How I built five new UX trends into my own SaaS product

AI is changing what UX designers design. Here’s what that looked like when I put the ideas to work in a real product.

Grab a soda and a bag of popcorn, friends: this one is about where I think UX is heading in 2027, and what happened when I tried to build those ideas into my own product, Adapt to Success. 🎉

I’ve written before that AI won’t replace UX designers. Designers who understand how to design for AI, automation, trust and human behaviour will replace the ones who don’t. I still believe that. What’s changed is how concrete it has become.

UX has moved from websites to apps, from screens to whole ecosystems, from static interfaces to personal experiences. Now it’s moving from traditional interfaces to AI-driven products. The question is no longer “How do we design this screen?” It’s “What should the product do for the person using it?”

The old flow looked like this:

User → Interface → Action → Result

The new one looks like this:

User → Intent → AI → Action → Result

I’m the founder of AIKON Media, and I launched my first digital prodct about a month ago entitled Adapt to Success (ats.rip), a tool that rebuilds a CV so applicant tracking systems can read it.

Here are five UX trends I think will shape 2027, and what each one looks like inside a product I built myself.

1. AI-native UX: the intent matters more than the interface

Many products still treat AI as a feature bolted on top. In 2027, more products will be built around AI from the start, and the interaction model will matter more than the screens.

In Adapt to Success: the visitor states an intent and gets a result. You upload your CV, and if you want it targeted to a specific job, you paste the job ad text into the textarea.
That’s it.
There’s no account to create, no settings to tune and no dashboard to learn. Everything the product needs to know is in the two things you give it.

2. Designing for trust and control becomes a core skill

AI brings something traditional interfaces rarely had to handle: uncertainty. A calculator gives you an answer. An AI system can give you an answer that sounds right but isn’t. And AI agents can now act, not just answer. That changes UX in two ways: the design has to show where an answer comes from, and it has to keep the person in control of what happens next.

In Adapt to Success: you see the whole optimised CV on screen, every page, before you pay anything. Two buttons let you switch between your own original and the new version, so you can check every line against what you wrote. The promise is that no fact is added that isn’t in your original CV. My articles cite their sources, and the front page says plainly what the product can’t do: “We can’t promise the call. We can promise a CV the software can read and the recruiter can find.”

The AI never acts without you either. It rewrites, you look, you decide. Nothing is sent on your behalf, Adapt to Success doesn’t keep your CV, and you only pay if you want to keep the result. The decision stays with the person whose career it is.

3. Design the AI’s behaviour, not just the user’s

Traditional UX research is about people: who they are, what they want, what stops them. With AI, designers also have to design the system’s behaviour. What happens when the AI misunderstands? What happens when it gives a wrong answer that looks right?

This is the part of my product I’ve spent the most time on, because a CV tool that invents things can cost someone a job.

In Adapt to Success:

• Rules for what the AI may never do. It must not invent employers, titles, dates, numbers or skills. A small example of how specific those rules have to be: a course called “Practical Leadership” cannot turn into “experience in practical leadership”. The course is a fact. The experience is a claim the CV never made.
• A fixed test set. Sixteen test cases built on twelve fictional CVs in seven languages, plus one that mixes Danish and English, with and without job ads, some with traps taken straight from those rules. Every answer is checked automatically for new numbers, new names, words copied from the job ad text and a CV that suddenly changes language. It runs before any change to the prompt or the model.
• A weekly watch. Every Monday an automated check scans the market for new AI models, tests the relevant ones against the same CVs and tells me whether anything should change. The product doesn’t get a new model because it’s new. It gets one only if it keeps the promise better.

4. Less mental effort as a competitive edge

The more intelligent interfaces become, the more it matters to understand human attention. AI can generate thousands of possible experiences, but people still have limited attention. A product that can do everything soon becomes a product nobody chooses, because it’s overwhelming.

In Adapt to Success: one task, three steps, no subscription. Upload your CV, preview the optimised version for free, save it as a Word file if you want it. You pay once, and only if you chose to save (download) it.

Just as important is what I chose not to build. Personalised interfaces and adaptive layouts are on every trend list for 2027, and they’re good ideas for many products. For a tool with one job, they would have made a simple product complicated. Sometimes the best UX decision is the feature you leave out.

5. Continuous insight instead of big research projects

AI already helps researchers analyse interviews, surveys, reviews and support cases. Research is moving from a big project every few months to a steady stream of insight.

In Adapt to Success: a cookie-free counter shows me each day where people drop off, from the front page to upload, preview, checkout and payment, with no personal data. And when someone goes to checkout and comes back without paying, the product asks one question with three buttons: was it the price, the result or a problem with the payment?

But patterns aren’t reasons. The counter tells me where people stop. Only people can tell me why. That’s still the researcher’s job, and in my case it’s mine: I answer every message myself.

The biggest UX trend in 2027 isn’t AI

Ironically, I don’t think the biggest UX trend in 2027 will be AI. I think it’ll be a turn back to human-centred thinking.

When technology becomes this powerful, the temptation is to use everything it can do. But users don’t care how advanced the technology is. They ask one thing: “Can you help me get done what I need to get done?”

AI can generate interfaces, build prototypes, analyse research, write code and draw user flows. Designers still have to decide what should exist, why it should exist, who stays in control, when the system should act and when the human should.

And most important of all: how do we make technology more human?

Try it with your own CV at ats.rip: you see the whole result before you pay. More of my work is at AIKON Media.
Feel free to follow me on LinkedIn, and if you found this article just a wee helpfull, please give it a clap or twenty: claps help the article reach other readers.😁

How I built five new UX trends into my own SaaS product was originally published in Bootcamp on Medium, where people are continuing the conversation by highlighting and responding to this story.