Your AI Visibility Problem Is a Positioning Problem

I built an audit that assesses how AI search tools find and describe a business. Then I ran it on my own.

The result wasn't what I expected. Both ChatGPT and Perplexity find Blandford Creative when the question names my city. Ask either one for a fractional marketing leader in Victoria, BC, and I come back. Ask for the best marketing consultants for engineering and professional services firms in Canada, with no city attached, and I disappear. Perplexity returned a ranked list of eight firms and put a competitor at the top, describing them as built specifically for engineering and industrial buyers.

That's my positioning. Someone else was getting credit for it. Not because they do better work, but because they'd said it plainly and I hadn't.

What made the finding uncomfortable was where the cause sat. My verticals were on my site. Cleantech, sustainability, engineering, energy, professional services, all named on the Who We Are page. They were missing from the service descriptions on What We Do, which is the page someone lands on when they're looking for exactly what I do. My home page headline described a feeling instead of a service. The site spoke in two voices: emotional and general on the pages selling the work, specific and credible on the blog.

Nothing technical was broken. Every page was indexed and readable. I'd simply never committed my specialty to the pages where it counted.

In an earlier post I described the two layers that decide whether AI tools surface a business. Technical signals determine whether a machine can read you. Narrative signals determine whether it can describe you. Most of the conversation about AI visibility lives in the technical layer. The narrative layer is where businesses lose.

The Layer Everyone Talks About

Let's start with the technical layer, because it's real and it's easy to miss.

I audited an architecture firm this summer with a national design award and a project AI systems already knew about. Asked for design-forward firms working in public realm and infrastructure in their city, ChatGPT surfaced their own award-winning project, unprompted, then returned a shortlist of seven firms. They weren't on it.

Part of the cause was a single setting. The page describing the practice, its principals, their professional designations and their awards carried a directive telling search engines and AI crawlers to skip it. That page was also left out of the sitemap. It rendered perfectly in a browser. Any visitor could read it. No machine could. The richest page on the site was the one page they'd switched off, and nobody at the firm knew.

That kind of failure is invisible from the front end, which is why it survives for years. If your technical foundations are broken, nothing else you do reaches a machine.

But switching that page back on wouldn't have been enough. Their home page had no body copy at all. No description of the practice, no statement of what they do, nothing a system could lift and quote. When Perplexity was asked for the best small firms in their city for mixed-use and community projects, it named two competitors and cited both from their own websites, because those firms had written plain sentences about the kind of work they take. The architecture firm hadn't. Their competitors didn't outrank them on merit. They outranked them by having written something down.

One page switched off, and one page with nothing on it to say. Both problems, on the same site, producing the same silence.

When The Foundations Are Fine And You Still Can't Be Found

Earlier this year I audited a craft distillery whose content is unusually strong. Named awards, two active sustainability certifications, specific technical detail about the systems behind their carbon-neutral commitment. Their site was well built and it read well. A person landing on it learned exactly what the business was and why it was unique.

Asked for sustainable or carbon-neutral Canadian distilleries, ChatGPT returned five competitors and didn't mention them. That was the query built most precisely around their strongest claim.

Part of that is scale. They're a small producer without the accumulated press coverage and directory presence that older competitors have built up, and AI systems lean on those outside sources. But part of it sat in the language, and that part was fixable. The site described the business the way the founders think about it. Buyers don't search that way. They search for award-winning BC gin, or a sustainable distillery on Vancouver Island. And the detail that makes the business distinctive, that it was founded by an engineer and a herbalist, appeared as a single line partway down the home page. It read as a nice touch rather than as the reason to choose them.

They've since worked through the roadmap the audit produced. The home page now opens by naming the craft, the founders' backgrounds, the place and the carbon-neutral commitment in a single sentence a machine can lift whole. The award images carry descriptive labels naming each competition, year and medal. Same business, same awards, same certifications. The difference is that the page now says what they are before it says how they feel about it.

That was the same failure I found in my own audit, in a completely different market. Both businesses described themselves the way they think about themselves instead of the way a buyer evaluates them.

Why Nobody Diagnoses This Correctly

This is where most businesses stall. It looks like a website problem, or a content problem, or a search problem. It's a positioning problem.

When a business describes itself differently across its website, its LinkedIn page, its directory listings and its press coverage, the usual response is to tighten the copy and align the messaging. But inconsistent description is rarely a writing failure. It happens because the positioning was never settled. Nobody decided what the business is, who it's for, and what it does better than the alternatives. Without that decision there's nothing to be consistent with, so every page, every writer and every year produces a slightly different version. The copy is downstream of a strategic decision nobody made.

The Objection To Take Seriously

There's a fair objection to all of this, and it comes from people who know search well. AI visibility, they argue, is an extension of SEO rather than something new.

They're right, and Google says so. In guidance published this year for website owners, Google states that its generative AI features are built on its core search ranking and quality systems, and that from Google's perspective, work aimed at AI search is still search work. The lineage runs deeper than that. PageRank, built on how pages link to one another, has been part of Google's core ranking systems since launch and remains part of them today, according to Google's own documentation. What other websites say about you has always mattered to search. That part isn't new.

What changed is what those outside sources get read for. A link is a vote, and votes get counted. What a source says about you is language, and language gets read. AI systems lean harder on the second.

Google is direct about which lever matters most. Its guidance says content people find unique and useful will influence a site's presence in AI search more than anything else it recommends, and it singles out having a distinctive point of view, because its systems look across many sources and a distinctive one stands out.

Read that as a business problem rather than a content problem. A distinctive point of view is positioning. The largest factor in your AI visibility, by Google's own account, is whether you've decided what you stand for and said it clearly.

One caution on scope. That guidance is written about Google's search features, and it names AI Overviews and AI Mode. It doesn't tell you how ChatGPT, Claude or Perplexity reach their conclusions, and it doesn't speak to Gemini as a standalone assistant either. These are separate systems built by different companies, and they don't all work the same way.

So Which Layer Is Costing You?

You can't tell from the outside, because both failures produce the same result. You ask an AI tool a question your business should own, and someone else's name comes back. Nothing on your website looks wrong. There's no error message and no alert. The only way to know is to test it and then look at what's causing it.

If you try this yourself, three things matter. Run the query logged out, in a private or incognito window, because AI tools use your history and a logged-in search about a business you've visited will flatter you. Ask the question a stranger would ask, not your own name. Something like: who are the best commercial electricians in Vancouver.

And run it more than once, on more than one platform. These systems don't return the same answer every time. The same question in two sessions can produce two different shortlists, and adding a location or a detail can change the result completely. One test tells you almost nothing. Several tests, phrased the way different buyers would phrase them, start to tell you something real.

That's what an AI Visibility Audit does properly. It tests what ChatGPT, Perplexity and other AI search tools currently say about you, identifies the specific gaps in both layers, and gives you a prioritized roadmap so you know what to fix first and what can wait.

If you want to know what a buyer sees when they ask an AI tool about a business like yours, get in touch. Send me the query you think you should own and a link to your site. I'll run it, and we'll look at what comes back and where the opportunities are.