Is Generative UI Helping or Hurting Your Customer Experience?

Generative UI promises every customer a personalized, dynamically-generated interface — but is it actually improving the experience, or quietly breaking it?

In this episode, Hashi cuts through the AI hype to explain what Generative UI (GenUI) really is, why companies are racing to adopt it, and where it goes wrong. From e-commerce checkouts to the Marriott app, he explores the real promise of AI-personalized interfaces — and the "unwritten contract" between a user and a button that too many companies are breaking.

You'll learn:

• What Generative UI is, in plain business language
• Why personalization drives conversion — and why habit beats personalization
• What a 2026 Qualtrics report reveals about the personalization-privacy tradeoff
• The #1 mistake companies make when deploying GenUI
• A practical framework for rolling out Generative UI safely: guardrails, data readiness, and controlled experiments

If you're a non-technical business leader trying to figure out where AI actually fits in your customer experience strategy, this one's for you.

Follow Hashi on Instagram: @aimehashi

Transcript

SPEAKER_00 0:04

Hey guys, I'm Hashi, and I make sense of AI for non-technical business leaders. What does that mean? There's a lot of AI jargon, a lot of AI hype out there. And it's really difficult if you're not technical, but you are a business leader trying to drive improvements and benefits to your company, and you just don't know where to start. And that's where I help. Today I'd like to talk about whether generative UI is helping or hurting the customer experience. Let's first define what generative UI actually is. Within the last couple of years, this has been a term that has come around and become a bit more popular, and thanks to Google as well. Generative UI or Gen UI is a user interface that is dynamically generated in real time by artificial intelligence. And what it's supposed to do is provide each user with a tailored, highly customized experience and not a one size fits all, like we're used to. The original promise with Gen UI was that everybody, regardless of where you came from or who you were, you would get a personalized interface. That means that whether you have a preference for large images, you prefer text, you prefer numbers, you prefer graphs, you would be able to get served a very personalized experience on an application or a website that was using Gen UI. And why companies rushed to do this was well, personalization is the name of the game. The more personalized you can create an experience, the more likely somebody's going to buy or return. At least that was the promise. If we look at e-commerce sites, for example, and we look at the application of generative UI there, the promise was we'd be able to get customers to check out a lot faster than if it wasn't using generative UI. Why? Because, well, who doesn't love stuff that's personalized to you? When you walk into a hotel and they know you by name or they call you by name, aren't you more likely to return to that hotel or even extend your stay because they treat you more personally? Same promise. If we are able to cater the website and move things around to adapt to the way that you like to look at a website, the thesis there was you get to check out faster. But the caveat there is we are also creatures of habit. We grew up with the close window button at the top right, with the save over there on the top left, the X to close out a pop-up, things like that. So the problem with generative UI is it starts to adapt and shift all those things on the page, thus it not being able to serve our habits. So let's take a look at another example of how Gen UI can be very beneficial. Let's imagine you love to stay at Marriott. You're a Marriott loyalty member and you use the Marriott app quite a lot. Now imagine you went to their profile page and they had a lot of questions about your behaviors, your preferences, which they do have some, but let's just say it was a lot more intensive. How you like to read, what your preferences with room temperature is, all of that stuff, and they they learn a lot of information about you. What you can do with GenUI then is the next time you open up the app, the app can prioritize what's most important to you in that visual representation. It can also prioritize important pieces of information. Maybe you're somebody who cares about the weather and a little less about the cost. So what the Marriott app could do is say, hey, there's great weather in Miami. I'm not even going to show you the cost. Would you like to come stay there? That's something that Gen UI can do. And granted, you don't necessarily need Gen UI to learn behaviors and follow an algorithm. However, doing it at scale allows us to personalize those experiences across millions of users across the same app. So as much as I love the promise of generative UI and how it can help provide really personalized experiences, I have a little bit of skepticism with me. First of all, while you can change pretty much everything, going back to my earlier point, that humans are essentially creatures of habit. We've been conditioned over decades on standard user experiences, where to look for certain buttons. Now, when you start to move and shift that layer dynamically through generative UI, all of a sudden an action that you were, you know, with muscle memory just able to go to the right and click on something is no longer there. I mean, do you remember when we started to move and shift menus on websites? That was a big problem. Right now, a lot of websites use standard menus. What happens when you go to a funky website that's moved the menu, I don't know, to the left and it's stacked vertically? You get confused. Same thing. It takes you a while to figure out your way and how to navigate. So this is where I'm a little hesitant on the promise of generative UI without guardrails around it. The same challenges with generative UI are going to exist regardless of the platform you're on. Whether you're on a website, whether you're on a mobile phone, app or some other SaaS platform, it's about the same. You know, we have been conditioned over many years, many decades to look for things in the same spot. Regardless of whether you're on a mobile app or whether you're on a website, if buttons and actions start to move around, it's going to take people a lot longer to learn how to interact with those applications. And that will lead to frustration. Frustration leads to people leaving, which is the exact opposite of what generative UI is trying to achieve. In a demo space, generative AI looks really great. But that's because you're controlling all the variables. Every demo, as we all know, works really well. But when rubber meets the road, there are so many more variables that are out there, different people's behaviors, patterns, needs, etc. And you just don't know until you get out to the wild. And that's where demos fall apart from reality. So the one principle that generative UI needs to maintain is the unwritten contract between a user and a button. Know where to expect it and know what it does. If you break that contract, you break the experience. In 2026, Qualtrics released a consumer trends report that actually said 64% of consumers say they prefer personalized experiences. Only 39% of that believe the benefits are worth what they're giving up. What does this mean? It means that people felt like they were giving up too much of their information, maybe too much privacy, to get a experience that was personalized but not amazing. And this is something we need to really think about before we dive head and feet and everything into generative UI. Can generative UI deliver on the promise? So the biggest single mistake I see companies making or attempting to make with generative UI is not governing it or not putting the right guardrails around it and just saying, hey, let's create a user experience that is totally generative UI and that disrupts everything, which, going back to my last point, you're breaking the contract, you're not catering to humans who are creatures of habit, and then you start to really disrupt their world, and you end up having the exact opposite effect. So what do you do? You actually should release generative UI in controlled spaces and within a controlled set of user experience guidelines. So don't change everything on the page, only change certain sections or certain components to drive personalization. See how that does. If it does well, then maybe give it a little bit more freedom. But don't just give it carte blanche. Before I would have a team start to work on anything related to generative UI, even though the promise sounds great, is I'd look and see if there's actually data that warrants that. Do I have enough data to make this actually useful for my end consumers? Or am I gathering enough data to understand the patterns and behaviors? And then I would figure out where I can have the most impact based on where there are critical drops in user experience. So if I'm tracking things like rage clicks or band-carts, you know, anything along the user journey, I'd want to study the data first. Because you know what? Good data, good AI. I want to leave you with this one final thought. Generative UI can be really powerful. But if you were a shop owner, you wouldn't rearrange your shelves when your customers are already in their shopping, right? You would rearrange them at some point, maybe overnight, before the shop opened, to try to make things faster and more efficient for your customers. And that's how you should approach generative UI in any of your user interfaces. If you like this content and you want to hear more from me, like, subscribe, and follow me on Instagram at AI Me Hushi. Thanks for watching, and I'll see you next time.