Intent is what actually killed search engines.

For thirty years, search engines and the SEO industry were built around one assumption: the customer knows what they want, types it into a box, and the engine tries to match. That assumption is wrong. It was always wrong. Customers rarely know exactly what they want — they know the symptom of what they want. AI search is the first system in history that can actually figure out what someone needs by asking. That single capability is why traditional search is finished. Not declining. Finished.

The quiet failure of keywords

Think about how you've used Google for the last twenty years. You had a problem. You translated that problem into a few words. You typed those words into a box. Google handed you back ten links — most of which were trying to sell you something tangentially related to the words you typed. You clicked one, scrolled, hit the back button, clicked another, eventually found something close enough, and moved on.

The entire SEO industry was built around helping businesses appear higher in those ten links. Keyword research. Ranking strategies. Content optimization. Backlink building. All of it was effort to insert a business into the answer to a search someone almost meant to make.

"Best plumber in Denver" is the canonical example. Every plumber in Denver, every SEO agency in the country, has spent the last two decades trying to rank for that phrase. But think about what that phrase actually tells you about the person who typed it.

Almost nothing.

You don't know if they have a clogged drain or a burst pipe. You don't know if they want emergency service or a planned remodel. You don't know if they want the cheapest option or the most experienced. You don't know if they're a homeowner or a property manager. You don't know if they're shopping or already decided. You don't even know what they actually need — you just know what they typed into a box.

Keywords aren't intent. Keywords are guesses at intent. The whole SEO game has been a thirty-year exercise in pretending that the words someone types are the same as what they actually want.

Keywords are guesses at intent. AI is the first system that can ask.

The conversation that changes everything

Here's what the same search actually looks like when a customer asks an AI platform instead of typing into a search bar. This isn't theoretical. This is a conversation any of these systems will have with a customer right now.

Look at what happened in that conversation. The customer started with "best plumber in Denver" — the same six-word keyword phrase that every Denver plumber has been fighting over for twenty years. By the time the conversation ended, they didn't want any plumber. They wanted a plumber who specializes in tankless water heater installation. Completely different customer. Completely different sale. Completely different business.

A traditional search engine could never have done that. It would have handed the customer ten plumber listings, most of which don't do tankless installs at all, and let them figure it out by clicking around.

The two things AI does that search can't

The dialogue above demonstrates two completely new capabilities that didn't exist before AI search arrived. Both are structurally impossible for keyword-based search to ever replicate.

1. It refines intent by asking.

The customer's stated keyword was "plumber." Their actual need was "someone who installs tankless water heaters in two-story homes." Those are not the same query. They produce different shortlists of businesses. They produce different decisions.

A search engine has exactly one shot to guess what the customer wants from the words they typed. AI gets to ask follow-up questions until it knows. By the third question, AI knows more about what the customer needs than the customer typed in any traditional search bar.

That's not a small improvement. It's a structural one. Keyword-based search is a one-shot transaction. AI search is a conversation. You cannot bolt the second one onto the first. They're different products.

2. It educates the customer in real time.

This is the deeper capability — and the one that should genuinely scare anyone who's spent the last twenty years buying Google Ads. Re-read the dialogue. Notice when the customer learned the word "tankless."

They didn't know it existed. They had a slow-hot-water problem. They never would have typed "tankless water heater installer Denver" into a search bar, because they didn't know that's what they needed. The AI taught them what they needed in the middle of the search.

This exposes a problem traditional search has always had, but nobody talked about because there was no alternative: you cannot search for something you don't know exists. If the customer in the dialogue above had been left to Google, they would have searched "best plumber in Denver," called whichever showed up first, paid for a service call, gotten a band-aid fix on their tank heater, and never learned that tankless was an option. The plumber who specializes in tankless conversions — the one whose business would actually serve that customer best — would never have been part of the conversation.

Traditional search can't fix this. It's structurally incapable of teaching. It can only match words to words.

What this means for small businesses

If you've been pouring money into Google Ads for the last decade, here's the part you need to sit with. The reason your cost-per-click has been climbing is not just inflation or competition. It's that the entire keyword-matching model is becoming less valuable. The customers who used to type "best plumber in Denver" into a search bar — your highest-converting customers — are increasingly asking ChatGPT instead. The ones who haven't switched yet, will. And when they do, your Google Ads spend doesn't follow them.

The customers asking AI aren't worse customers. They're better customers. By the time they get a recommendation, they've had a four-turn conversation that established their actual need. The AI didn't hand them a list — it handed them one business name. Yours, if you're the right fit. Nobody else's, if you're not.

Which means: every dollar of attention you used to spend trying to win in keyword search is being silently re-routed to a system where the question isn't how high do I rank for "plumber Denver". The question is: does AI know, with confidence, that I'm the plumber who installs tankless systems in Denver?

That's a question about how clearly your website declares what you do, where you do it, and what you specialize in — in language an AI model can actually read. It's not about keywords. It's not about backlinks. It's not about content volume. It's about whether your business is legible to AI as a specific, well-defined answer to a specific, well-understood need.

The honest path forward

If you accept the argument above — and I'd encourage you to test it; ask ChatGPT or Perplexity or Claude for a recommendation in your own category right now and see what comes back — then the work in front of you is pretty clear.

You need your website to declare, in code AI can read, exactly what kind of business you are, exactly what problems you solve, exactly where you operate, and exactly what makes your specialty different. Not in marketing copy. Not in keyword-stuffed paragraphs. In structured data: the machine-readable layer underneath your visible site that tells AI platforms who you are with no ambiguity.

That's the work. It's not glamorous. It doesn't involve bidding on keywords. There's no monthly invoice from a platform. But it is the work that determines whether the next thousand customers in your area, asking AI for someone like you, hear your business name or somebody else's.

The phonebook ended. The yellow pages ended. Google's ten blue links are ending. What replaces them is a single conversation, between one customer and one AI, that ends with one recommendation. You either earn that recommendation or you don't.

The good news: earning it is well within reach. The signals AI reads are entirely on your own website — work you control, work that doesn't expire, work that compounds. The opportunity is sitting right in front of every small business in America that hasn't realized yet what just happened.

The bad news: someone in your market will figure it out first.