to identify direct competitors, summarize their features and positioning, and even uncover gaps. AI also assists in synthesizing academic papers and industry reports, significantly accelerating the discovery and understanding of the topic, as illustrated in Image 1.</p><p><strong>Pitfall to Avoid:</strong> Over-relying on AI summaries without verifying sources can introduce inaccuracies. Always double-check facts.</p><p><strong>Tools:</strong> ChatGPT, Gemini, Perplexity, Copilot</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1ir4KDPl8YuCBbyqe477mQ.png" /><figcaption>Image 1: I used deep research to help synthesize an understanding of the relationship between explainability, trust, control, and feedback</figcaption></figure><h4>2) Mind Mapping &amp; Listening for Insight</h4><p>For complex problem spaces, I generate mind maps with AI during the Discovery &amp; Alignment stage to visualize connections. Some tools can read articles aloud or convert them into podcast-like audio, allowing me to absorb ideas while context-switching. While I’m working, I value any AI tool that transforms static text into dynamic, shareable formats.<br>In the world of attention deficiency, I value bite-sized, visually given information.</p><blockquote><strong>Insight:</strong> Notebook LM can explain complex concepts in simple terms, provide real-world examples, and reinforce understanding.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7mLycjootcQOV0kDIdTY6g.png" /><figcaption>Notebook LM’s mind mapping feature can help you spot logical gaps in your article and see what’s clearly conveyed versus what needs improvement. This article!</figcaption></figure><h4>3) Brainstorming &amp; Ideation</h4><p>During Ideation &amp; Concepting, AI provides rapid bursts of raw ideas. This includes generating user scenarios, headline options, or even varied social media captions to test tone. The key is crafting prompts with clear context, constraints, and examples. I sometimes ask the AI to self-evaluate its outputs or compare ideas based on certain criteria. It’s excellent for quickly generating ideas and for sense-checking different directions before design takes shape.</p><p><strong>Pitfall to Avoid:</strong> AI can produce polished text that appears final but lacks substance. Keep ideas rough until validated with users or your team.</p><p><strong>Tools:</strong> ChatGPT, Gemini, Copilot, Perplexity</p><h4>4) Refining Ideas &amp; Prototyping</h4><p>I use AI to refine rough drafts into workable design prompts during Prototyping &amp; Testing. I’ll feed messy notes or half-baked flows into Copilot or Chat GPT as seen in Image 2, allowing it to ask clarifying questions until the output is ready for tools like Lovable.ai or UX Pilot. AI’s iterative questioning often reveals gaps I might miss independently. I’ve even experimented with turning large Figma dumps into structured slide decks. Images 3 &amp; 4 are examples of my use of AI for the exploration of a topic I am working on. I prototyped using fake data and scenarios to explore multiple ideas, edge cases, and tools.</p><p><strong>Pitfall to Avoid:</strong> Be aware of the hidden cost of iterative prompting, in terms of time, money, and computation. Prompts can become lengthy and difficult to manage.</p><blockquote><strong>Insight:</strong> What did this process yield? Initially, perhaps not ‘peace of mind,’ but after innumerable iterations across different AIs and AI products, gained a crucial understanding: the more precisely I shared exact visuals and defined interactions using a general-purpose AI, the closer I got to the actual desired results in these prototyping apps. (Image 3, Image 4)</blockquote><p><strong>Tools:</strong> ChatGPT, Figma AI, UX Pilot, Lovable.ai, Replit, Figma Make</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0BVbGoCILmK3e4eqvhM6vQ.png" /><figcaption>Image 2: Prompted first using Copilot, but quickly realized the Copilot wasn’t great at prompt writing and so shifted to using Chat GPT.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yJRROp35IA6zaLh0SVJJkw.png" /><figcaption>Image 3: Multiple prompts, some growing larger than I can process, I also used multiple tools to build this, only to realize the same prompts, even when specific, can yield different results in different tools.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7DeKOWSTJBnO2YyBqO8itw.png" /><figcaption>Image 4: Same prompt, different results in Lovable, Ux pilot, Replit. I also tried other UX design tools, but they were not great.</figcaption></figure><h4>5) Getting AI to Critique &amp; Play Roles</h4><p>Before stakeholder reviews, I used AI for role-playing. I ask it to act as a design lead or a skeptical product owner to get feedback. This helps me spot weak points early and present them with more confidence. Some AIs are surprisingly effective at asking tough questions or identifying logical gaps.</p><p><strong>Pitfall to Avoid:</strong> AI feedback should not entirely replace human critique. Over-reliance on AI risks diminishing your judgment and critical thinking skills. <br>Another pitfall to avoid is removing any confidential information from the content.</p><p><strong>Tools:</strong> ChatGPT, Gemini</p><h3>Closing Thought: AI is a Partner</h3><p>If there’s one thing I’ve learned, it’s this: Generative AI shouldn’t be treated like a vending machine for instant answers. It’s more like a creative partner, a sparring buddy that’s only as good as the prompts you feed it and the context you give it.</p><p>No single AI model can do everything. That’s why the best results come when we <strong>mix tools, experiment boldly, template what works</strong>, and keep our critical thinking sharper than the output we get.</p><p>So, take these ideas, adapt them to your practice, and don’t be afraid to prototype, test, and evolve your prompts as you learn. Because with the right balance of trust, control, and curiosity, AI becomes so much more than just a feature.</p><p>PS: There is a part two of this article that shares templates and exercises for AI Design workflows</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b455d44fe22c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Not Sound Like an AI]]></title>
            <link>https://medium.com/@syashigupta/how-to-not-sound-like-an-ai-b50cea7ff183?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/b50cea7ff183</guid>
            <category><![CDATA[ai-detection-tool]]></category>
            <category><![CDATA[humanize-ai-text]]></category>
            <category><![CDATA[design-writing]]></category>
            <category><![CDATA[fact-check-content]]></category>
            <category><![CDATA[ai-writing-tool]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Mon, 09 Jun 2025 20:58:31 GMT</pubDate>
            <atom:updated>2025-06-12T18:43:36.543Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nbKhRzUzYivUG9nWZ0_VTA.jpeg" /><figcaption>Generated using firefly(adobe)</figcaption></figure><p>You refined your text (any length of words) using AI, and now you want to improve it to sound more natural, not like AI. I mean, I am not judging! I do it too. Most of the time, I want help expressing my thoughts more clearly. My writing process involves writing a messy draft and then letting an AI assistant help polish it.</p><p>But here’s the problem— when text is refined using AI, it can set off AI detectors or sometimes be spotted as AI-written or feel unnatural, like something you wouldn’t say out loud. So, how do we fix that?</p><p>No one wants their ideas to be flagged as “not human enough.” Especially when, ironically, the original thought was all you.</p><h3>Why does this matter?</h3><p>Because there are limits to these AI detectors. Some studies suggest <a href="https://www.cell.com/patterns/fulltext/S2666-3899(23)00130-7">AI-detection tools are biased, especially against non-native English speakers.</a> Others say these t<a href="https://arxiv.org/pdf/2306.15666">ools are right only about 80% of the time</a>.</p><p>Additionally, using AI gets some wrinkled-nose reactions, so it can feel shameful, and as a result, people might even feel guilty using it. But I want to pose a rhetorical question here — Would you want to hire someone to edit your résumé? Probably not. You’d likely prefer to refine it using AI. The same goes for writing cover letters tailored to specific companies or articles like this one. These high-impact use cases, like résumés or cover letters, shouldn’t cause stress, especially for non-native speakers.</p><p>Please note: I am not trying to promote AI-generated work or research, but rather AI’s use to augment a writer’s ideas.</p><p>So, where does that leave us?</p><p>By the time you finish this, AI models may have evolved again. Still, you can take steps to make AI-assisted writing sound more like you. These aren’t perfect solutions, but they can make a big difference, especially when it comes to sounding authentic.</p><h3><strong>What to do?</strong></h3><p>As a writer, one may use these to improve their AI-refined text to sound more natural.</p><h4>1. Break up long sentences</h4><p>AI-generated text often leans formal, stringing complex clauses together into one long, breathless sentence. Break those into two. Use simpler words. Swap “utilize” for “use.” Your writing should breathe<em>.</em></p><h4><strong>2. Ease up on bullet points</strong></h4><p>I love bullet points. They’re easy to follow and fast to write. But when overused, they appear automated. Try mixing in full paragraphs to keep the rhythm varied and more natural.</p><h4>3. Double-check your dates and sources</h4><p>AI has a weird blind spot for dates, even when they’re right in front of it. I once asked a chatbot to cite a speech using APA style. I gave it the correct date. It still got it wrong. Another AI tool repeated the same mistake. Turns out the model may have been trained before the speech even took place.</p><p>Bottom line? Don’t trust AI with the details. Always verify dates, links, author names, and publication info yourself.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pAvnwNdVoeuZ5qGDgwzfhg.png" /><figcaption>This YouTube video was uploaded in 2025, and I attended the talk in person, so I was curious why ChatGPT got the date wrong.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SnfQ7pDWaoCJS89F9mfRxQ.png" /><figcaption>Used google studio for the same and got a different incorrect answer</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_IBVeqidgW-raveYgG_Log.png" /><figcaption>Asked for date got time instead as a reply</figcaption></figure><h4>4. Fact-check everything</h4><p>AI can confidently invent sources, statistics, and even article titles. Sometimes it cites paywalled papers or ones that don’t exist. Always check numbers, sources, and capitalization.</p><h4>5. Watch for odd punctuation, especially em dashes</h4><p>AI loves em dashes, and sometimes uses them oddly. If a paragraph had some “:” or “-”, try rewriting it with commas or consider splitting it up.</p><h4><strong>6. Edit AI Text and Question Its Reasoning</strong></h4><p>AI-generated text can feel “off” in a few key ways. First, there’s the writing itself, which is often too polished, too balanced. No typos, no quirks, no rhythm shifts. But human writing has texture. Some sentences are short, others run long. That’s why it’s important to heavily edit AI-written text to bring back that human feel.</p><p>Second, there’s the reasoning. <a href="https://arxiv.org/abs/2506.02126?utm_source=alphasignal">AI can be surprisingly off in its logic or suggestions</a>. I’ve found that asking the AI to explain its edits or recommendations helps surface its reasoning. If that reasoning doesn’t make sense, it’s a cue that the writer needs to step in and revise. Always question AI outputs, both for tone and for thought process.</p><h4>7. Use document history</h4><p>Tools like Google Docs’ version history let you track your process, refine ideas over time, and retain your voice. Keep an eye on the “before” and “after” to ensure AI edits don’t stray too far from your original wording.</p><h3>So what can you do?</h3><p>Stay informed. Use AI if it helps you clarify your thoughts, but edit the output with a human touch using some guiding ideas above.</p><p>I am a huge believer that originality of ideas is key, but we don’t need to push away these tools that can augment and assist us in so many ways, and more importantly, enhance an authored text more intelligibly. We should leverage them to our advantage without carrying the guilt for taking AI’s help.</p><h4><strong>Some notes</strong></h4><p><strong>Detection in generative AI</strong><br>It used to be easy to spot AI content, and the uncanny valley gave it away. But now, AI-generated images, videos, and text are far more convincing. There’s no unique signal that proves if AI wrote something. Tools like GPTZero, Originality.ai, or Perplexity can offer clues but not certainty. Even advanced detection methods, like metadata tracing or pixel-level analysis in images, aren’t foolproof.</p><p>Some researchers are exploring watermarking AI-generated text, embedding invisible signals into language. But that only works if everyone agrees to play by the same rules, which isn’t likely anytime soon.</p><p>// Something else that I found and it&#39;s pretty cool about maintaining authorship -<a href="https://ia.net/topics/ia-writer-7">https://ia.net/topics/ia-writer-7</a><br>//Grammarly also has something very similar called Grammarly -Authorship, and it is an API that can be used on any web document writer!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b50cea7ff183" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Theory to look at if you want to have better health behavior!]]></title>
            <link>https://medium.com/@syashigupta/theory-to-look-at-if-you-want-to-have-better-health-behavior-6a4ede64ef50?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/6a4ede64ef50</guid>
            <category><![CDATA[five-stages-of-change]]></category>
            <category><![CDATA[healthy-lifestyle]]></category>
            <category><![CDATA[theory]]></category>
            <category><![CDATA[ttm]]></category>
            <category><![CDATA[behavior-change]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Sun, 08 Mar 2020 16:17:46 GMT</pubDate>
            <atom:updated>2020-03-08T16:17:46.381Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Do you want to exercise regularly, quit smoking, cut down sugar, leave crunchy amazing chips, or cultivate some other healthy behavior? TTM is the theory you should be looking at if you haven’t.</em></p><p>For some time I have been struggling with the idea of going to the gym because I always had an excuse to not exercise. And in time I was given the topic Transtheoretical Model(TTM) or stages of change theory to research for a class and this got me interested in drilling down deep and learning theory’s practical applications. The TTM theory suggests that changes in healthy behavior occur in <strong><em>six stages</em></strong>, and that mechanism of change incorporates <strong><em>cognitive and behavioral processes</em></strong> that individuals engage in at different stages. In addition, individuals will weigh the <strong><em>pros and cons of engaging(decisional balance) </em></strong>in an activity and will generally experience <strong><em>self-efficacy</em></strong> as they advance through the stages of activity behavioral change. The framework looks like:-</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*sFWzHc-4EzNfvPVnPftCqg.png" /><figcaption>Framework constructs- Stages of change, processes of change, decisional balance, and self- efficacy</figcaption></figure><p>To explain this model better, let me use an example. Say for example I am obese because of my diet from the last 3 years of a lot of sugary and trans-fat food. I am at this pre-contemplation phase where it doesn’t matter to me if I am obese or not. I haven’t thought about exercising or eating healthy but recently I start to hear from my friends, doctors, and family who make remarks about how I need to change my diet and workout. This hits me and eventually in some months by moving through the three processes of change:- <em>Consciousness-Raising, Dramatic Relief, and Environmental Re-evaluation, I enter a new stage</em>:- <strong>Contemplation stage</strong>. Additionally, I weigh the pros and cons(decisional balance)of moving from one stage to another. I switch stages and move through processes of change when the pros of starting to thinking about healthy behavior at the pre-contemplation stage are more than cons. Along with decisional balance, I measure my ability to make a change. Unless I feel I am capable of changing and going through the stages of change I won’t do it. Similarly, I move from stages of contemplation to preparation, from preparation to action, and finally from action to management by moving through other processes of change, decisional balance, self-efficacy. In TTM there is a sixth stage called — <em>termination stage</em> or the<em> relapse stage </em>which is<em> </em>used for addictive behaviors like drug abuse, smoking, or obesity, etc. For health behaviors, like actively researching about their own health or, taking recommendations on phishing attacks, there is no need for a termination or relapse stage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qq44zjubjdIUD0n_xclvQg.png" /></figure><p>Original research paper by Prochaska and Diclemente suggested that people move from one stage to another through some processes of change. But from the research papers, what became evident was that people don’t necessarily use exactly these processes of change or to move from one stage to another in a linear/cyclic fashion! Basically suggesting that one can have a distinct trajectory of change.</p><p>TTM is used as a method to suggest interventions and hence is used mostly by clinics/doctors. Clinics/doctors try to gauge their participants/patients&#39; stage of change and try to push them through to the stages of action and maintenance. Many clinical applications directed to target a specific group of patients can use TTM to push patients forward into managing their own health by knowing their current stage and then using targeted interventions- to get specific how about Fitbit using TTM!!.</p><p>There is no theory without limitations and TTM comes with its own set of problems but irrespective of that I personally think TTM can be used well along with other models in a wide variety of fields to direct behavior change- like finance, nutrition, technology, etc. TTM has also been used in learning models to see if a child is in action/ preparation stages to perform better at school and direct interventions like gaming to push them to perform better. Through this article, I am suggesting that teachers/ coaches/ instructors can guide people by actually understanding which stage of change they are in and pushing them through to other stages. Or even people who understand this model can create a plan for themselves. You don’t require an app or software for it. You know yourself better!</p><p>I realized that after getting to know TTM I became more cognizant of my behaviors and through a self-directed plan, I could bring myself to go to the gym at least 5 days a week(been 3 weeks now)! In the original papers, it was suggested that people who quit smoking without any therapy, had a pattern and they could make a change to healthy behavior by moving through the stages of change. If you have read till this point, I am sure you are now familiar with this theory so what are you waiting for? I know you can make a change of being better than you are but of course more consciously this time.</p><p><em>I am not an expert on this topic. But please do let me know in the comments if you have some constructive criticism!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a4ede64ef50" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is personalization making you narrow-minded?]]></title>
            <link>https://medium.com/@syashigupta/is-personalization-making-you-narrow-minded-4769c2875a5?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/4769c2875a5</guid>
            <category><![CDATA[algorithmic-bias]]></category>
            <category><![CDATA[filter-bubble]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[cognitive-bias]]></category>
            <category><![CDATA[personalization]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Tue, 04 Feb 2020 23:31:29 GMT</pubDate>
            <atom:updated>2020-02-04T23:31:29.843Z</atom:updated>
            <content:encoded><![CDATA[<p><em>When Medium says” suggested articles based on your previous likes”. Personalization is now everywhere. It is in the apps that you use- Spotify, YouTube, Facebook, Instagram, Google Home, and so on many more… What is this personalization doing to us? Is it good or bad or poisonous?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/792/1*q2TZ69JpDxoxbSh0ORYbjA.png" /><figcaption>Personalization</figcaption></figure><p>This question came up in my head after I watched this <a href="https://www.youtube.com/watch?v=B8ofWFx525s&amp;ab_channel=TED"><strong>Ted video</strong></a> on the filter bubble. though the video came out in 2011 it is still very relevant today in 2020. This video made me take note and list down all things that had happened recently to me since personalization began to rule my life and I began to see the implications. I had to delete my Instagram application because I was overwhelmed with the kind of information I was getting on it. I read in one of the stories about protests in India and followed that thread for 1–2 days and then what happened next was horrifying. My whole feed, all the stories I saw were full of posts on protests. It was an emotional and cognitive overload. <em>There is only so much that a person can feel and consume</em>. After actively consuming it for 15–20 days I saw I was getting anxious and stressed and I decided to get out of Instagram. In those 15 days, where did the artistic feed that I thoroughly used to enjoy every day on Instagram disappeared? I had a similar case with YouTube where it was stuck on the recent songs that I had heard and played them on the loop till I got really bored with all of them.</p><blockquote>First, Machine learning feeds off me and then it ends up feeding me similar things to the ones I liked</blockquote><p>The technology driving this personalization is <strong>machine learning</strong>. First, Machine learning feeds off me and then it ends up feeding me similar things to the ones I liked. Does this kind of personalized feed make people happy? Yes most of the time, instead of a world view on social media, I get a partial view. Just like Spotify has a matrix matching model for producing your loved Discover weekly. We(as population) unknowingly are being categorized by algorithms into groups. There is another poignant thing that was mentioned in the Ted video was on how long before ML there were newspaper editors that did a similar task of providing us edited information. I agree but I disagree too. The extent to which it is happening now along with the repercussions it has on data privacy and human behaviors is scary. The data being collected on me that then that collected data decides the kind of data that reaches me is something I don&#39;t know. Machine learning is generating vicious feedback loops that are inhibiting our growth as a society.</p><p>When you see everyone around you saying the same “thing” as you, and you start to feel validated thinking that “thing” to be the norm? What kind of effect does this validation have on us? Do you feel empowered? It is that positive affirmation that also makes us narrow-minded. With just one world view presented how can we think of expanding and learning? If you are a liberal or a conservative you will remain so because that is what you will be surrounded by on your internet world. If you love pop songs you will get an endless feed of pop songs similar to what you listen to. Another thing that is happening that I noticed is this whole personalization is making us a bad listener. Think about it. Eg: If 99% of people around me are saying “AA” is great then that 1 % of people telling me otherwise wouldn’t matter. Unless I make an open attempt to listen to them and try to understand their point.</p><p>The point above also treads into the idea of <em>cognitive bias</em> in psychology. A cognitive bias is a systematic error in how we think, as opposed to random error/ignorance. A cognitive bias distorts the truth and it can’t be fixed by getting more data, especially data like personalized data, their effects can add up over time. Even if you want to try to <em>de-bias</em> yourself how do you do it? If you can’t trust yourself, or the ML algorithms, or the people around then who do you trust? Let me give you an example my friend discussed with me:- <strong>“Is coronavirus a biological attack from China?”</strong>. Interesting to think about right? But as soon as this topic was touched upon, in a quick second I told my friend “This is a conspiracy theory!”. Whatever articles I have read to date none of them ever made any claim like that. It wasn’t until my friend said it. He further went on to tell me what he reads on google scholar and I still have trouble believing him. Because what if his google scholar is just trained to show him articles like these? At this point, however, you should know that even I have lost sense of what is true and what isn’t!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/792/1*AtqxFP0cHDUlI-wlHd5oyA.png" /><figcaption>Blinders on</figcaption></figure><p>How do we save ourselves from putting those blinders on?<br>I don’t know! I do know that one should keep clearing cookies on a web browser and use incognito mode to search. But what about other social media mobile applications that are personalized for me?</p><p>I never intended to leave this article with an unresolved question but I don’t know the answer to this. Next time when Netflix recommends you a series just pause and think for a minute is this actually good for me?</p><p><em>I would love to know your point of view and do share if you have any interesting insights on this :)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4769c2875a5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to make a chatbot in Dialogflow?]]></title>
            <link>https://blog.chatbotslife.com/how-to-make-a-chatbot-in-dialogflow-5e5f4cd30ba3?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/5e5f4cd30ba3</guid>
            <category><![CDATA[dialogflow]]></category>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[conversational-ui]]></category>
            <category><![CDATA[bots]]></category>
            <category><![CDATA[chatbot-design]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Sun, 02 Feb 2020 04:37:40 GMT</pubDate>
            <atom:updated>2020-02-04T15:05:20.041Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Intents, contexts, entities…. What are these? How do I convert a simple conversation into a Dialogflow Intent, entity, sub-intent? I made a chatbot for my </em><a href="http://www.syashigupta.com/"><em>website</em></a><em> using dialogflow and here is how I made it.</em></p><p>Dialogflow is a robust platform to make a <a href="https://chatbotslife.com/">chatbot</a> on, but only a primitive one. I will share my challenges working with Dialogflow towards the end. If you are a designer and you want to make your own conversational bot you can do so in Dialogflow. This product can also serve as a brief introduction to conversational interfaces and the underlying technology- machine learning(training). My intent for this article is to share <strong>my mental schema of converting a conversation flow</strong> into <strong>intents, sub-intents, and entities</strong> on Dialogflow. I was struggling while explaining this to my project partner so I thought of breaking it down through this post.</p><figure><a href="https://www.chatbotconference.com/"><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5QA4jEOW_w_mKaAW" /></a></figure><p>This article already assumes that you are familiar with Dialogflow. If not then please watch this Linkedin learning video:- <a href="https://www.linkedin.com/learning/building-chatbots-using-google-dialogflow?u=53565897">Building Chatbots Using Google Dialogflow</a>. To create your own chatbot first make a conversation flow diagram (participants should take turns in a conversation flow because it is a very primitive chatbot!).</p><h3>Trending Chatbot Tutorials</h3><blockquote><a href="https://chatbotslife.com/building-a-successful-chatbot-d210bec69773">1. Building a Successful Chatbot</a></blockquote><blockquote><a href="https://chatbotslife.com/write-a-serverless-slack-chat-bot-using-aws-e2d2432c380e">2. Write a serverless Slack chat bot using AWS</a></blockquote><blockquote><a href="https://chatbotslife.com/chatbot-using-django-rest-framework-api-ai-slack-part-1-3-69c7e38b7b1e">3. Chatbot using django rest framework + api.ai + slack</a></blockquote><blockquote><a href="https://chatbotslife.com/chatbot-conference-2020-e56430926042">4. Chatbot Conference 2020</a></blockquote><p>I made a conversation flow for my website chatbot and it looks like this:- Don’t judge me, please. I was just roughly laying out my bot. If your conversation flow is sequential and not as spread out like mine don’t worry. You can think of dividing your sequence into broader intents. I will explain, follow along :)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zwIz50BLGTbPNs3OjnIqqg.png" /><figcaption>Conversation flow for chatbot on my website</figcaption></figure><p>Let’s begin:- Once you have a simple conversation flow laid out you can easily convert it into Dialogflow by following this procedure:-</p><ol><li><strong>Top Questions or queries that the user asks your bot</strong> changes into <strong>INTENT </strong>in Dialogflow<strong>.</strong></li><li><strong>Questions or Queries one level down</strong> form <strong>SUB-INTENTS</strong> in Dialogflow.</li><li><strong>Questions or queries that the user asks your bot</strong> make up the <strong>TRAINING PHRASES</strong> in that particular <strong>INTENT.</strong></li><li><strong>TRAINING PHRASES</strong> will have <strong>ENTITY</strong> property- this is how the DialogFlow system extracts certain words and then matches them to an intent.</li><li><strong>Your bot’s answers</strong> are <strong>RESPONSES.</strong></li><li><strong>CONTEXT</strong> is basically like a broad category. Ex. If you have a skincare intent then skincare products and all sub-intents below the skincare intent will come under one context.</li><li>Remember you can only go <strong>4 levels</strong> after an intent. Let me explain this with an example:- skincare intent-&gt;face care subintent-&gt; lip care subsubintent-&gt; Lip problems subsubsubintent-&gt; sell lip care product subsubsubsubintent. (Overall 5 levels as shown in the image below)</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*swQbNejmG3Aj-3aCS5aWLA.png" /><figcaption>Entities, context, intents, sub-intents. I have used another sequential conversation flow(bottom right) to explain how intents and sub-intents can be found in sequential flows.</figcaption></figure><p>I have tried to break down the process and make it as easy as I could. Here is an example intent from my website <a href="https://chatbotslife.com/">chatbot</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NABEAnEy946Ss4-E075Hnw.png" /><figcaption>About you Intent- when a user asks about me and the training phrases</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ygjz04MMqO5esI9w9BZvYA.png" /><figcaption>The entity(Aboutyou) used for training and the response.</figcaption></figure><p><strong>CHALLENGES</strong></p><ol><li>Conversations happen differently than text conversations. People say differently than they write. So need to keep that in mind while training the <a href="https://chatbotslife.com/">bot</a> with phrases. example- supercalifragilisticexpialidocious. If someone is texting they will write this text as it is but when speaking out loud they will break it into smaller words.</li><li>There will be a question and then there will be a response. There is no way a chatbot can ask questions or respond unless prompted by the user. So an external prompt is needed and the conversations will always happen in turn-taking sequence.</li></ol><p>If you find any errors or if you have suggestions for me to improve please put them in the comments below. I am always looking for constructive feedback :D</p><p><em>I would like to thank my professor Helen for introducing me to CUI, ML/AI- and everything in this bucket.</em></p><h3>Don’t forget to give us your 👏 !</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/780/0*ZxaGcKC3UQEydQIX" /></figure><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fupscri.be%2F3e59ef%3Fas_embed%3Dtrue&amp;dntp=1&amp;url=https%3A%2F%2Fupscri.be%2F3e59ef&amp;image=http%3A%2F%2Fapi.screenshotlayer.com%2Fapi%2Fcapture%3Faccess_key%3Dfe59908dad3baab69ffab249a2224b03%26viewport%3D1024x612%26width%3D1000%26url%3Dhttps%253A%252F%252Fupscri.be%252F3e59ef%253Fscreenshot&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=upscri" width="800" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/7078d8ad19192c4c53d3bf199468e4ab/href">https://medium.com/media/7078d8ad19192c4c53d3bf199468e4ab/href</a></iframe><figure><a href="https://chatbotslife.com/bot-communities-mastermind-group-d2dae9876709#.53x0py6ou"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*6XUspT9JOSq0w0Fi35HIaA.png" /></a></figure><figure><a href="https://m.me/ChatbotsLife"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*c1LDMH5vbnIz9rmAka8Hwg.png" /></a></figure><figure><a href="https://chatbotslife.com/chatbot-development-hire-top-ai-chatbot-developers-c8b45ef7f207"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*D0Jf3dI6ZThtqcfwDYY7mg.png" /></a></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5e5f4cd30ba3" width="1" height="1" alt=""><hr><p><a href="https://blog.chatbotslife.com/how-to-make-a-chatbot-in-dialogflow-5e5f4cd30ba3">How to make a chatbot in Dialogflow?</a> was originally published in <a href="https://blog.chatbotslife.com/">Chatbots Life</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Conversational Interfaces?]]></title>
            <link>https://medium.com/@syashigupta/why-conversational-interfaces-f5f994d9225a?source=rss-7763a4c26639------2</link>
            <guid isPermaLink="false">https://medium.com/p/f5f994d9225a</guid>
            <category><![CDATA[chatbot-design]]></category>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[conversational-ai]]></category>
            <category><![CDATA[conversational-ui]]></category>
            <dc:creator><![CDATA[Syashi Gupta]]></dc:creator>
            <pubDate>Sat, 01 Feb 2020 20:00:55 GMT</pubDate>
            <atom:updated>2020-02-02T01:20:05.539Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/792/1*5NDgQfRKlhf7nzqO-3ehzQ.png" /><figcaption>Remember how Eliza was a big hype? When it came out. Even though people knew it was a computer program still they wanted to be alone with this computer!</figcaption></figure><blockquote>“Gartner Research predicts that 50% of businesses will spend more on conversational applications than mobile apps by 2021.”</blockquote><p>Let’s try to break down <strong>why is there such a big rush after conversational interfaces?</strong></p><p><strong>Better than humans</strong>- For one we think chatbots are better than humans at the job that is defined for them. Sherry Turkle beautifully puts these lines in the book reclaiming conversation. She says <em>“whatever people do if a robot can take over that role, it was by definition, not specifically human. And over time, this has come to include the roles of conversation, companionship, and caretaking. We redefine what is human by what technology can’t do. But as Alan Turing puts it, the computer conversation is an imitation game”</em>. Companies bring in these conversational bots to actually satisfy customers and save spending money on humans helping customers.</p><p><strong>Reduction in Cognitive Load</strong>- We are trying to take away that cognitive load from people’s brains so that people can target their energies on other crucial things. So this conversational interface actually finds the information, does the analysis, and gets you the results. We are running away from seeking meaning or putting in efforts. Anything that requires energy like a human relationship, we seek better alternatives(CUIs). For anything that makes people read and search they prefer an assistant/interface to tell them what they need to know. They want to free themselves from acting/ thinking/ doing/ meaning. I know this is a very dystopian view and this isn’t going to happen anytime soon. But still, essentially that is the kind of experience we ultimately are designing for. CUIs as companions, or as social agents.</p><blockquote><em>To put things in perspective, today you’ll find at least </em><a href="https://www.sitebuilderreport.com/best-website-builder"><em>40 website builder tools</em></a><em>, more than </em><a href="https://financesonline.com/help-desk-statistics-analysis-of-trends-data-and-market-share/"><em>35 helpdesk platforms</em></a><em>, and over </em><a href="https://www.capterra.com/customer-relationship-management-software/user-research"><em>350 CRM options</em></a><em>. Meanwhile, Gartner identified at least </em><a href="https://uxmag.com/articles/a-six-minute-breakdown-of-gartner%E2%80%99s-top-conversational-ai-platforms"><em>1,000 conversational platforms</em></a><em> in the market. That’s just platforms for building conversational applications; the number of actual conversational applications is exponentially greater.</em></blockquote><p><strong>Cross-platform and reduces friction</strong> — It is easy to integrate CUIs on all interfaces, whether it is desktop, web, mobile, tablet. CUIs are being made better with time and people can easily ask CUIs for help on any interface they have trouble with. Ask Siri! and it will explain it to you. So first it is taking away the cognitive load and then making the technology more acceptable for users.</p><p><strong>Accessibility </strong>— Ever had your keyboard and trackpad taken away from you and all you have is a mouse and accessibility keyboard? It was my worst nightmare and at that moment I wanted a very smart CUI to do some of my tasks. I spent 35 mins trying to sign in to google drive. CUIs can be a really big step in accessibility direction.</p><p><strong>User Acceptance</strong> — It has been found that people love talking to robots! (:O) Humans forget what it means to be human when they are talking to robots. I feel it is just how societies function now. Human-machine interaction rather than human-human! Our mental models have shifted a little. We are redefining trust, judgment, and intimacy. We as a culture treat technology as infallible and we trust the products that technology produces. This is a piece of good news for companies and designers designing for CUIs.</p><p>I think there is more to CUIs than this list. CUIs are changing our behaviors and needs. CUIs are everywhere and available at all times that no human can match that ability. These perfected superhuman powers are desirable in the market! Its only natural that the demands for CUIs will only grow. But just like a bubble which grows and then bursts. I am excited to know what new technology will be coming up next after CUIs?</p><p><strong>References:-</strong><br>https://uxmag.com/articles/why-this-ux-role-will-grow-fastest-in-2020</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f5f994d9225a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>