How to check if ChatGPT recommends your brand — a 5-minute test

You don't need a $1,000-per-month enterprise tool to find out whether AI assistants mention your brand. You can do a meaningful first-pass audit in five minutes with nothing but a browser. Here's exactly how.

By the end of this post you'll have a starting AI Visibility Score and a clear answer to the question: "Are buyers hearing my name when they ask AI for a recommendation in my category?"

Step 1 — Write down five buyer-intent queries

Pick the five questions a real buyer in your category would type into ChatGPT or Perplexity when researching whether to buy from you. Quality matters more than quantity.

Cover these four query types:

• Category recommendation — "best [category] for [persona] in 2026" (e.g. "best CRM for dental practices")
• Comparison — "[your brand] vs [competitor]" (e.g. "Salesforce vs HubSpot")
• Problem-to-solution — "how do I [problem you solve]" (e.g. "how do I reduce no-shows in my dental practice")
• Brand-aware — "is [your brand] any good"

Read each one out loud. If you can't picture a specific buyer typing it, rewrite it. Conversational beats keyword-stuffed every time.

Step 2 — Run each query in ChatGPT (with browsing on)

Open chat.openai.com, start a fresh chat, and enable web browsing if your account supports it. Run your first query exactly as written.

For each answer, capture three things in a simple spreadsheet:

• Brand mentioned? Yes or no.
• Position — primary recommendation, listed favorably in a top-3 set, mentioned as an alternative, or mentioned negatively.
• Sources cited — which URLs ChatGPT linked to in the answer (when browsing is on).

Repeat for all five queries. AI answers are non-deterministic, so for the most important queries run them 2–3 times and average the result.

Step 3 — Run each query in Perplexity

Open perplexity.ai (no sign-in needed for most queries) and rerun all five queries. Perplexity is the easiest engine to audit because it always shows its citations explicitly — you can immediately see which sources informed the answer.

Pay particular attention to the citations. If three of the five queries cite the same domain (Capterra, G2, a specific industry blog), that domain is your highest-leverage GEO target. You need to be present there, framed favorably.

Step 4 — Score yourself on the four metrics

Now translate your raw notes into the four GEO metrics:

Visibility Rate

Number of queries where your brand was mentioned ÷ 5 × 100. Below 20% means effectively invisible. 20–40% means present but inconsistent. Above 60% means category-leading.

Position Score (simplified)

For queries where you appeared, average the position weight: 1.0 if you were the primary recommendation, 0.7 if listed favorably in a top-3, 0.4 if mentioned among many alternatives, 0.0 if framed negatively.

Citation Share (rough)

Count how many citations across all queries pointed to your domain or to a third-party source where you were named favorably, then divide by total citations × 100.

Share of Voice

Total mentions of your brand ÷ (your mentions + all competitor mentions) × 100.

Step 5 — Calculate your AI Visibility Score

Blend the four into a single 0–100 number using these weights (yes, this is the same formula we use in our paid audits):

AI Visibility Index = (Visibility Rate × 0.3) + (Position Score × 100 × 0.3) + (Citation Share × 0.2) + (Share of Voice × 0.2)

Interpretation:

• 0–25: You are effectively invisible. Foundation work needed.
• 26–45: Lagging. A focused 90-day plan moves the needle fast at this level.
• 46–65: Solid baseline. Refinement and authority work compound.
• 66+: Category-leading. Defend the position and expand to adjacent queries.

What to do next based on your score

If you scored below 25, there's a clear and large opportunity. Most of your competitors are in the same boat — whoever moves first owns the category in AI search. Start with quick wins: Schema markup, FAQ pages targeted at the five queries you just tested, and claiming/optimizing third-party listings (G2, Capterra, GBP).

If you scored between 25 and 65, you have meaningful presence but inconsistent. The next move is a full 30-query audit across all four engines to find the patterns — which query types you win, which you lose, which third-party sources are doing the heaviest lifting for your category.

If you scored above 65, you're already winning. Now the work is defensive — monitor for competitive moves, expand to adjacent buyer personas, and build a moat through original research that gets cited.

The limits of a 5-minute test

This test is a directional starting point, not a complete picture. It doesn't probe Google AI Overviews or Gemini, doesn't run enough queries to be statistically meaningful, and can't track movement over time. If the score motivates you to invest, the next step is a full audit — 30 queries, 4 engines, multiple runs per query, complete citation analysis.

That's exactly what BetteRankings delivers.