Most companies start their AI visibility work with a score. How often does ChatGPT mention us? Are we ahead of our competitors? Those are useful questions. But they skip a more important one. What do we want AI engines to know us for?
AI tools do not remember brands in the abstract. They connect companies to subjects. When a buyer asks a question, the model looks for the organisations most closely and credibly linked to that subject. If your company is linked to nothing in particular, it is rarely the answer.
This is why we start every engagement by choosing a pillar topic.
What a pillar topic is
A pillar topic is the broad territory you want to be known for. It should be wide enough to matter to buyers, journalists and policymakers. It should also be narrow enough for you to lead it credibly.
“Digital transformation” is too broad. Thousands of companies claim it. “Reducing call volumes in local government” is much closer. A buyer can picture the problem, and a journalist can picture the story.
Beneath the pillar sit sub-topics. These are the specific problems, outcomes and questions your buyers ask. Every piece of coverage, case study or article should strengthen one of them.
Four tests for a good pillar topic
Buyers ask about it. Write down the questions your customers ask before they buy. Run them through ChatGPT, Gemini and Claude. If the topic never comes up, it may matter more to you than to them.
You have proof. AI engines favour evidence. Look for customer results, data or named examples you can share. A topic without proof is just a slogan.
Nobody owns it yet. Check who the engines cite today. If one competitor dominates every answer, a narrower sub-topic may be a better place to start.
You can keep talking about it. Authority builds over months, not weeks. Choose a topic you can return to with fresh stories, research and commentary for at least a year.
Three common mistakes
The first is choosing a topic that describes your product instead of your buyer’s problem. Buyers rarely ask AI about features. They ask how to fix something.
The second is choosing too many. Three pillar topics will stretch a small team thin. One well-supported pillar usually does more than three neglected ones.
The third is treating the topic as a marketing exercise alone. The strongest signals come from independent sources. Trade press, industry reports and expert commentary carry more weight than your own website.
The research points the same way. A peer-reviewed study led by researchers at Princeton University found that adding citations, quotations and statistics from credible sources lifted a source’s visibility in generative engine answers by up to 40%. [1] AI engines reward evidence they can check.
Turning a topic into visibility
Once the pillar is agreed, the work becomes much clearer. Every pitch, case study and article has a job. It either answers a buyer question beneath the pillar or it does not.
We track progress in the same way. Each month we test the questions linked to each sub-topic. We see whether your company appears, how it is described and which sources the engines rely on. Over time you can watch a topic shift from your competitors to you.
The stakes are rising. Pew Research Center found that when Google showed an AI summary, users clicked a traditional search result in just 8% of visits, compared with 15% when no summary appeared. [2] If buyers stop clicking, being named inside the answer is what counts. This is why one audit is never enough. Monitoring needs to be ongoing.
One-off coverage fades. A connected body of work on a clear topic keeps growing in value.
Where to start this week
Pick the three questions your best customers asked before they signed. Test them in ChatGPT, Gemini and Claude. Note who is cited and why. If your company is missing, you have found your first sub-topic.
Want help choosing your pillar topic? Book a free 30-minute call and we will test your buyers’ questions live.
Sources
[1] Aggarwal et al., GEO: Generative Engine Optimization, Princeton University, ACM SIGKDD 2024
[2] Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results, July 2025