Why we’re trading 17 opens tabs for one good conversation

If you’ve ever lost an evening toggling between look-alike air purifiers wondering why one costs so much more than the other, you know the truth: online shopping is… broken. We don’t call it that, of course. We call it research.

But with the recent flood of shopping assistants and agentic browsers, we’ve effectively been shown the light. You mean to tell me I can simply describe what I’m looking for—a handheld vacuum for a studio apartment with a shedding dog, or a gift for a dad whose only hobby is “silence”—and get back a tailored buying guide generated through conversation? It’s enough to make you wonder why we settled for keyword searches for so long.

And so, somewhere between the comparison charts that require at least surface-level context to decipher and the "spec sheet archaeology" of digging through technical details, the humble product page has begun to lose its shine. It isn't broken per se, it’s just built for a slower world.

The centre of gravity has shifted from browsing to dialogue. From ‘Here’s everything we think you should know’ to ‘Let’s figure out what actually matters to you’. And that shift—from multiple choice to conversation, is what’s slowly dismantling the product page as we know it.

Static pages in a dynamic world

Despite all the personalisation hype of the last decade, the product page remains a relatively static, unexciting entity. It assumes that a casual browser and a desperate buyer want the same information, arranged in the same sequence.

But human decision-making is messy. I might not care about durability in general, but if I have a flight scheduled tomorrow, suddenly all I care about is shipping speed.

Traditional pages don't pivot; they just stare back at you with their bullet points. AI agents, however, thrive on this messiness. Instead of forcing you to scroll past the chip specs to find the battery life, an agent starts with the basics: What are you trying to do?

It interprets your intent and serves up the trade-offs. It says, ‘This one is charges faster, but users complain it heats up’, or ‘This one is overkill unless you’re planning to game a lot’. It feels less like reading a manual and more like asking a knowledgeable friend who just so happens to have memorised the entire internet.

Good luck trying to replicate that via a traditional product page. You’d have to scroll through an endless sea of user-generated content to find the one guy who used the product the way you intend to.

Preference stewardship, not personalisation theatre

For years, personalisation in many contexts was aggressive guessing based on context clues. Click on a stroller once and the internet has decided you are a professional stroller collector and has, subsequently, showed you ads for prams until your child was in pre-primary.

Agents replace this with something I like to call ‘preference stewardship’.

Agents treat preferences as fluid, not immutable. You can say, ‘I want running shoes,’ then, moments later, ‘Actually, keep it under INR 12,000’. You could also add, ‘I’ll pay more if my knees won’t hurt’. The system adapts to each prompt without having a spasm. It anchors itself in what matters now, not what you clicked on three weeks ago by accident.

We actually learned the value of this years ago, way before generative AI hit the scene. In the mobile marketing world, we found that intelligence isn't always about information; often, it's just about context and timing.

One of our clients, a financial services brand, generated over a million dollars in monthly revenue simply by tweaking when they spoke to users. They moved from brute-force blasts to triggering reminders at the precise window a user naturally tended to pay bills. The “intelligence” there was just good timing.

Agents are a high-level evolution of that principle. They don't just know when to speak; they know what to say based on the immediate context of the conversation.

The trust contract gets rewritten

Here’s the catch, though: the moment an AI starts filtering reality for you, trust becomes the only currency that matters.

When a product page is confusing, you blame the UX designer and feel like they could’ve done better. Simple enough. When an AI agent misleads you, you feel betrayed and misled. If an agent recommends a product because it has a higher margin, or hallucinates a feature that doesn't exist, you won't just close the tab — you’ll burn down the bridge.

Transparency is the new UX. Agents will need to show their work:

• “I’m recommending this because you said budget was priority #1.”
• “This is the best-rated option, though it lacks the waterproofing you asked for.”

We’ve already seen how easily this “consent capital” is lost. We learned it with notification fatigue and creepy ad retargeting.

If agents feel like sales bots rather than unbiased assistants, the revolution ends before it begins. The winners will be the ones who treat this intimacy as a privilege, not a resource to mine.

Commerce without the catalogue

Am I sounding the death bell for the product page? Not quite.

We will always need a place to check the fine print. But the product page is accepting a demotion. From being the front door to all traffic, it will become a quieter reference point—something you visit when you need to double-check the details. And this shift solves the problems we’ve been trying to fix with UI tweaks for years.

Decision fatigue will drop because trade-offs are explained, not hidden in correlations across tabs. Journeys will adapt mid-stream, just like real conversations. And, most importantly, signals are volunteered, not guessed. You won't need to track my cookies if I just tell you what I want here.

The static page won’t vanish overnight, but the real decision-making—the human kind—is moving elsewhere.

Because when it comes to parting with our money, most of us don’t actually want “all the information.” We want confidence. We want to know we aren't making a mistake. And when AI agents behave responsibly, that is exactly what they deliver: the confidence to close the tabs and finally click ‘Buy’.

-Jacob Joseph, vice-president – data science, CleverTap