Four Leads Walked In. AI Sent Every One of Them.
Something happened at our firm recently that I am still thinking about.
Within a single week, four serious prospects reached out to us - and what made them stand out from our usual flow of inbound leads was how they got to us. Not tire-kickers, either: real companies with real budgets and real challenges. When our business development leads asked each of them how they had found us, the answer was the same: an AI told them about us. Some had been talking to ChatGPT. others to Claude. And in a couple of cases, when we asked what exactly they had searched for, the person said, “You know what, I’ll just send you the whole conversation.”
So, we read the conversations. And that is where it got interesting.
None of them had typed anything like “best reputation management firm” or “who is Five Blocks.” That is the old behavior - the keyword behavior, the behavior we have all spent twenty years optimizing for. In the old world, we thought of search as a ranked list. The queries were short, and they didn’t carry much intent. Instead, each of these people had told the AI a story. “I run a fund. Last year we went through something difficult. We had some challenging coverage. I’m looking for a firm that can help with the following kinds of things.” Paragraphs, not keywords. A situation, described in some detail, the way you would explain it to a trusted advisor.
And the AI “listened” to that story, asked some follow-up questions, went out into the world, found us, and “referred” them to Five Blocks.
Content is no longer just read. It is matched.
Here is the way I have started explaining this to clients, and it has stuck.
In the old world, search resulted in a list of relevant sources - you typed a few words, a machine handed you ten blue links, and you, the human, skimmed them top to bottom, clicked a couple, and formed your own opinion. The content sat there passively, waiting to be browsed.
The newer behavior doesn’t work like that. When someone describes their situation to an AI, the model isn’t just ranking pages. It is also looking for fit. Think of the person’s story as a molecule covered in very specific bumps and spikes - I run a fund, we had this particular issue, I need these particular things. The AI scans the available content looking for other shapes whose spikes lock into theirs. Where there is a match, you surface. Where there is no match, you may simply not come up in that conversation - no matter how well you “rank.”
I want to be careful here, because this doesn’t mean being ranked as a top company in your field stops mattering. It matters. Part of what lets you fit the lock is that you are highly rated and well regarded overall. It’s just that, increasingly, being highly ranked is not enough on its own. The real winner is the fit.
This is also truer for some kinds of searches than others. Where it really shows up is when someone is evaluating reputation from a specific point of view - they have a particular situation and they’re sizing you up against it. If a person asks a more general question - “tell me about this company” - those bumpy spikes matter less, and the old rules look more familiar.
General content has less to grab onto
The smooth, comprehensive, we-do-it-all corporate page - the one that took six rounds of internal review to sand down to perfection - has flatter spikes. There is less for a specific story to bind to. It reads beautifully to a human skimming for thirty seconds and offers a machine less to work with when it’s trying to match a real set of requirements.
The content that wins tends to be more specific than you would expect. It answers the actual shape of a real situation: the question a real searcher asks at 11 p.m., in their words, about their needs.
In our own work we see this constantly - and it’s worth being precise about what’s doing the work. It isn’t really the page. It’s the narrative. A detailed, deeply specific narrative outperforms a more standard one, because the standard one had less for the model to hold.
What this adds for communicators
The brief doesn’t flip so much as grow. For years the content question was, “Did we describe what we do?” That question still matters. The new one sits on top of it: “Did we also answer the specific situations our buyers describe to a machine when no one is watching?” It’s an addition, not a replacement.
A few practical moves follow from that:
Map the predicaments, not just the services. Make a list of the real situations your buyers find themselves in - the ones that send them to an AI in the first place - and write to those, in their language, not your internal language.
Give content spikes. Specific, scenario-shaped answers bind. Broad mission statements give a model less to lock onto.
Density as well as polish. You actually have to satisfy two audiences now: the AI that’s matching, and the human who still shows up and reads. You need both.
Write for the matcher and the skimmer. Your audience increasingly includes an AI that will go and speak to the human on your behalf. Win the match and you win the introduction - but the human still arrives, so don’t lose them either.
None of this means search, websites, or earned media stop mattering. People still go to your website. It’s just that the same material now serves a second purpose - it’s the raw material the AI reads. The job of that content hasn’t switched from being browsed by people to being matched by machines; it’s doing both at once. And earned media can arguably play an even bigger role than before as it adds legitimacy and authority to the specific descriptions of your company or products.
And it’s worth naming who the new influencers are. We used to spend our energy on the outlets and the analysts and the journalists who shaped a story. They still matter. But four of the most influential “readers” of your content today are ChatGPT, Claude, Perplexity, and Gemini - and increasingly they are the ones making the introduction.
Four leads taught us that the introduction now often happens before anyone ever visits your site. And here’s the part that stayed with me: of those four, two signed up for services within a few days. The question is whether your story has the right shape to be found.
