How to see which AI search prompts mention my competitors instead of us

To see which AI search prompts mention your competitors instead of you, build a fixed list of the questions your buyers ask, run each one across ChatGPT, Gemini, Perplexity and the others several times, and record every brand named and every site cited. Then filter for prompts where a competitor appears and you do not. A spreadsheet works for 20 prompts; past that, an AI visibility tracker with a competitor gap view saves hours each week.
What exactly am I looking for?
"Mentioned instead of us" covers four different gaps, and each needs a different fix.
Named, you absent. The answer recommends a competitor and never names you.
Both named, they rank first. You appear, but lower in the list or as an afterthought.
You named, not linked. The answer names you but cites a competitor's page or a third-party article as its source.
Cited source favours them. The answer links a comparison article or review page where the competitor is the top pick.
Keep mentions (your name in the text) and citations (your site among the sources) as separate columns. Many tools count only one of them.
How do I build the prompt list?
The list decides what you can see, so spend time here.
Start from real demand. Pull question-style queries from Google Search Console, the questions in sales and support calls, and People Also Ask boxes for your category.
Write them the way people ask assistants. Longer, with context: "best payroll software for a 40-person company in India" rather than "payroll software".
Cover the buying intents. Category ("best X for Y"), comparison ("X vs Y"), pricing, problem ("how do I fix Z"), integration and vendor evaluation.
Keep most prompts unbranded. A branded prompt ("is Acme good?") tells you little about discovery. Peec AI suggests branded prompts be about 15% to 20% of a set (Peec AI docs).
Size it to your capacity. 30 to 50 prompts is enough to start. Group them by product line or persona so a gap points to a team.
Freeze the list for at least a quarter. Changing prompts and judging results in the same month makes trends meaningless.
How do I check by hand?
For a first pass with no budget:
Use clean sessions. Sign out, or use a private window, so your history does not tilt the answer.
Ask each prompt in each engine you care about: ChatGPT, Gemini, Perplexity, Claude, Grok, Copilot and Google's AI Overviews and AI Mode.
Run each prompt three times on different days. Answers vary, and one run proves little.
Log every answer in a sheet with these columns: prompt, engine, date, brands named in order, sites cited, you named (yes or no), you cited (yes or no), top competitor.
Filter for rows where a competitor is named and you are not. Count how many runs out of three each prompt shows that pattern.
The result is a list of prompts ranked by how consistently a competitor wins. Fifty prompts across six engines and three runs is 900 answers, which is why most teams move to a tool after the first month.
How do I read the results?
Look for patterns before you look at single prompts.
By intent. Losing all the comparison prompts points to missing comparison pages. Losing pricing prompts points to unpublished prices.
By engine. Winning on Gemini but losing on ChatGPT often means the sources each engine trusts differ. Profound's July 2026 analysis found brand sites were 69% of Gemini citations against 47% on ChatGPT, where earned media carried more weight (Profound).
By cited source. If the same three review pages or listicles appear whenever a competitor wins, those pages are your target.
By position. Peec AI's win rate is a useful idea here: the share of answers where a brand is ranked first, which can never exceed how often it appears (Peec AI docs). High visibility with a low win rate means you are present but not preferred.
Which tools show competitor gaps, and what do they cost?
All prices were checked on 24 October 2026 on each vendor's own site.
Semrush AI Visibility Toolkit
What it is: AI visibility module inside Semrush, with Competitor Research.
Best for: teams already on Semrush who want gaps beside SEO data.
Gap view: Competitor Research identifies "the prompts and topics where others are getting cited but you aren't" (Semrush KB).
Pricing: $99 a month with 25 tracked prompts, no free trial; Semrush One Starter $199 a month with 50 prompts tracked daily (Semrush pricing).
Limitation: Prompt Tracking names Google AI Mode and ChatGPT as examples; confirm the engines you need.
Ahrefs Brand Radar
What it is: an AI visibility index of hundreds of millions of prompts plus custom prompts.
Best for: researching any brand's AI presence without building a prompt list first.
Gap view: AI Share of Voice benchmarks you against competitors; the page does not describe a per-prompt "they appear, you do not" filter (Ahrefs Brand Radar).
Pricing: Brand Radar AI add-on from $199 a month; custom prompt packages from $50 a month (Ahrefs pricing).
Limitation: each prompt uses one check per platform, location and update, so costs scale with frequency.
Peec AI
What it is: AI search analytics with competitor visibility, position, share of voice and win rate.
Best for: SEO and content teams who want per-prompt competitor detail.
Gap view: suggests competitors once they appear twice beside you, and its Actions list shows competitor evidence and sources (Peec AI docs).
Pricing: Starter, Pro and Advanced plans are listed, but no figures appeared on its pricing page when we fetched it (Peec AI pricing).
Limitation: confirm prices and engines in a demo.
Profound
What it is: enterprise AI visibility platform with Answer Engine Insights and agents.
Best for: large brands tracking many engines and regions.
Gap view: tracks competitive presence across engines.
Pricing: free 7-day trial with 50 prompts on ChatGPT, Gemini and AI Overviews; Enterprise custom with up to 9 engines (Profound pricing).
Limitation: trial prompts cannot be customised.
AthenaHQ
What it is: AI visibility platform priced in credits.
Best for: teams wanting many engines on one self-serve plan.
Gap view: "Unlimited competitor tracking, impersonation, and benchmarking" on Starter.
Pricing: Essential free with 300 credits; Starter $295 a month with 3,600 credits, where 1 credit is one AI response (AthenaHQ plans).
Limitation: credits run down faster as you add engines and prompts.
HubSpot AEO
What it is: a standalone AEO tool from HubSpot.
Best for: HubSpot users who want a low-cost start.
Gap view: competitor share of voice and citation analysis on ChatGPT, Gemini and Perplexity.
Pricing: $50 a month, or $45 a month billed annually, with 25 prompts (HubSpot AEO).
Limitation: three engines.
What do I do with a gap once I find it?
Open the sources the answer cited when the competitor won. They tell you what to fix.
The competitor's own page is cited. Compare it with yours. Usually it states the answer more directly, covers the exact use case or publishes a price you do not. Write the better page.
A listicle or comparison article is cited. Ask to be considered for the next update, with facts the author can check.
A review site is cited. Build up detailed reviews there and correct your profile.
A forum thread is cited. Answer genuinely where you have something useful to add, as a named company account.
Nothing is cited. The answer comes from the model's memory. That moves slowly; broad, consistent coverage over months is the fix.
Then re-run the same prompts four to six weeks later and compare.
What mistakes make the data misleading?
Counting raw citations as the panel grows. More prompts produce more citations without you being more visible. Compare rates, not counts.
One run per prompt. A single answer is a sample of one.
Brand names that are ordinary words. A tool matching "Pilot" will count every pilot. Add aliases and check matches.
Ignoring the untracked winner. The brand beating you may not be on your competitor list. Look at every brand named, not only the ones you track.
Mixing locations and languages. Answers differ by country. Keep each prompt set to one market.
Where does Nagent fit?
We make Nagent. Its DRIS agent is built around this exact question.
A fixed prompt panel. You accept the prompts per page DRIS watches; each carries its intent and where it came from (Search Console, People Also Ask, written by hand or by DRIS). Only accepted prompts are asked, across ChatGPT, Claude, Gemini, Perplexity and Grok.
Answer-engine gaps. Visibility lists questions where competitors appear and you do not.
Beat us on. For each competitor, it counts the prompts where an answer cites them and no answer cites you, naming the first few.
Untracked winners. It shows the ten sites cited most beside you that are not on your list, and, in Beta, the brands answers name without linking.
Fixes for approval. It proposes page fixes, and a person approves any change to a live page.
What it does not do: it does not chase third-party reviews or outreach for you, and how often the engines are asked depends on your plan. An AEO page audit is about 10 credits; see pricing.
Frequently Asked Questions
How do I track whether ChatGPT and Perplexity mention my brand?
Fix a list of buyer prompts, run them on a schedule, and log whether each answer names you, links you and who it names instead. Track rates over weeks, not single answers.
How do I find the prompts my buyers ask AI assistants?
Start from Search Console question queries, sales and support call notes and People Also Ask, then rewrite them as full questions with the buyer's context. Several trackers also suggest prompts from your category.
Profound vs Peec AI vs AthenaHQ for tracking AI citations: which suits whom?
Profound suits enterprises that need many engines and regions. Peec AI suits SEO teams wanting per-prompt competitor detail. AthenaHQ suits teams wanting a self-serve plan across many engines, priced in credits.
How accurate is competitor data from AI visibility tools?
Directionally useful, never exact. Answers vary by run, location and model version, so trust trends across many prompts and weeks more than any single figure.
Want to see the prompts where competitors win in ChatGPT, Gemini and Perplexity, with fixes ranked? Book a Nagent demo and ask about DRIS.