A few weeks ago, I was looking through our Google Search Console data and noticed a group of searches that didn’t look like searches at all.
They were unusually specific, written almost like mini buyer profiles:
“about me: B2B technology consultant evaluating CPQ modernization options for mid-market clients, looking for implementation partners”
“about me: Salesforce admin responsible for Revenue Cloud rollout, researching vendors with deep CPQ integration experience”
There were roughly 15–20 variations like this, and Mountain Point was showing up near the top of the results for many of them.
What caught my attention wasn’t the ranking. It was the format.
Nobody searches Google like that.
Or at least, people didn’t used to.
The buyer may not be doing the research anymore
We are getting used to asking ChatGPT, Claude, Perplexity, Gemini, and other AI tools questions we previously would have taken to Google ourselves.
“Find me three implementation partners that understand complex CPQ.”
“Compare Salesforce Revenue Cloud partners that have manufacturing experience.”
“Who could help us modernize our quote-to-order process?”
The AI tool then does some combination of searching, reading, comparing, and synthesizing before giving the user an answer.
That changes something important for companies like ours.
The buyer may still be researching you, but they may never personally perform the search.
An AI system is doing the first round of diligence on their behalf.
Those strange Search Console queries were the first time I had seen what that might look like from the other side.
Then I looked at the traffic
Around the same time, we started seeing unusual referral behavior from ChatGPT.
One visit to our CPQ modernization content lasted more than 80 minutes and included 11 pageviews. Another session generated 50 pageviews over roughly 78 minutes. A visit to one of our customer stories generated more than 60 pageviews in about half an hour.
I can’t say with certainty what happened inside those sessions. Analytics data has limitations, and AI-driven browsing doesn’t always behave like a normal human visit.
But the pattern was hard to ignore.
Something was moving systematically through our content, looking at service pages, expertise, customer stories, and supporting material.
That is exactly what I would want someone—or something—to do if they were evaluating Mountain Point as a potential partner.
And increasingly, I think that distinction matters less.
Your website has a new audience
We have always written websites for people.
That is still important. But now there is another audience reading the site before the buyer ever gets there.
AI systems need to understand what you do, where you have expertise, which customers you serve, what problems you solve, and why you are different.
If that information is vague, buried in marketing language, or spread across disconnected pages, a human may eventually piece it together.
An AI system may simply move on.
This is where I think the conversation around “AI SEO” sometimes gets too tactical. The opportunity is bigger than adding schema or writing more FAQs.
Companies need to make themselves easier to understand.
For us, that means being explicit about things like CPQ modernization, Salesforce Revenue Cloud, ServiceNow CRM and CPQ, quote-to-order, manufacturing, complex configuration, and the kinds of projects we have actually delivered.
The clearer the signal, the easier it is for both humans and AI systems to understand where we fit.
Traffic may become the wrong measurement
This also creates an uncomfortable issue for traditional marketing metrics.
We rank well for a number of very specific questions around Salesforce Revenue Cloud and CPQ. Increasingly, some of those searches are answered directly inside search results or AI-generated responses.
The user gets the answer without visiting our website.
Historically, we would call that a zero-click search and treat it as a loss.
I’m not sure that’s the right way to think about it anymore.
If an AI system reads our content, uses it to answer a buyer’s question, and begins associating Mountain Point with a particular area of expertise, the website may have done its job even though Google Analytics shows no conversion.
The buyer might never consume the original content.
They may simply hear:
“Here are three firms you should consider.”
Being one of those three may ultimately matter much more than getting another pageview.
This changes what good content looks like
There is still plenty we don’t know about how different AI platforms discover, evaluate, and prioritize information. I would be cautious about anyone claiming they have already figured out the new ranking algorithm.
But the direction seems fairly clear.
Specific beats vague.
Evidence beats claims.
Clear expertise beats broad positioning.
Customer stories matter because they establish that you have actually done the work.
Direct answers matter because they are easier to extract and reuse.
And external references increasingly matter because your own website is only one source an AI system can use to decide whether you are credible.
In other words, a lot of the fundamentals haven’t changed.
The audience has.
The part I’m paying attention to
For years, B2B marketing has been built around getting a buyer to your website, capturing their information, and nurturing them through a funnel.
AI puts another layer in front of that process.
A buyer can now ask an AI tool to understand the market, define the options, compare vendors, identify risks, and potentially create the shortlist before visiting a single vendor website.
That means part of the buying decision may be happening before you ever know a buying process exists.
The strange Search Console queries are what made that real for me.
Someone—or some system—was asking Google a question about exactly the type of work Mountain Point does.
The query wasn’t written for a human to read.
But we still needed to be in the answer.
That may be one of the biggest changes coming to B2B marketing: your next prospect may research you extensively without ever visiting your website.
The question is no longer just whether people can find you.
It is whether the systems helping them make decisions can understand why you belong on the shortlist.