How It Works
How It Works
This page answers the questions we hear most: why AI recommends some companies and not others, what you can actually do about it, and what working with us looks like if you decide you want help.
This page answers the questions we hear most: why AI recommends some companies and not others, what you can actually do about it, and what working with us looks like if you decide you want help.
Why doesn’t my company show up when someone asks AI for a recommendation?
Why doesn’t my company show up when someone asks AI for a recommendation?
It starts with understanding that being known and being recommended by AI are two different jobs.
AI models are trained on a huge amount of text, so most established B2B companies are “in there” somewhere. Ask a model to describe your company by name and it usually can. Ask it to recommend a company for the problem you solve, without naming you, and it’s clear there’s a different test. The model has to decide, from everything it knows, who to put forward. That decision leans on a specific claim (what exactly do you do, for whom) and independent confirmation (has anyone besides you said so). Miss either one and you can be accurately described and still never recommended.
It starts with understanding that being known and being recommended by AI are two different jobs.
AI models are trained on a huge amount of text, so most established B2B companies are “in there” somewhere. Ask a model to describe your company by name and it usually can. Ask it to recommend a company for the problem you solve, without naming you, and it’s clear there’s a different test. The model has to decide, from everything it knows, who to put forward. That decision leans on a specific claim (what exactly do you do, for whom) and independent confirmation (has anyone besides you said so). Miss either one and you can be accurately described and still never recommended.
How do AI systems decide which companies to mention?
How do AI systems decide which companies to mention?
Mostly two things: how clearly a company’s role is stated, and how many independent sources repeat it the same way.
“Independent” is the key word. If the only place your positioning appears is your own website, that’s one source telling the model one thing. If it shows up consistently across press, reviews, directories, and your own site, that’s several sources agreeing, and models weight agreement heavily. This is also why two companies with similar products or services can get very different treatment.
Mostly two things: how clearly a company’s role is stated, and how many independent sources repeat it the same way.
“Independent” is the key word. If the only place your positioning appears is your own website, that’s one source telling the model one thing. If it shows up consistently across press, reviews, directories, and your own site, that’s several sources agreeing, and models weight agreement heavily. This is also why two companies with similar products or services can get very different treatment.
Is there a system for fixing this?
Is there a system for fixing this?
There is a system. I built this framework after watching companies solve for the wrong thing: publishing more content when the real gap was that nobody outside that company was confirming their story, or investing in press before the underlying story was even right.
Here is how I think about understanding, activating, and governing AEO and GEO:
Measure. Ruth the baseline. See what AI is actually saying about you right now.
Diagnose. Trace the finding back to its cause. Is the story unclear, or is it clear by unconfirmed by anyone besides you?
Translate. If the story itself is the problem, fix that first. A sharper claim, said the same way everywhere it appears.
Activate. If the story is fine but nothing independent backs it up, this is where that gets built through the authority sources identified as trustworthy.
Monitor. Now watch it and keep watching it. AI’s answers shift as the market changes and as the models themselves change. Monitor share of voice along with your other performance metrics.
There is a system. I built this framework after watching companies solve for the wrong thing: publishing more content when the real gap was that nobody outside that company was confirming their story, or investing in press before the underlying story was even right.
Here is how I think about understanding, activating, and governing AEO and GEO:
Measure. Ruth the baseline. See what AI is actually saying about you right now.
Diagnose. Trace the finding back to its cause. Is the story unclear, or is it clear by unconfirmed by anyone besides you?
Translate. If the story itself is the problem, fix that first. A sharper claim, said the same way everywhere it appears.
Activate. If the story is fine but nothing independent backs it up, this is where that gets built through the authority sources identified as trustworthy.
Monitor. Now watch it and keep watching it. AI’s answers shift as the market changes and as the models themselves change. Monitor share of voice along with your other performance metrics.
What should I fix first?
What should I fix first?
Whichever one is actually broken!
Most companies assume the fix is “more or better content” and go straight there. Often the real problem sits upstream: the positioning itself isn’t specific enough to repeat. If that’s the case, publishing more of a vague message just spreads the vagueness further. Other times the positioning is genuinely sharp, and the real problem is that almost nothing independent confirms it. What moves that is getting the same claim repeated somewhere you don’t own: a review, a press mention, an analyst note, a community answer.
Same symptom, low or no recommendation, two very different causes, and two very different fixes. Which one you have isn’t really guessable from the outside so it’s the first thing we test for.
Whichever one is actually broken!
Most companies assume the fix is “more or better content” and go straight there. Often the real problem sits upstream: the positioning itself isn’t specific enough to repeat. If that’s the case, publishing more of a vague message just spreads the vagueness further. Other times the positioning is genuinely sharp, and the real problem is that almost nothing independent confirms it. What moves that is getting the same claim repeated somewhere you don’t own: a review, a press mention, an analyst note, a community answer.
Same symptom, low or no recommendation, two very different causes, and two very different fixes. Which one you have isn’t really guessable from the outside so it’s the first thing we test for.
What is an AI Representation Baseline, and how does it help?
What is an AI Representation Baseline, and how does it help?
An AI Representation Baseline helps you to “measure” and “diagnose” what AI is saying about you. This is where everything starts, for a plain reason: you can’t fix a story you haven’t read.
It is a one-week test of what ChatGPT, Perplexity, Google and/or Claude actually say about you. What comes back is two kinds of insights:
Quality scores about your brand. Is your messaging using buyer language? Are you distinct from competitors? Are you visible?
Baseline business metrics. You get a starting point on four numbers: Representation Gap, Narrative Consistency, Share of Answer, and Source Composition.
An AI Representation Baseline helps you to “measure” and “diagnose” what AI is saying about you. This is where everything starts, for a plain reason: you can’t fix a story you haven’t read.
It is a one-week test of what ChatGPT, Perplexity, Google and/or Claude actually say about you. What comes back is two kinds of insights:
Quality scores about your brand. Is your messaging using buyer language? Are you distinct from competitors? Are you visible?
Baseline business metrics. You get a starting point on four numbers: Representation Gap, Narrative Consistency, Share of Answer, and Source Composition.
How do you measure whether it’s actually working?
How do you measure whether it’s actually working?
The four number the baseline sets your starting point:
Representation Gap. The distance between your desired positioning and how AI represents you.
Narrative Consistency. Whether the description holds the same way across sources and platforms.
Share of Answer. How often you show up when no brand name is queried.
Source Composition. Where the story is coming from: your owned site or channels, or a real mix of authoratative, independent sources.
The four number the baseline sets your starting point:
Representation Gap. The distance between your desired positioning and how AI represents you.
Narrative Consistency. Whether the description holds the same way across sources and platforms.
Share of Answer. How often you show up when no brand name is queried.
Source Composition. Where the story is coming from: your owned site or channels, or a real mix of authoratative, independent sources.
Most companies have never looked at how AI describes them.
Contact us to see what AI says about you →
Most companies have never looked at how AI describes them.
See what AI says about you
—>