Why your brand is invisible to ChatGPT (and how to fix it)

The shift from search engines to AI systems is moving faster than most marketing teams realise. A growing share of discovery journeys now start with a question to ChatGPT or Claude, not a Google search. Queries like “what's the best CRM for a 20-person sales team” or “top project management tools for agencies” increasingly get answered by AI systems that synthesise information and make direct recommendations.

If your brand isn't in those recommendations, you're invisible at the moment of consideration. Not low-ranking, but absent. And unlike Google, where you can see your position for any keyword, most brands have no visibility into whether or how they appear in AI answers.

Here are the five most common reasons brands are invisible to AI recommendation engines, and the concrete steps to fix each one.

1You optimised for keywords, not entities

Google rewards pages that rank for specific keywords. AI systems work differently. They build a model of the world in terms of entities (brands, products, people, concepts) and the relationships between them. A brand that has 50 pages targeting "project management software" keywords but almost no third-party content that mentions the brand name as an entity will rank well on Google and be invisible on ChatGPT. The fix isn't to abandon keyword content. It's to build entity presence alongside it.

2Your brand doesn't appear in the sources AI trusts

AI systems are trained on large corpora of web text, but not all sources are weighted equally. Authoritative review sites (G2, Capterra, Trustpilot), Wikipedia, tech publications (TechCrunch, Product Hunt, Hacker News), and community platforms (Reddit, Quora) carry far more weight than brand-owned content. A brand with 10,000 blog posts on its own domain but no presence on any of these sources is invisible to AI recommendation engines. Building presence on third-party sources is the highest-leverage AI visibility tactic that most brand teams aren't doing.

3You have no machine-readable identity signal

When an AI system crawls your site, it pieces together who you are from your page text. If your homepage says "We help teams work better together" and your about page says "The operating system for modern teams" and your G2 listing says "Project management tool for remote work", the AI system forms an inconsistent model of your brand. Deploying an agent.json file at your domain root solves this. It gives AI systems a single, structured, first-party source of truth about who you are, what you do, and why you should be recommended.

4Your category is dominated by established players with higher entity authority

If you're competing in a category with players like Notion, Asana, or Monday.com, those brands have years of training data, thousands of third-party mentions, millions of user discussions, and deep entity presence across every authoritative source. A new entrant doesn't beat that overnight. But "less authority" is different from "no authority". Niche positioning helps: if you serve a specific vertical or use case that the big players don't emphasise, you can own that sub-category in AI recommendations even without matching their overall entity authority.

5You've never measured it

This is the most common reason. Most brands assume their Google SEO translates to AI visibility. It almost never does. Google SEO optimises for a ranking algorithm. AI recommendation engines run on fundamentally different signals. Brands that haven't audited their AI visibility are flying blind. They don't know their score, which platforms mention them, how often, at what position, or with what sentiment. The first step to fixing AI invisibility is measuring it. Everything else follows from that baseline.

The fix: measure first, then act

Every fix above starts with the same step: measure your current AI visibility so you know where you actually stand. Without a baseline, you're guessing at which of the five problems you have.