Meet Alex. A busy professional and frequent traveler, Alex has been relying on the same battered suitcase for years. As he pulls it out of storage to prepare for an upcoming trip, the handle creaks ominously, and one corner seems to have collapsed entirely. “I think this thing’s finally giving up,” Alex says aloud, half to himself, half to his ever-present AI assistant, Eko.

Eko chimes in immediately. “It might be time for a replacement. Would you like me to find one for your trip?”

Alex nods. “Yeah, good idea. Something lightweight but durable. Big enough for the essentials without being oversized.”

“I see you’re traveling to Reykjavik next Friday for five days,” Eko replies, pulling the information from Alex’s calendar. “What are you thinking of packing? Layers for the cold? Hiking boots? Anything bulky I should account for?”

Alex thinks for a moment. “Yeah, a couple of sweaters, some waterproofs, and hiking gear. Maybe room for a tripod too.”

“Got it. A carry-on size might be too small for that. Let me check medium-sized options that meet airline size requirements.”

Within seconds, Eko says, “I’ve found three options that meet your needs. Would you like to see them?” Alex’s phone buzzes, and he lifts it to see the choices projected side by side in augmented reality, scaled to size and rendered in crisp detail. “You can swipe to cycle through available colours and styles,” Eko explains. “This style in black won’t arrive in time for your trip, but the grey and blue will.”

Alex swipes through the options, inspecting the details. He quickly settles on the grey recycled-material suitcase. “Let’s go with this one.”

“Good choice,” Eko replies. “Would you like to add an extended warranty for £20? Your house insurance already covers accidental damage and theft, so it might not be necessary.”

“No, skip the warranty,” Alex decides.

“I’ll select the standard warranty for you,” Eko confirms. “Your new suitcase will arrive tomorrow evening.”

The next day, the suitcase is delivered on time. Eko notifies Alex and offers a quick rundown of its features. “You can initialise the lock combination by setting the dials to 000 and pressing the reset button. Would you like me to remember the combination?”

“Yes, make it 258,” Alex replies, fiddling with the lock.

“Done. I’ve just registered your product with the manufacturer, so you’re good to go.”

A week later, returning from his trip, Alex notices one of the suitcase’s wheels has developed a squeak and feels loose. He holds his phone close so Eko can see it and says, “Eko, can this be fixed?”

“I’ll handle it,” Eko responds. Moments later, it updates him. “The manufacturer is sending a replacement wheel assembly. It should arrive in two days. It’ll be easy to replace - I’ll walk you through the process when you’re ready.”

As Alex unpacks, Eko casually asks, “So what do you think of your new suitcase? I’m sure the manufacturer will be interested to hear.”

Alex considers for a moment and says, “Umm, yes it did well. I’m glad I got it. It has a quality feel to it - aside from the problematic wheel of course. I really like the multiple small compartments - great for toiletries and various bits. I guess the zip on the main compartment did snag once or twice - it didn’t bother me too much to be honest.”

Eko replies, “I’ll let them know. By the way, the manufacturer recommends wiping the case down with a dry cloth after every trip and storing it in the microfibre bag it shipped with. Would you like me to send you the full care instructions?”

Alex smiles. “No need. That’s simple enough. Thanks, Eko.”

Analysis

This scenario may feel futuristic, but the technology to make it real is already here or imminent. Large language models with web search, multi-modal communication, agentic AI, augmented reality, and APIs are paving the way for personal AI assistants like Eko.

This has profound implications for the traditional practice of designing, developing, and delivering products and services. For decades, we’ve honed the craft of creating interfaces to engage users—crafting compelling websites, useful apps, and intuitive customer journeys. But as personal AI assistants emerge they’ll be able to bypass the elaborate brand and information environments we’ve painstakingly built and served up.

In a world where every potential customer’s personal AI assistant evaluates all products based on granular data, the primary path to brand success is by delivering on user needs and expectations. Once a product manufacturer or service provider can connect (considerately, with permission!) to each of their customers’ personal AI assistants, they will be able to analyse and improve every aspect of their customer experience in full. A richly detailed and fully representative body of insight around consumer intent, purchase and usage will have a profound and disruptive effect on market forces.

An AI’s consideration set can be the entire market, and will prioritise the attributes (from the manufacturer’s API) and qualities (from trusted feedback mechanisms) that will meet requirements and satisfy an individual’s preferences. Personal AI assistants can therefore be purposefully blind to marketing effort—resisting attempts to manipulate via cognitive biases and emotional appeal. Mass advertising campaigns and mindshare-building efforts will fade as a brand realises the power of engaging its customers (actual and prospective) in a 1:1 conversation 24/7.

Once people start to delegate tasks to their AI assistants, the need for our current multiverse of human-to-system interfaces—websites, apps, support centres and dashboards—will decline sharply. This will manifest as a fundamental pivot from brands owning and serving their own interface(s) to enabling their customers’ primary interface: their personal AI assistant. Brands which offer robust, well-documented APIs that communicate their offering and open up their processes to agentic AI will thrive.

Consider all the thousands of interaction patterns we’ve developed over time, the design systems, front-end frameworks, information architecture, brand systems and so forth. What happens when the user can describe what they need and a preferred way to consume the information which informs their decision? One user with accessibility needs might benefit from a full voice modality. Another might require a purely visual modality. Yet another might want suitcase options described in a way that speaks to them. Suitcase 2 is like a Ford Focus - the sensible, family-friendly choice that won’t break the bank but probably won’t excite.

This aspect of the future teases a hyper-personalisation of digital experience: people will pull what they need from the world and consume it in a distilled form that suits them. Our current practices of endlessly rebuilding and serving up one-size-tries-to-fit-all interfaces, informed by broad segmentation of needs and requiring people to make the effort to visit could be facing extinction.

Thinking points