Startup Corner #14: Search is dying, but discovery isn’t
AI isn’t just changing how people search — it’s changing how they decide. That’s a massive opportunity.
Hey friends,
If you’ve searched for a product recently and ended up talking to a chatbot instead of scrolling through blue links… you’re not alone.
Nearly 60% of online shoppers now say they’re using AI chatbots instead of search engines to get product advice (Source).
And it’s not a niche behavior — we’re talking 1,300% increases in AI-driven site visits during peak shopping periods.
In travel, that number’s jumped even higher — some companies have seen 1,700% spikes in AI-attributed traffic.
This isn’t just about replacing Google. It’s about replacing the entire behavior of how people discover, compare, and decide what to buy.
And that shift? It has huge implications for startups.
So let’s unpack what’s really happening — and where the opportunity lives now.
For the last 20 years, we’ve lived in a world where search equals discovery.
If you wanted to find a product, a tool, a review, or a comparison — you Googled it.
Businesses competed for keywords, backlinks, schema tags. SEO was a growth function.
But with generative AI, the model’s flipping.
People are now asking,
“What’s the best carry-on suitcase for international travel under $300 that fits in Delta overhead bins and doesn’t look like it’s from 2010?”
And instead of digging through forums and listicles, they’re just getting a straight-up answer from a chatbot — or an LLM-powered assistant baked into Amazon, Google, or somewhere in between.
That answer?
It’s becoming the new click.
So what does that mean for startups?
First, it changes how products get found.
When language models become the front door to shopping, you’re no longer competing for search position — you’re competing for model memory.
In other words: it’s not about being the top link. It’s about being the most trusted, structured, and referable answer.
That means:
Clean, structured product data
Verified reviews from high-signal platforms
Consistent mention in communities where people talk in natural language (Reddit, TikTok, Discord, Quora — not blog spam)
And ideally, being integrated directly into the knowledge layer that AI is pulling from
In a way, SEO becomes a trust game — not just a technical game. The question becomes: how do I make my product the one the AI wants to recommend?
Second, this shift compresses the buying journey — which is both an opportunity and a threat.
When an AI assistant shortens the funnel, people move from curiosity to decision faster.
That’s great if you’re in the top recommendation.
But brutal if you’re second or third, or not listed at all.
Startups that rely on long-tail discovery — think affiliate content sites, niche comparison tools, even some DTC funnels — are going to feel this.
There’s less room to be “found” mid-funnel if the funnel is collapsing into a conversation.
But there’s also upside here.
If you’re building tools that help buyers make high-context decisions — product advisors, transparent aggregators, verified review platforms, even AI wrappers on existing marketplaces — you’re suddenly more relevant than ever.
Because people still want to make smart, personal, confident choices.
They’re just not clicking through five tabs to get there anymore.
They’re asking a question and expecting a complete answer.
Startups that can surface structured truth fast — especially for complex or emotional purchases (think health, education, finance, travel) — will have a real edge.
Third — and this might be the biggest unlock — the interface is shifting from page to conversation.
And that means the old rules for UI, landing pages, even branding, don’t apply in quite the same way.
For example: when someone talks to Amazon’s AI assistant Rufus, they’re not “browsing” anymore — they’re interviewing the store.
If your product can’t speak clearly through that assistant — if your content isn’t AI-readable, your reviews aren’t verifiable, your data isn’t structured — you’re invisible.
The same applies to Google’s Gemini integrations, which now pull from a 45-billion-product graph to generate personalized shopping advice, deal tracking, and recommendations — all without requiring the user to bounce between 10 sites.
So if you're building a startup that depends on being seen — and let’s be honest, that’s all of us — you need to think about how your product shows up in conversations.
This might mean:
Publishing clean, structured product data that’s easily digestible
Creating conversational content that aligns with how people ask questions, not just how they search
Building integrations that get you inside the platforms users are already chatting with
And long-term? Maybe even training your own LLM on your product ecosystem so you can be the first party in the discovery chain, not just a passive listing.
Bottom line: the way people shop, search, and decide is being rebuilt — not from scratch, but from conversation.
And if you’re building a startup right now, especially in ecommerce, discovery, AI tooling, content infrastructure, or product data — this shift is your map.
Because even if search is fading, discovery isn’t going anywhere.
People still want help deciding.
The job now is to show up as the most helpful voice in the room — even if that room is an AI prompt box.
Until next time,
— RB
Startup Corner