TL;DR: Want to know what ChatGPT and Gemini say about your business? Copy the ten prompts below, replace the brackets with your details, and run them in a logged-out browser. In ten minutes you’ll know whether AI recommends you, describes you accurately, and how you compare to competitors. Log the results and repeat monthly.
Before you start (30 seconds)
Open ChatGPT and Gemini in a logged-out or incognito window you want the answers a stranger gets, not ones shaped by your history. Keep a note open to record who gets named. Replace [category], [city], [area], [business name] and [competitor] with your real details as you go.
The 10 prompts
Visibility — are you in the answer?
Best [category] in [city]Recommend a reliable [category] near [area]I need [the problem you solve] in [city]. Who should I contact?Top 5 [category] in [city] and what each is known for
Accuracy — is the AI right about you?
What is [business name] in [city]?What services does [business name] offer and what do customers say?Is [business name] trustworthy? What should I know before using them?
Competition & the money question
Compare [business name] and [competitor]. Which is better and why?Why do people choose [competitor]?I have [budget] for [category] in [city]. Give me ONE recommendation and why.
Bonus: 3 prompts for your industry
If you’re local, add these, they surface how the AI handles “near me” intent, which is where most Pakistani buying questions live:
[category] near me open now(test from your actual location)affordable [category] in [city] with good reviewsis [business name] still open? what are their hours?
Prompt 13 is a quiet trap-detector if the AI reports wrong hours or “permanently closed,” you’ve caught a live problem worth fixing today.
How to read your results in 2 minutes
Prompts 1-4 and 10 (visibility): Named in most? Strong protect it. Named sometimes? On the bubble small improvements tip you in. Never named while competitors are? You’ve found a silent leak. Nobody local named? Your market’s AI space is empty, first mover wins it. Prompt 10 matters most forcing a single pick reveals your true ranking.
Prompts 5-7 (accuracy): Every wrong fact is an urgent fix. AI errors trace to inconsistent public information an old listing, a mismatched profile, a stale page. Note each error and where it likely comes from.
Prompts 8-9 (competition): Write down every strength the AI credits to competitors reviews, specific services, years established. That list is your to-do list: those are the signals the machines reward in your category.
What to do with what you find
The prompts diagnose; the fixes are elsewhere on this blog. Absent or unreadable? Work through the 7 reasons you’re invisible on ChatGPT and the AI Visibility Framework. Visible but never quoted? Answer-first writing and schema. Wrong facts? The correction playbook. Losing to competitors? Build Google Business Profile and review authority.
Then do the one thing most people skip
Save your results with today’s date, and put a reminder to run these again in 30 days. A single test is a snapshot; the monthly log is the trend line that shows whether your fixes are working and catches problems early. Ten minutes a month is the cheapest business intelligence you’ll ever gather the fuller version is the 30-minute AI brand audit.
Prefer we run it for you?
DigiMSM will run these prompts plus the deeper technical checks and hand you a prioritized plan free. Pakistan’s 1st AI-powered digital marketing agency. WhatsApp +92 335 4155677 or digimsm.com/contact-us/.
Published by the DigiMSM editorial team, Islamabad. Last updated July 2026.