From Keywords to Concept Discovery: How AI Changed Search Before SEO Noticed
Years ago, I sat in an in-person meeting between a client of mine and a large SEO agency. It did not go well.
My client insisted that his core service—what actually made his business work—was winter vestibules. The agency pushed back hard. Their CEO essentially said, “Nobody searches for that. If I’m wrong, I’ll give you the keys to my company.”
Voices were raised. Tempers flared. Everyone left frustrated.
At the time, the agency wasn’t wrong—at least not by traditional SEO standards. There was little to no keyword volume. No obvious data to optimize around. From a national SEO perspective, “winter vestibules” barely registered.
But years later, I realized something important.
They weren’t having an SEO argument. They were having a language and geography argument.
When the “Experts” Leave and You Have to Learn SEO the Hard Way
We ended up working with that agency for about a month before dropping them. My client still wanted SEO—but he wanted it done by someone who understood his business and respected the fact that he was a leader in his industry.
So I had to learn SEO the hard way: trial, error, testing, and listening closely to the client.
He wasn’t guessing. He knew his market. He knew the buildings. He knew the buyers. And he knew that winter vestibules were a real, established solution in New York City—even if national keyword tools didn’t reflect it.
We built a site around that exact term.
In my client’s words, the site “looked like a doodle on a bar napkin.” He wasn’t wrong—I was new to design at the time.
But here’s the thing.
It ranked. It converted. And it still performs well.
Because it spoke directly to the people who already understood the problem—and gave language to the ones who didn’t.
This Wasn’t a Keyword Problem — It Was a Market Problem
Winter vestibules aren’t new. In New York City, they’re common. Many older buildings and restaurants were built without enclosed lobbies, so every time a door opens in winter, freezing air whips straight through the space.
Inside that market, the pain is obvious. The solution is known. The terminology, however, is fragmented.
Some people say “temporary vestibule.” Others say “winter enclosure.” Many just describe the experience: cold air blasting into the restaurant every time the door opens.
From the outside, this looks like low demand.
From the inside, it’s high demand with fractured language.
Traditional SEO only knows how to chase terms that already exist at scale.
I Saw the Same Pattern Again — This Time in Medicine
Years later, I ran into the same problem personally.
I had a very specific set of GI symptoms: upper-left abdominal pain, pressure, and audible internal noises. Google was useless. No matter how I phrased it, there were no meaningful results.
Eventually, I bypassed Google entirely and searched medical journals. That’s where I found a term I had never seen surface in search results before: splenic flexure.
Once I had the right noun, everything changed.
Research papers made sense. Doctors took the symptoms seriously. Online discussions aligned. Search actually worked.
The problem wasn’t a lack of information.
It was a lack of language discovery.
From Search Optimization to Concept Discovery
Traditional SEO assumes something critical: that people already know what to search for.
That assumption no longer holds.
People experience situations, not keywords.
They say things like “cold air hits customers every time the door opens,” “my roof keeps leaking near the chimney,” or “I hear loud noises in my upper left abdomen.”
They don’t know the correct term until someone gives it to them.
Generative AI changes this.
AI doesn’t just return pages. It maps experience to concept to language.
Once humans have the right words, search demand follows.
How This Applies Beyond Niche Industries (Like Roofing)
This isn’t just a niche-industry problem.
Take roofing as an example.
Most people don’t search for “step flashing failure” or “ice dam formation due to insulation gaps.” They search for water stains on the ceiling, leaks near the chimney, or ice forming at the roof edge.
AEO-driven content doesn’t just optimize for roofing terms. It connects symptoms to causes, in plain language, tied to geography and conditions.
This allows AI systems to understand the problem, suggest the right solution, and surface the right business—even when the customer never uses industry terminology.
The New Optimization Framework: AEO, GEO, and AIO
AEO — Answer Engine Optimization
AEO isn’t about ranking pages. It’s about making answers extractable and trustworthy.
Clear summaries. One concept per page. Plain language. No filler.
If AI can’t confidently summarize it, it won’t surface it.
GEO — Generative Engine Optimization
GEO focuses on retrievability, especially for regional and situational demand.
It asks whether AI can map lived experience to the correct concept, whether geography is explicit, and whether the entity is unambiguous.
AIO — AI Input Optimization
AIO improves the inputs that shape AI outputs.
Clean terminology. Structured content. Reduced ambiguity.
You’re not training models—you’re improving the ecosystem they retrieve from.
How to Design Content for This New Era
Stop asking, “What keywords should we target?”
Start asking, “What problem does someone experience before they know what to call it?”
Then design content that names the problem clearly, introduces the correct concept, anchors it to context and location, validates the experience, and guides the next step.
This is why hub-based content works. This is why answer-first summaries matter. This is why bloated SEO pages fail.
The Big Insight
People don’t lack problems. They lack the language to search for solutions.
Generative search doesn’t replace SEO. It reveals where SEO was incomplete.
The future isn’t about chasing keywords. It’s about helping people name what they’re experiencing.
Once that happens, everything downstream—search, AI answers, and conversion—starts working again.
Why This Matters Now
Whether you serve niche regional markets, specialized industries, or broad service categories, this shift affects you.
If your customers describe their problem better than your website does, AI will route around you.
The businesses that win will understand the problem before the keyword, design for discovery rather than volume, and think in systems instead of tactics.
That’s the work we focus on at Open Space Digital.
