Give AI one clear story to tell about you.

Ask ChatGPT what your company does. Then ask Gemini and Perplexity. If the three answers disagree, or sound like a competitor, you do not have a content problem. You have a positioning problem that AI has made public.

AI engines recommend the brands they can describe in one sentence: who they are for, what they are best at, and why they are different. Fuzzy positioning gets fuzzy answers, or no answer at all. That is why our team starts where great marketing starts, with your ICP, buyer journey, positioning and messaging.

This page explains what our positioning and messaging audit covers and what you get at the end. It also shows how we test whether the new story changed what AI says about you.

What is a positioning and messaging audit for AI search?

A positioning and messaging audit checks whether the story you tell about your company is clear, consistent and repeatable by AI engines. We review your ICP, personas, buyer journey, positioning, USPs and messaging. Then we compare them with how buyers really ask AI and with what AI says about you today. The output is a rewritten core narrative, a message house and a plan for where each message must live.

It is brand strategy for the AI-search era. The tools are familiar to any good marketer: segmentation, jobs-to-be-done, a positioning statement, a message house. What changed is the audience. Your buyer still reads your site, but an AI engine now reads it first, along with your docs, reviews, community threads and press.

Why a classic positioning project is not enough

A classic positioning project ends with a deck and a tagline. That works when humans are the only readers. AI engines are different readers. They piece your story together from dozens of sources, and they trust a fact more when several sources agree on it.

So the audit has to go further than the deck. It has to cover every place an engine looks, and it has to end with a test. Without the test, you only have an opinion about your positioning.

Why does AI search start with positioning, not content?

AI search starts with positioning because an engine can only repeat a story it can understand. If your site, reviews and press describe you in five different ways, the engine picks one, blends them, or names someone clearer. More content does not fix that. A clear position, stated the same way across sources, gives every engine one sentence to repeat.

In Positioning: The Battle for Your Mind (1981), Al Ries and Jack Trout argued that positioning is not what you do to a product. It is what you do to the mind of the prospect. That idea still holds. The mind just changed. Your prospect now asks an AI to build the shortlist, so the battle for the mind includes the model.

Fuzzy in, fuzzy out

We see the same pattern again and again. A company calls itself a platform on the homepage, a tool in its docs and an agency on LinkedIn. Its reviews praise a feature the homepage never mentions. A buyer asks AI for the best option in the category. The engine has no single story to repeat, so it hedges or moves on.

When ChatGPT names you, it stakes its own credibility on you. That is a trust transfer, and you cannot buy your way into it. You earn it by being easy to describe and easy to verify.

"Do the brand work first (ICP, journey, positioning, messaging), learn exactly how each engine reads it, then test whether AI now tells your story."

Krishna Kaanth M, Founder, MaximusLabs

What exactly do we audit?

We audit seven connected parts of your brand strategy. They are the ICP, the buying committee, pains and jobs-to-be-done, the buyer journey, positioning, USPs and messaging. Each one answers a question AI also has to answer about you. If any of them is vague or out of date, the engine fills the gap with guesses or with a competitor's story.

Where the inputs come from

We build the ICP and personas from evidence, not workshops alone. That means your sales calls, CRM data, win and loss notes, and time with your founders. A persona that only lives in a slide deck rarely matches the words buyers type into ChatGPT.

We keep each part as a living document, dated and linked to the call or data it came from. An approved decision outranks a passing remark in a later meeting. That way your positioning does not drift every time someone has a new idea on a call.

How do we compare your story with buyers and with AI?

We line up three versions of your story side by side. First, the story you tell on your site, in sales decks and on calls. Second, the questions buyers actually ask AI at each stage of the journey. Third, what AI says about you today, taken from our AI perception audit. The gaps between those three versions become the work plan.

The third input comes from our AI perception audit. It shows how each engine describes you, attribute by attribute, who it compares you with and which sources it trusts. We do not guess at what AI thinks. We read the answers.

The six stages of the buyer journey

Buyers ask different questions at each stage, so we map prompts stage by stage. Here is how that looks for Acme, our fictional sample brand:

• Problem: "Why does our forecast keep missing?"
• Solution: "How do RevOps teams fix forecast drift?"
• Shortlist: "Best data platforms for Series B RevOps teams?"
• Compare: "Acme vs Lumen for HubSpot users?"
• Validate: "Is Acme SOC 2 compliant? What do users say?"
• Decide: "Which one should a 40-person RevOps team pick?"

Most brands only check the shortlist stage. Deals are often lost earlier, when AI frames the problem around someone else, or later, when it cannot find proof.

The four gaps we look for

• Missing. Buyers ask about a use case you serve, and AI never mentions you for it.
• Drift. Your sources describe you in different ways, so AI blends them into something vague.
• Wrong category. AI files you next to products you do not compete with, or below your real tier.
• Unproven claims. You say you are the best fit for a segment, but no third-party source backs it up.

Each gap gets a fix and an owner. Some fixes are a sentence on a pricing page. Others need a new comparison page, a refreshed review profile or a founder essay that finally states your point of view.

What do you get at the end of the audit?

You get six working documents your team can use the next day. A positioning statement. ICP and persona briefs. A message house. A proof-point inventory. Page-by-page messaging fixes for your site. A source-distribution plan that maps each message to the places AI reads. Each one is written to be used by sales, product marketing and content, not filed away.

Written for people and engines

Every deliverable is written in plain sentences that a human can say out loud and an engine can quote. We avoid slogans that only make sense with a designer's layout around them. If a line cannot stand alone in an AI answer, we rewrite it.

How does the message house work?

A message house is a one-page map of what you say. The roof is your core claim. The pillars are the three or four reasons a buyer should believe it. Under each pillar sit the proof points and the honest answers to common objections. For AI search, we add one more column: the sources where each message must appear, so engines find the same facts everywhere.

Here is a short example for Acme, the fictional brand we use in our samples:

Drift is the enemy

The most common finding is drift. One old press kit still calls the company an enrichment tool. A partner listing uses last year's tagline. Each looks small. Together they give AI a second story to repeat, so we track every drift and fix it at the source.

Where does each message need to live so AI picks it up?

Each message needs to live in the places AI engines read and trust, not only on your website. That means your site and docs, review sites, communities such as Reddit and LinkedIn, YouTube, and the press. Different engines lean on different sources, so a message on one surface is a rumour. The same message on several independent surfaces starts to look like a fact.

The source-distribution plan turns this table into a to-do list with owners. Our GEO and AEO work then ships those changes, from site copy to earned citations.

Each engine has its own sources

A brand can appear in ChatGPT and still miss Google's top results, because each engine trusts different sources. That is why we never plan distribution for "AI" in general. We plan it for ChatGPT, Perplexity, Gemini, Google AI Mode and Claude, one by one.

Should you create a new category or win an existing one?

Most companies should win a clear, existing category first and then narrow it. A new category is a long, costly bet. With AI search it carries an extra risk: if buyers never ask for your new label, engines have no prompt to answer with your name. We test both paths against real buyer prompts before you commit to either one.

Category design is a real discipline. Play Bigger (2016) was written by Al Ramadan, Dave Peterson, Christopher Lochhead and Kevin Maney. It made the case that category kings take most of the value in their market. We agree that owning a category is the prize. We disagree that a new name is the fastest way to get there.

The bridge approach

Our usual advice is a bridge. Anchor yourself to a category buyers already search for, then add the angle only you can claim. "The data platform built for RevOps" works because "data platform" is a known shelf and "built for RevOps" is the difference.

Ries and Trout made a related point decades ago. It is easier to be first in a narrow category in the prospect's mind than to be second in a broad one. AI engines behave the same way. They name the brand that clearly owns the narrow slot.

How do we know the new positioning landed?

We test it. Before any change ships, we record a baseline: a fixed set of custom prompts, run on every tracked engine. After the new positioning goes live, we rerun the same prompts and compare. Each prompt runs several times on each engine, and we call a change only when it clears the noise band. That is how a positioning audit ends in proof, not opinion.

We run these tests as AI search experiments on Max Visibility, our own measurement tool. It is message testing for AI. A tracker tells you that you appeared in a share of answers. Our experiments answer a better question: did the new positioning land, on which engines, and because of which sources?

What Max Visibility shows us

• Prompts: the baseline prompt set, tagged by journey stage and persona.
• Chats: each engine's full answer to each prompt, so we can read how it describes you.
• Ranking: where you appear in each answer, next to the brands you compete with.
• Sources: the domains and URLs each engine cites, so we can see which source moved the answer.
• Insights and Brands: patterns across prompts, measured against your competitor set.

A worked example

Say AI calls Acme "a cheap enrichment tool" when it wants to be known as a RevOps data platform. The hypothesis is simple. If the site, G2 profile and press kit all use the new label, engines will drop the old one.

We record the baseline, ship the changes, then rerun the same prompts. Perhaps Gemini now uses the new label and Perplexity does not. The Sources screen shows which pages Perplexity still cites, so we know which source to fix next.

Max Visibility tracks ChatGPT, Gemini, Perplexity, Google AI Overviews and Google AI Mode, across countries and languages. It tracks Claude too, as an add-on model. Max Visibility is included for full-service GEO clients.

Who is this audit for?

This audit is for B2B SaaS and AI companies, roughly $5M to $100M in ARR, whose marketing leader answers for pipeline. It fits best when AI describes you wrongly or when the market has not caught up with a repositioning. It also fits when prospects keep confusing you with someone else. It is a poor fit if you want results this week.

Signs you need it now

• ChatGPT, Gemini and Perplexity describe you in three different ways.
• You launched a new product or moved upmarket, and AI still describes the old company.
• Your homepage, review profiles and press boilerplate use different category labels.
• Competitors with weaker products get named first in AI shortlists.
• Your team cannot agree on the one sentence that describes the company.

It is not the right service if you want paid ads, ABM or outbound. We are inbound only. We earn your place in AI answers. We never buy it.

If you sell in a specific market, our industry pages show how this work applies to AI and SaaS, fintech and e-commerce brands.

How do you get started?

Start with a GEO strategy call. We look at what AI says about you today and talk through your ICP and positioning. Then we tell you honestly whether this audit is the right next step. Most clients begin with the AI perception audit, move into positioning and messaging, and then test the changes with AI search experiments on Max Visibility.

The order matters. First we see the story AI tells about you. Then we decide the story you want it to tell. Then we ship it everywhere AI looks and prove that it changed.

Agencies sell content volume. Tools sell dashboards. We sell the story AI tells about you, and proof that it changed.

Book a GEO strategy call