<?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 Burak Şahin on Medium]]></title>
        <description><![CDATA[Stories by Burak Şahin on Medium]]></description>
        <link>https://medium.com/@burak.sahin?source=rss-ae83fa59e7ea------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*NvaRlInpKjeQkk2YvTH7sA@2x.jpeg</url>
            <title>Stories by Burak Şahin on Medium</title>
            <link>https://medium.com/@burak.sahin?source=rss-ae83fa59e7ea------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 10:01:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@burak.sahin/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[The Design Decisions AI Needs to Know]]></title>
            <link>https://medium.com/@burak.sahin/the-design-decisions-ai-needs-to-know-acc2adf20a7b?source=rss-ae83fa59e7ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/acc2adf20a7b</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[design-systems]]></category>
            <dc:creator><![CDATA[Burak Şahin]]></dc:creator>
            <pubDate>Mon, 09 Nov 2026 07:31:01 GMT</pubDate>
            <atom:updated>2026-11-09T07:31:01.623Z</atom:updated>
            <content:encoded><![CDATA[<p><em>As more people use AI to build products, shared decisions become more important.</em></p><figure><img alt="Two designers and a square-faced AI character gather around a shared design guide showing a circle, a square, and a rounded button." src="https://cdn-images-1.medium.com/max/1024/1*xA7gmzUQaYgYursn0PJYFg.png" /><figcaption>Shared design decisions give people and AI a common starting point.</figcaption></figure><p>A designer can work with AI and create a consistent experience without a formal design system.</p><p>With clear direction, useful examples, and enough context, the screens can look and feel like they belong together. The designer can guide the process, notice differences, and make adjustments along the way.</p><p>But what happens when another designer starts working on a different part of the same product?</p><p>They may describe the problem differently, provide different examples, or make different assumptions. Both designers can get good results. Their work may even be consistent within each flow.</p><p>But when those flows come together, the product can start to feel fragmented.</p><p>A button behaves differently on two pages. One flow explains an error clearly, while another only says “Something went wrong.” A familiar action moves to a different place on a new screen.</p><p>None of these differences has to come from a bad decision. They can simply come from decisions being made separately.</p><h3>The Decisions Behind the Screens</h3><p>Every product has design decisions, whether or not they are documented in a formal system.</p><p>Some are easy to see: colors, typography, spacing, and components. Others are less visible: why a certain action needs confirmation, how much information belongs on a page, or how the product should speak when something goes wrong.</p><p>These decisions often live in different places.</p><p>A designer might remember a discussion from six months ago. A developer might know which component is used across the product. Another team might have documented a similar pattern in a file that nobody else has seen.</p><p>When people work with different pieces of context, AI receives different pieces of context too.</p><p>A shared design system and clear documentation can give the team a common starting point. They can also help each person explain the product more clearly to the tools they use.</p><h3>The Decisions Worth Sharing</h3><p>This does not mean every team needs to document every detail before using AI.</p><p>A useful place to start is with the decisions that come up again and again. Here are a few questions I would look at.</p><h3>1. What Are the Visual Foundations?</h3><p>Colors, typography, spacing, borders, and layout give a product its visual rhythm.</p><p>Imagine two designers building separate settings pages with AI. Both pages look clean, but each uses a different spacing scale and a slightly different approach to headings.</p><p>Neither page feels wrong by itself. The difference becomes more noticeable when someone moves between them.</p><p>A shared set of visual foundations gives both designers a reference. Even a small set of agreed values can help, especially when it is connected to the styles and tokens used in code.</p><h3>2. Which Components Should We Use, and When?</h3><p>A component library can show what is available. It also helps to explain why and when to use each component.</p><p>For example, one designer might ask AI to create a confirmation step before deleting an item. Another might ask for the same thing in a different flow.</p><p>If the product has an existing confirmation pattern but its purpose is not clear, the two results may use different dialogs, language, and actions.</p><p>A useful description could answer:</p><ul><li>What problem does this component solve?</li><li>When should we use it?</li><li>When is another pattern a better choice?</li><li>Is there an existing version in code?</li></ul><p>The answer does not need to be long. A few examples can make the decision easier for both people and AI to understand.</p><h3>3. What Happens Beyond the Happy Path?</h3><p>A screen may look complete when everything works. The rest of the experience still needs attention.</p><p>What happens while data is loading? What does the user see when there are no results? What happens when something fails?</p><p>Imagine two teams building search experiences. One shows a helpful empty state with a suggestion. The other leaves the results area blank.</p><p>Both teams may have focused on making search work. But the person using the product experiences two different approaches to the same situation.</p><p>Examples of loading, empty, error, and success states can help teams build a more connected experience.</p><h3>4. How Does the Product Speak?</h3><p>Visual consistency is only one part of the experience. The language matters too.</p><p>One screen might say “Something went wrong.” Another says “Oops! We hit a snag.” A third shows a technical error with no explanation.</p><p>A shared content guide can describe the product’s tone and provide examples of button labels, confirmations, and error messages.</p><p>“Be clear and friendly” is a useful direction. A few real examples make it easier to apply.</p><h3>5. How Should the Experience Adapt?</h3><p>Different designers may make different assumptions about smaller screens, keyboard use, focus states, or long content.</p><p>A table might work well on a laptop but become difficult to use on a phone. A custom menu might look right but be hard to navigate with a keyboard.</p><p>Documenting expected responsive and accessibility behavior gives the team a better starting point. Existing components and automated checks can help, but the experience still needs to be reviewed and tested by people.</p><h3>6. Why Was the Decision Made?</h3><p>This might be the most valuable detail, and the easiest one to leave out.</p><p>A rule can tell us what to do. The reason behind it helps us understand when it applies.</p><p>For example, a team may choose a small confirmation dialog for destructive actions because it lets people understand the consequence without losing their place.</p><p>If a new situation is different, knowing the reason helps the designer and developer decide whether to follow the same pattern or adapt it.</p><p>The documentation does not need to capture every conversation. A short note about the problem, the decision, and the reason behind it can be enough.</p><h3>A Shared Starting Point, Not a Fixed Answer</h3><p>A design system should help people make decisions. It cannot make every decision for them.</p><p>Products change. New problems appear. Sometimes an existing component does not fit, and the right answer is to improve the system rather than force the work into it.</p><p>This is also why documentation needs to stay connected to the product. If the design file, code, and written guidance describe different things, people may follow the wrong source — even when they are trying to do the right thing.</p><p>Designers and developers can help maintain that connection together. AI can support the work by surfacing gaps or suggesting updates, but the team still needs to review those suggestions.</p><h3>Start With What the Team Needs</h3><p>A small team may not need a large design system. A growing product may need more structure.</p><p>The useful question is: <strong>which decisions do we keep making separately, and where would shared guidance help?</strong></p><p>Maybe the first step is agreeing on a few visual foundations. Maybe it is documenting three common components. Maybe it is writing down how the product handles errors.</p><p>The system can grow from there, as the team learns what it needs.</p><h3>More Ways to Build, More Need to Share</h3><p>What interests me is that AI gives more people a way to turn ideas into working experiences.</p><p>A single designer can go far with the right context. As more people build different parts of the same product, sharing that context becomes more important.</p><p>The goal is not to make every designer work in exactly the same way. It is to give everyone enough understanding of the product to make decisions that work well together.</p><p>I’m still thinking about what the right amount of documentation looks like. Too little can leave teams making the same decisions separately. Too much can become difficult to maintain.</p><p>If your team is using AI with a design system, I’d love to hear what has helped. Which decisions have been worth documenting, and where do you still rely on conversation and judgment?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=acc2adf20a7b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Designers Can Build More of What They Design]]></title>
            <link>https://medium.com/@burak.sahin/designers-can-build-more-of-what-they-design-b685f152aaae?source=rss-ae83fa59e7ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/b685f152aaae</guid>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[design-thinking]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Burak Şahin]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 22:01:58 GMT</pubDate>
            <atom:updated>2026-10-29T22:01:58.537Z</atom:updated>
            <content:encoded><![CDATA[<p><em>AI is bringing designers closer to working products. It also gives us more to think about.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oGvJqN9xq2fz3EemOSacTg.png" /></figure><p>For a long time, a design could feel finished before anyone could use it.</p><p>The screens were ready. The prototype worked. The details were documented. But the product still had to be built, and many small decisions were waiting on the other side of that process.</p><p>How should this behave when the content is longer? What happens while the page is loading? Does the interaction still feel right on a phone?</p><p>These questions have always been part of design. What is changing is how directly we can work on the answers.</p><h3>What Happens After the Prototype?</h3><p>With AI helping us write and change code, more designers can turn an idea into something that actually runs. For some projects, that can mean taking a small product all the way to a live page.</p><p>That is a meaningful change. A prototype shows an intention. A working product lets us experience the consequences of it.</p><p>A layout that looked clear with three cards may feel crowded with twenty. A nice animation may slow down a simple task. An empty state may make sense in a design file but leave a real person unsure what to do next.</p><p>Getting closer to the working version makes these gaps easier to see.</p><blockquote><strong>The design continues after the screen looks finished.</strong></blockquote><h3>A Smaller Gap Between an Idea and a Test</h3><p>Imagine a small internal tool that helps a team find the right document.</p><p>A designer can explore the flow, use AI to build an initial version, and put it in front of a few colleagues. They can try it with realistic content and notice where people hesitate. The next change can happen in the same working version.</p><p>This does not make every project simple. It makes one part of the process more direct: we can test some decisions in the place where people will experience them.</p><p>A loading state becomes something we can wait through. Keyboard navigation becomes something we can try. A confusing label becomes a problem we can fix and check again.</p><p>The useful part is the shorter distance between a question and something we can learn from.</p><figure><img alt="Design. Build. Learn. A red arrow leads back to design. When people use what we build, we see what needs to change." src="https://cdn-images-1.medium.com/max/1024/1*vTl7ClRNv4v3q_srIDNTVA.png" /><figcaption>Design continues in the working product.</figcaption></figure><h3>Where Do Developers Fit In?</h3><p>I think this makes the relationship with developers more interesting.</p><p>A designer can bring a working idea to the conversation. A developer can question how it is built, help it fit the existing system, and spot problems that are easy to miss from the interface alone.</p><p>We can discuss the same thing while using it. The conversation becomes more concrete: this interaction feels slow, this component already exists, this data should come from somewhere else.</p><p>Developers bring experience with architecture, security, performance, and maintenance. AI-generated code still needs that judgment. A page opening successfully tells us very little about how it will behave over time.</p><p>Designers bring an understanding of the user, the flow, and the small details that make an experience clear. Having more ability to build gives us another way to contribute that understanding.</p><blockquote><strong>There is more room to work on the product together.</strong></blockquote><figure><img alt="Different skills. One product. Equal margins align design and engineering contributions beside a centered red product block." src="https://cdn-images-1.medium.com/max/1024/1*PLtae-FlQPeg6UAbV67Ggg.png" /><figcaption>Different contributions, shared responsibility.</figcaption></figure><h3>Building More Means Owning More</h3><p>Taking a design further also means staying with more of its consequences.</p><p>If we help release a tool, we need to think beyond how it looks on the first day. Can people use it with a keyboard? What happens when a request fails? Who will fix it when something changes?</p><p>The amount of review depends on the project. A small page and a feature connected to customer data need different levels of care. Being able to generate the code does not answer those questions for us.</p><p>This is where ownership starts to feel different. The work includes checking what we built, listening to people using it, and agreeing with the team on how it will be supported.</p><h3>Start With Something Small Enough to Understand</h3><p>A focused page, a small internal tool, or one interaction can be a useful place to begin.</p><p>Choose a clear problem. Build a version you can explain. Review it with a developer. Test it with the people it is for. Release it through the team’s normal process, then watch what happens.</p><p>Understanding the result matters. When AI changes something, we should be able to explain what changed and why. Otherwise, every fix becomes another guess.</p><p>We also still need to decide whether something deserves to be built. Faster implementation can easily turn a small idea into five extra features. The ability to add more makes focus even more valuable.</p><h3>The Role Is Getting Wider</h3><p>What interests me most is the possibility of staying closer to an idea as it becomes a real experience.</p><p>Designers can take responsibility for more of that journey, while working with developers on the decisions that make the product reliable and maintainable.</p><p>I do not think every designer needs to take the same path. Some will go deeper into code. Others will use these tools to explore and communicate ideas more clearly.</p><p>But the space between designing and building is becoming easier to enter.</p><blockquote><strong><em>And once we enter it, there is a useful question waiting for us:</em> how much of the experience are we ready to take care of?</strong></blockquote><p>This is a change I’m still thinking through.</p><p>As designers get closer to building, I wonder how our everyday work with developers will change. What can we take on ourselves, and where should we work more closely together?</p><p>I don’t have a complete answer yet. I’d like to explore this more as these tools become part of how we work.</p><p>If you’re trying this in your team, as a designer or a developer, I’d love to hear about your experience in the comments. What has become easier, and what still needs figuring out?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b685f152aaae" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When Building Became Easy, Interfaces Got Worse]]></title>
            <link>https://medium.com/@burak.sahin/when-building-became-easy-interfaces-got-worse-10bced674189?source=rss-ae83fa59e7ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/10bced674189</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Burak Şahin]]></dc:creator>
            <pubDate>Mon, 25 May 2026 10:01:12 GMT</pubDate>
            <atom:updated>2026-05-25T10:10:39.752Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*vJ98NJDfyI1LJOZcnjJP-w@2x.jpeg" /><figcaption>Overloaded software interface filled with features, dashboards, and AI tools competing for attention.</figcaption></figure><h2>AI Is Making Interfaces Worse</h2><p>Not because AI is bad.</p><p>Because building things became too easy.</p><p>A few years ago, adding something to a product was a decision that carried weight. A new dashboard card meant design work. A new filter meant development time. A new settings page meant maintenance and support costs. Every addition consumed resources, so teams naturally asked an important question:</p><blockquote>“Do we really need this?”</blockquote><p>Today, that question is asked less often.</p><p>With AI-assisted design and development, creating something new is incredibly fast. A designer can generate multiple concepts in minutes. A developer can build a new component in a fraction of the time it used to take. Product teams can prototype and test ideas almost instantly.</p><p>As a result, adding things to a product has become easier than ever. Unfortunately, removing things has not.</p><h3>Why Does Every Product Suddenly Feel Busy?</h3><p>I’ve noticed the same pattern in many products recently. A small analytics card appears because it might be useful. Then another one gets added. An AI summary sounds valuable, so it finds its place on the screen too. Suggested actions appear below it, a side panel arrives later, and eventually another shortcut is added somewhere else in the interface.</p><p>None of these decisions are wrong on their own. In fact, most of them make perfect sense during a product review meeting. The problem is that almost nothing gets removed. Everything stays, and after a few months the interface starts looking less like a carefully designed experience and more like a collection of good ideas stacked on top of each other.</p><blockquote>Good ideas rarely hurt a product on their own.</blockquote><blockquote>Too many good ideas often do.</blockquote><h3>The Feature Nobody Wanted to Remove</h3><p>Users experience something very different. They open the product with a simple goal in mind – <em>creating a form, reviewing a report, answering an email, or completing a task.</em> Instead of immediately understanding where to focus, they are greeted by charts, metrics, notifications, recommendations, assistants, shortcuts, and actions all competing for attention.</p><p>When everything on the screen asks for attention, prioritization shifts from the designer to the user. And that’s where the experience starts breaking down.</p><blockquote>When everything feels important, nothing really is.</blockquote><h3>AI Didn’t Change the Question</h3><p>AI has accelerated this problem because it removes many of the barriers that once forced teams to prioritize. In the past, implementing ten ideas was expensive, so teams carefully selected the two that mattered most. Today, implementing all ten is often possible, which means all ten frequently make it into the product.</p><p>The difficult part was never building features. It was deciding which features deserved a place in the experience. AI made the first challenge easier, but the second one remains exactly the same.</p><blockquote>Building became easier. Choosing didn’t.</blockquote><h3>Simplicity Is Still a Decision</h3><p>Ironically, AI could actually make interfaces simpler.</p><p>Instead of displaying every available action, it could surface the right action at the right moment. Instead of showing ten charts, it could explain the one insight that matters. Instead of exposing hundreds of settings, it could handle complexity behind the scenes.</p><h4>That future feels promising.</h4><p>Yet many products are moving in the opposite direction. Rather than replacing complexity, they are simply adding AI on top of existing complexity. The interface becomes larger, busier, and harder to understand.</p><p>The result may be more capable software, but not necessarily a better experience.</p><blockquote>Simplicity doesn’t happen automatically. Someone has to choose it.</blockquote><h3>The Return of Prioritization</h3><p>For years, software companies competed by shipping more features than their competitors. Today, almost everyone can build features quickly. That advantage is becoming less meaningful.</p><p>The next competitive advantage may not be speed. It may not even be AI itself.<strong> It may simply be clarity.</strong></p><p>The products that stand out will not necessarily be the ones with the most features. They will be the ones that help users focus on what matters and quietly remove everything else.</p><p>In a world where adding things is becoming easier every day, prioritization is no longer a constraint imposed by technology. It is becoming a design choice again.</p><p>Antoine de Saint-Exupéry once wrote:</p><blockquote>Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.</blockquote><p>For decades, technical limitations forced us to follow that principle. We removed things because we had to. Today, those limitations are disappearing. We can add almost anything we want.</p><p>The challenge is no longer building more. The challenge is knowing what to leave out.</p><blockquote>In a world where adding is almost free, removing becomes the real craft.</blockquote><p>This article is really just an observation I’ve been thinking about lately.</p><p>As AI continues to reduce the cost of building software, I wonder whether our biggest challenge is shifting from <em>how do we build this? to should we build this at all?</em></p><p>I don’t have a definitive answer, and I’d like to explore this topic in more depth in future articles. There are many angles that deserve a closer look.</p><p>If you have thoughts, examples, or a completely different perspective, I’d love to hear them in the comments. The best discussions usually start with an unfinished idea.</p><p>And this one still feels unfinished.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=10bced674189" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>