#27. The magic ingredient for AI-powered experience design

 

This newsletter belongs to “The Sunday Tales Summer Edition”.
These summer letters are a bit lighter, a bit slower, just like the season.
But they’re also an opportunity: a chance to treat these months as an open campus, to learn something new, reflect on my process, and find better ways to do the work I care about.


It seems that all companies are AI companies now. Like a circus performer, AI is rolled out for various tasks, in the ultimate hope of making the user's life easier. But how often is that really the case?

I think of this as the “FOMO" syndrome of AI. AI is now part of every app I use, yet there doesn’t seem to have been much thought about where and when the user will actually need it. To me, it feels like tech companies are panicking: if they don’t incorporate AI into their apps, they will be left behind.

There are heavy environmental costs associated with this technology. This, as well as the arbitrary way in which AI is used, has made me question whether there might be better, kinder ways for designers to assist these companies. I asked myself: how can designers help incorporate AI in ways that feel relevant, so its use doesn’t unnecessarily deplete our precious natural resources?

Here's my answer. When designing experiences with AI, context is everything. In other words, it’s important to think about what real and specific differences this technology can make, while consuming energy as sparingly as possible.

 

Context = my car distributor!

Granted, it’s a little old, but lately my little car has spent more time with the mechanic than on the road. In summer it acts up more than usual, possibly due to the hot weather.

The mechanic explained that, this time, my car had a malfunctioning distributor. The distributor controls when and where the spark happens, so that the engine can start. So, when it’s not working (like now!) the cylinders aren’t sparked at the right time, or in the right order, when the engine turns.

Scene from the mechanic. Some of you may know that for the past 10 years, I have been driving a cute Mini Cooper from 1977.

Scene from the mechanic. Some of you may know that for the past 10 years, I have been driving a cute Mini Cooper from 1977.

 

The way I see it, that distributor is the context element of my car. It provides the “where and when” of my driving experience.

The where and when are essential elements of storytelling, because they set the story’s rhythm. Specifically, “where" – or pace – focuses on the places where content is revealed in the narrative. This keeps the audience engaged, without feeling overwhelmed or bored.

“When" is the sequence, or the order in which events or information are presented in the story. This helps to build understanding, anticipation, or emotional impact.

Together, pace and sequence determine how the story flows, and how effectively the message resonates with the audience.

For example, context builds flow, like it does in the recent adaptation of the cartoon Lilo & Stitch. While Stitch is seen earlier in the movie, landing on Earth like a comet, he doesn't encounter Lilo immediately. Later, at the dog shelter, a lonely Lilo mistakes him for a dog and adopts him. This encounter is the inciting incident (the spark) that allows the story to move forward, with more hilarious scenes ensuing between the two characters.

Have you watched Lilo & Stitch, the movie? For me, it’s not as funny as the cartoon, but it still provides plenty of laughs!

 

Context in the time of AI

If we put context at the forefront of our design thinking, we may then see that some features are completely out of place in the user experience.

The introduction of AI has made context even more important. While AI seems like an essential tool, its context is equally important as any other feature to be added to any product.

Considering that AI can be more of a complicated nuisance than a help, I think we must now consider “context" more deeply and accurately.

To explain further, here are some examples of AI in action on WhatsApp and Slack, two messaging apps with different audiences and uses.

 

WhatsApp: what's that ring over there?

Part of Meta, WhatsApp has lately changed its policy to add many subtle functionalities, many of which didn't disrupt the user experience. That was until the introduction of Meta AI.

Because Meta AI is a major part of Meta’s other apps, it has also been added to WhatsApp, for no discernible reason.

 
Hello Meta Ai. Meta AI was unleashed a few months ago, creating lots of disorientation between WhatsApp users.

Hello Meta AI. Meta AI was unleashed a few months ago, creating lots of disorientation between WhatsApp users.

 

The amount of wonder this little button has created is obvious on the internet, where you can find many videos that instruct you how to deactivate this “useless" tool.

Considering that WhatsApp is a simple messaging app, used by families and among friends to communicate quickly and conveniently, the Meta AI companion isn’t just inconvenient, but annoying. It’s also more than a little worrying for some people. (Is this thing reading my messages?" a panicked friend asked me!).

For many users, it feels supremely out of place, leading many to activate it by mistake (fat finger error, or accidental clicks in UX terms).

In a WhatsApp context, Meta AI is like that annoying and unwanted character in a soap opera who once appears, never goes away. Despite Meta’s description of the service as “optional,” the AI button appears automatically and cannot be turned off or removed, though you can choose not to interact with it (unless, of course, you click on it by mistake!).

 

Slack: keep me up with team news!

The messaging app Slack has a very specific audience of companies and large teams working together.

Their AI companion was released in February 2024. From June 2025 onwards, Slack made thread and conversation AI summaries available to all users on paid plans, eliminating the need for a separate AI add-on subscription. For this reason, you may start to see it here and there in the app.

Slack understands that teams may text different messages at different times of the day, from different timezones, between different channels. Some users may also need to keep up with messages on their return from holidays.

Threads (nested messages comprising questions-and-answers between members on similar topics) are a useful feature that I often use to avoid long channel-talks and keep everything nicely organised by topic.

It’s in one of these threads that I noticed Slack’s AI companion had smartly provided a helpful resume. I thought this was the right context, as from my own experience, some threads can become long and cumbersome, meaning a quick recap is needed.

 

Welcome Slack AI. Slack AI is a new feature that, even though it was released in February 2024, is now part of every plan.

 

I am sure that Slack has added other AI features elsewhere. However, I found this specific integration to be timely, well-integrated, and most of all useful.

 

Final thought: add an "AI spark" to your story with thoughtful context!

Undoubtedly, there is a strong tendency to infuse AI into everything we design, like the latest ‘must-have’ kitchen gadget. But AI is an expensive tool, both for the user (as they ultimately pay for it) and the environment.

While we can't avoid embracing it at least in part, I would like us designers to be more like storytellers, consciously adding to their story in a way that helps the hero move forward in their journey.

I think a better and smarter way to design AI integration is to start by focusing on “context", so its addition doesn't feel awkward or prevent the user from finding their “spark" in the experience.

So next time a AI-powered feature is added to your product, ask yourself:

  • Is it useful in this specific part of the experience (when)?

  • Could it be introduced in a different setting (where)?

  • Can AI be sprinkled here and there in the app for a more customised experience, so it doesn't feel omnipresent?

My advice? Don't underestimate this way of thinking. While AI can do many things, the ultimate question must centre on how we wish to frame the user experience, while taking the best care of our dear planet… one spark at a time.

 

Want to read the first chapters of my book – for free?

I’m currently writing a book about storytelling for designers. It’s been years in the making, and I’m finally ready to share it, one chapter at a time.

Join my newsletter The Sunday Tales to become one of the first readers. You’ll get exclusive access to the early drafts, plus the chance to help shape the book through your thoughts and feedback.

Previous
Previous

#28. Design is not important until it becomes essential

Next
Next

#26. Purpose vs. Resolution: Designing for depth (and singing for my soul)