Short answer: we ask real buyer questions in ChatGPT, Gemini, Claude, Grok and Perplexity, logged out, and write down exactly what each one says. Same questions every month. No screenshots taken out of context, no claims without a date.
AI answers change from day to day and from one account to the next. So any claim about “being visible in ChatGPT” means nothing unless it says which question, which platform and which day. This page explains the method behind every result AdsBeing reports, for clients and for ourselves.
1. Pick the questions buyers actually ask
We start with 15 questions a real buyer would type. For AdsBeing they include “best performance marketing consultant in Bengaluru”, “independent marketing consultant for B2B SaaS in Bengaluru”, “why are my Google Ads leads low quality” and two identity checks: “who is Mohammad Imran Hossen” and “what is AdsBeing Connect Digitally”. For a client, we build the list with them from sales calls and search data.
2. Ask in a clean session
Logged-in AI tools remember who you are. Ask Perplexity “who is Mohammad Imran Hossen” while logged in as Imran and it answers “you are”. That tells you nothing. So every check runs in a temporary or private chat, logged out where the platform allows it. If a run is contaminated by an earlier chat, we discard it and say so in the log.
3. Record what the answer says, word for word where it matters
For each question and platform we log the date, whether the brand is named, whether the person is named, which other names appear, and how the brand is described. The description matters as much as the mention. If an AI names you but calls a consultancy an “agency”, that is a problem to fix, not a win.
4. Find the source behind the answer
Most AI answers lean on pages they can find: directories, review sites, Google Maps listings, articles. When an answer is wrong, we trace which page taught it. In our own first check, the “agency” label came from third-party directory pages, not from our website. Fixing those pages is part of the work.
5. Repeat on the same day each month
One check is a snapshot. We run the same questions on the same platforms once a month and compare. A change only counts if it holds across runs.
What we will not claim
- No ranking promises. Nobody controls what an AI says.
- No “visible in ChatGPT” without the question, the platform and the date next to it.
- No logged-in or personalised runs counted as results.
- Results are judged over at least 3 months, because one month is noise.
Frequently asked questions
Which AI platforms do you check?
ChatGPT, Gemini, Claude, Grok and Perplexity. The AEO package reports on ChatGPT, Gemini and Claude.
How many questions do you track?
Up to 15 buyer questions a month in the AEO package. The free check covers 5.
Can you guarantee my brand will be recommended?
No. We can improve the pages AI tools learn from and measure what changes. We cannot promise a placement.
How long before results show?
Judge it over at least 3 months. In Imran’s employer role at Spacez, the brand appeared in ChatGPT and Gemini answers for selected Bengaluru luxury-stay searches within 3 months of this work.
Written by Mohammad Imran Hossen, senior performance marketing consultant in Bengaluru and founder of AdsBeing Connect Digitally.