AI search visibility · AI governance software
Why do some companies get recommended by AI while others don't?
Why do some companies get recommended by AI while others don't? I decided to investigate this in one specific sector: AI governance and compliance software.
I tested 5 AI assistants (ChatGPT, Claude, Gemini, Perplexity and DeepSeek) using 10 buyer-intent questions, analysing 50 answers, 1,293 brand mentions and 179 companies. I also reviewed 30 third-party ranking articles and 20 company websites.
- AI assistants
- 5
- AI assistants
- buyer-intent questions
- 10
- buyer-intent questions
- answers analysed
- 50
- answers analysed
- brand mentions
- 1,293
- brand mentions
- companies
- 179
- companies
- ranking articles
- 30
- ranking articles
- company websites
- 20
- company websites
7 findings that surprised me
01Mentions ≠ recommendations
Seven companies appeared in at least 15 answers but never reached the top five. One was mentioned 38 times without a single top-five placement.
02A few companies dominated
Just 5 companies captured 55% of all top-five placements. Only 51 of 179 companies reached the top five.
03AI assistants disagree
On a typical question, two assistants shared only around 32% of their top-ten recommendations.
04Technical SEO isn't everything
Some less-visible companies had llms.txt, structured data and accessible websites. Technical foundations matter, but they don't guarantee recommendations.
05Positioning matters
Some companies had extensive content but described themselves inconsistently. One company's old name kept appearing in AI answers after a rebrand.
06Great content isn't enough
One company had the strongest measured content among dedicated vendors, yet was absent from two assistants and appeared in just 4 of 30 ranking articles. Analyst recognition also didn't guarantee top placements.
07Visibility extends beyond your website
Third-party article coverage correlated strongly with AI mentions. Leading companies generally linked a YouTube channel from their websites; less-visible companies had no YouTube presence, even when they published webinars or podcasts.
These are observed patterns, not proof of causation.
What should businesses do?
- Test real buyer questions across multiple AI assistants.
- Track top-three and top-five placements, not just mentions.
- Make your positioning and use cases crystal clear.
- Audit crawlability, sitemaps and structured data.
- Build credibility through genuine reviews, comparison articles, analyst coverage and discoverable video.
Repeat the tests and measure changes over time.
The takeaway
This exploratory study changed how I think about AI visibility. It's not just about technical optimisation or publishing more content.
Are you building a great product, or making it easy for buyers and AI assistants to understand, compare and recommend it?
The next visibility challenge may be getting onto the shortlist when a buyer asks AI what to choose.
Source: 5 AI assistants × 10 buyer questions (50 answers), 30 ranking articles, 20 company websites. October 2026.