You don't need a budget, a tool, or a consultant to start understanding your AI presence. You need an hour, access to one or two assistants, and the right questions — about the outside world (what AI says) and about your own operation (what you are or aren't doing). This is a twenty-question diagnostic, split into four blocks. Answer honestly; the discomfort is the point.
Block 1: What AI Says About You
1. When you ask "what are the best companies in [your industry]?", does your brand appear?
2. Does AI correctly describe what your company does, or get it wrong / confuse you with another?
3. Is the price or business model AI mentions about you up to date?
4. When AI talks about your brand, is the tone positive, neutral, or full of caveats?
5. Does AI cite any information about you that's no longer true (an old phase, a discontinued feature)?
Block 2: How You Compare
6. In an industry recommendation question, do you appear before or after competitors?
7. In the direct matchup "[competitor] or [you]?", which side does AI lean toward?
8. Do you appear as an alternative when someone asks for options to the market leader?
9. Is there a competitor that appears in almost everything while you vanish?
10. Does AI attribute a clear differentiator to you, or describe you generically?
Block 3: Consistency Across Channels and Models
11. Does your presence change a lot from one model to another (strong in one, absent in another)?
12. Is the information about you the same on your site, your profiles, and what AI says?
13. Do your main pages answer real questions or just describe the company?
14. Is your value proposition in clear text, or does it depend on image/video/design?
15. Is your reference content current, or published and abandoned?
Block 4: Your Measurement Maturity
16. Can someone at your company answer, with data, "how does AI talk about us?"
17. Do you have a baseline — a recorded snapshot of your presence at a defined moment?
18. Do you know which third-party sources feed what AI says about you?
19. Do you track how that presence changes over time, or only check now and then?
20. If AI started describing your brand incorrectly tomorrow, how long would it take you to find out?
How to Interpret Your Answers
There's no score to calculate. The value is in the patterns. If you stalled on Block 1, you have a presence or accuracy problem that needs diagnosis before action. If Block 2 stung, a competitive front is being lost without you tracking it. If Block 3 exposed inconsistencies, you have quick wins available — they're errors under your direct control. And if you couldn't answer almost anything in Block 4, that's where to start: you're operating in the dark.
Question 20 tends to be the most revealing. "How long until you find out about an error?" For most companies, the honest answer is "maybe never, or only when a customer complains." In an era where AI talks about your brand at scale, with the trust users place in it, that blind spot is expensive.
Running these twenty questions by hand, today, is already a huge step — it pulls your AI presence out of the realm of "I think" and into "I saw it myself." Turning that one-off diagnosis into continuous tracking, across every model, with a baseline and alerts, is the next step. That's exactly the work Genoma automates, so that the answer to question 20 stops being "maybe never" and becomes "the same day."