How to Track If ChatGPT and Perplexity Mention You
To track whether AI engines mention your brand, write down the exact prompts your buyers type, run them on each engine on a fixed schedule, and record three things: whether your brand name appears, whether your domain is cited, and which competitors show up instead. This is not a one-time audit. It is a repeatable measurement system, and the teams that build it now will own the category before their competitors realize the game has changed.
Why does AI visibility matter more than search ranking right now?
AI answer engines are replacing the first page of search results for a growing share of buying decisions.
When a buyer asks an AI engine what the best tool for their problem is, the engine does not return ten blue links. It names two or three brands and stops. If your brand is not in that shortlist, you are invisible, full stop. Traditional SEO tells you how many people visited your site. It tells you nothing about how many buyers asked an AI about your category and heard only your competitor's name. That gap is where revenue is silently leaking.
The fix starts with measurement. You cannot improve what you do not track.
What prompts should you track?
Start with the exact questions your buyers paste into AI engines, not the keywords you rank for on Google.
There is a meaningful difference. A keyword is "project management software." A buyer prompt asks which project management tool a 50-person agency should use when moving off spreadsheets. The second form is what buyers actually type. Collect these from sales call recordings, support tickets, and your own search query reports. Aim for 20 to 30 prompts that cover your category, your use cases, and your competitors' names. These become your tracked prompt set.
Group them into three types:
Category prompts: asking what the best tools are for your category.
Use-case prompts: asking how to solve a specific problem your product addresses.
Competitor comparison prompts: asking how one competitor compares to another.
The third group is where surprises live. Buyers comparing competitors often get a recommendation that names a third option. That third option should be you.
How do you run the tracking process manually?
Run each prompt on a fixed schedule, record the raw output, and score three signals per result.
Pick a cadence. Weekly is realistic for most teams; bi-weekly is the minimum to catch meaningful shifts. For each prompt, open a fresh session on each engine. Do not use a logged-in session with conversation history, because personalization will corrupt your data. Run the prompt, copy the full response, and score it:
Brand mention: Is your brand name in the response? Yes or no.
Domain citation: Is your URL linked or named as a source? Yes or no.
Competitor presence: Which other brands appear, and in what position?
Log every run in a spreadsheet with columns for date, engine, prompt, brand mentioned, domain cited, and competitors named. After four to six weeks you will have a baseline. After three months you will see trends: which engines favor you, which prompts are your blind spots, and whether a content change you made shifted your visibility.
The manual process works. It is also tedious, inconsistent, and easy to drop when the team gets busy.
How does DRIS automate this across ChatGPT, Claude, Gemini, Perplexity and Grok?
DRIS runs your tracked prompts across multiple AI engines on a schedule and returns structured visibility data per prompt, per engine, per run.
Instead of opening browser tabs and copying responses by hand, DRIS sends each prompt to multiple AI engines, then records the output against the three signals: brand mention, domain citation, and competitor presence. Every result is stored in a decision ledger, so you can compare this week's run against last week's without rebuilding the spreadsheet.
Because DRIS connects to the tools your team already uses, the data surfaces where your team actually works, not in a dashboard nobody opens. You set the prompt list once. The agent runs on schedule. You review the delta.
The deeper value is pattern recognition across engines. Different AI engines do not weight sources the same way. A brand that appears consistently on one engine may be absent from another for the same prompt. Knowing which engine is the weak spot tells you exactly where to focus your content and citation-building work.
What signals tell you your visibility is improving?
Watch for three leading indicators before you see traffic or pipeline impact.
First, mention rate. Divide the number of runs where your brand appeared by total runs, per engine. A mention rate that climbs steadily on a specific engine over eight weeks is a real signal, not noise.
Second, citation rate. A mention without a domain citation is weaker than one with it. AI engines that cite your domain are treating your content as a source, which compounds over time. Track citation rate separately from mention rate.
Third, competitor displacement. When a prompt that previously returned only a rival brand now returns your brand alongside it, that is a win even if you are not yet in first position. Track the number of prompts where you appear alongside a named competitor versus prompts where you are absent entirely.
These three signals give you a leading indicator of revenue impact, weeks before it shows up in pipeline data.
What content changes actually move these numbers?
The AI engines that cite brands consistently are citing brands whose content directly answers the prompts buyers ask.
This is the core insight. If your tracked prompt asks how to cut video ad production costs, and your site has no page that answers that question directly, no AI engine will cite you for it. The content gap is the visibility gap. Once you know your tracked prompts, you know your content brief.
For each prompt where your mention rate is below 50%, create a page or post that answers it in the first paragraph, uses the exact language of the prompt in the heading, and includes specific outcomes with numbers where the evidence supports them. Then run DRIS again four weeks later and check whether the citation rate moved.
This is the feedback loop that compounds. Prompt tracking reveals gaps. Content fills gaps. DRIS measures whether the content worked. The loop repeats.
Related reading
Frequently Asked Questions
How often should I run prompts to track brand mentions on ChatGPT and Perplexity?
Weekly runs give you enough data to spot trends without overwhelming your team. The minimum viable cadence is bi-weekly, but anything less frequent makes it hard to connect a content change to a visibility shift.
Do I need a separate prompt set for each AI engine?
Start with one shared prompt set across all engines. The same prompts reveal how differently each engine weights sources and names brands, which is exactly the comparison you need. You can add engine-specific prompts later once you have a baseline.
What is the difference between a brand mention and a domain citation in AI engine tracking?
A brand mention means the AI named your company in its response. A domain citation means it linked or named your URL as a source. Citations carry more weight because they signal the engine is treating your content as authoritative, not just recognizing your name from training data.
Can tracking brand mentions on ChatGPT and Perplexity replace traditional SEO monitoring?
No. The two systems measure different things. Traditional SEO monitoring tracks clicks and rankings on search engines. AI visibility tracking measures whether your brand appears in answers generated by AI engines. Both matter, and the teams winning right now are running both in parallel.
How does DRIS handle prompt tracking across multiple AI engines at once?
DRIS sends each tracked prompt to multiple AI engines on your set schedule, records the structured output, and logs brand mentions, domain citations, and competitor appearances per run. The result is a consistent dataset you can trend over time without manual copy-paste work.
What should you do next?
Write your first 10 tracked prompts today, run them manually on the AI engines your buyers use, and see where you stand. Then book a demo with Nagent to see how DRIS automates the full process across all five major AI engines.