The fastest way to see which competitors AI search engines recommend instead of you is to combine two things: a tracking tool that runs a fixed set of prompts on a schedule and logs every brand named in the answers, and a manual prompt-sampling check you can run today for free. Trackers give you trends and share-of-voice numbers. Manual sampling gives you the actual wording, the cited sources, and the context behind each mention. Used together, they show you who is winning the recommendation, why, and what content you need to publish to get into the answer.
This guide covers the types of tools available, a step-by-step free method, how to read the results without fooling yourself, and how to turn what you find into a content plan.
Why "who gets named instead of me" is a different question than rankings
In traditional search, a competitor outranking you is visible: you check a results page and count positions. In AI answers, there is no page one. ChatGPT, Claude, and Perplexity each generate a response that might name three brands, five brands, or none, and the answer can change from one run to the next.
That changes what you measure. Instead of position, you track:
- Mention rate: the percentage of relevant prompts where a brand is named at all.
- AI share of voice: your mentions divided by total brand mentions across the same prompt set, compared with each competitor.
- Competitor citations: the specific URLs the engine cites as sources when it recommends someone else.
- Recommendation context: whether a brand is listed as the top pick, a runner-up, or a passing mention.
The competitor citations are the most actionable piece. If an engine keeps naming a rival and linking to the same comparison article or directory, you've found a source the model trusts, and a place you may need to appear.
The tool types that reveal competitor mentions
You don't need a specific product to understand the landscape. Tools in this space fall into a few categories, each with trade-offs.
1. Dedicated AI visibility trackers
These platforms run a library of prompts against multiple AI engines on a schedule, parse the answers, and report which brands appear. Typical outputs include mention rate, share of voice by competitor, sentiment, and the sources cited.
Best for: ongoing monitoring, trend lines, and reporting to a team. Watch out for: prompt sets that don't match how your real customers ask questions, and pricing that scales with the number of prompts or engines tracked.
2. SEO suites with AI add-ons
Several established SEO platforms now include an AI-visibility module alongside keyword and backlink data. The advantage is having everything in one dashboard; the limitation is that the AI features are often newer and shallower than in dedicated tools.
Best for: teams already living in an SEO suite who want a baseline. Watch out for: limited control over the exact prompts and locations being sampled.
3. Managed AEO services with built-in tracking
Some services bundle tracking with the work of fixing what the tracking reveals. At AEO Juice, for example, the Pro and Prime tiers include weekly LLM-visibility tracking alongside an automated content calendar and AI-generated fixes, so the data points straight to the next action instead of sitting in a dashboard.
Best for: founders and small teams who don't have time to interpret data and then plan content separately. Watch out for: make sure you can still see the underlying prompts and answers, not just a score.
4. Manual prompt sampling (free)
You ask the engines your customers' questions yourself and record what comes back. It's slower and less statistically tidy, but it costs nothing and gives you the richest qualitative detail. It's also the best way to validate whatever a paid tool tells you.
5. Brand-mention and referral analytics
Your web analytics can show visits referred from chat.openai.com, perplexity.ai, and similar sources. This won't reveal competitors, but it shows whether AI engines are sending you traffic at all, which is a useful reality check next to your competitor data.
The free method: a 60-minute prompt-sampling audit
Here's a repeatable process you can finish in about an hour. All you need is a spreadsheet and access to ChatGPT, Claude, and Perplexity.
Step 1: Write 15–20 realistic prompts
Pull them from how buyers actually talk. Good sources are sales call notes, support tickets, Reddit threads, and the "People also ask" boxes in Google. Mix three types:
- Category prompts: "What are the best invoicing tools for freelancers?"
- Problem prompts: "How do I stop chasing late client payments?"
- Constraint prompts: "Which bookkeeping software works for a two-person agency under $30 a month?"
Avoid putting your own brand name in the prompts. The goal is to see who gets named when nobody is steering.
Step 2: Run each prompt three times per engine
AI answers vary. A single run can mislead you, so run each prompt three times on each engine, ideally in fresh conversations without prior context. With 15 prompts, three engines, and three runs, that's 135 answers. If that's too many, cut to 10 prompts and two runs. Consistency matters more than volume.
Step 3: Log everything in a simple sheet
Create columns for:
- Prompt
- Engine
- Run number
- Brands named (in order)
- Your brand mentioned? (yes/no)
- Cited sources (URLs)
- Notes (e.g., "listed as top pick," "described as expensive")
Step 4: Tally mentions and sources
Count how many times each brand appears, then divide by the total number of answers to get mention rate. Sum all brand mentions to calculate rough AI share of voice. Then list every cited URL and count which domains show up most.
Step 5: Repeat monthly
Same prompts, same method. The first pass is your baseline; the real value comes from seeing movement over time.
How to read the results without fooling yourself
Raw counts are easy to misread. Keep these points in mind.
Treat small samples as directional. If a competitor appears in 8 of 45 answers and you appear in 5, that's a signal to investigate, not proof of a precise gap.
Separate engines. Perplexity leans heavily on live web sources and shows citations by default. ChatGPT and Claude may draw more on training data unless browsing is active. A competitor dominating one engine but absent from another points to different causes.
Look at position and framing, not just presence. Being named first with a clear reason ("best for small teams") is worth more than being the fifth item in a list.
Notice who's missing. If the same two competitors appear in nearly every answer and a third one you worry about rarely does, your real AI-era competitive set may differ from your sales-team one. Sometimes the brands AI names are directories or publishers rather than direct rivals, and those are content opportunities, not threats.
Check for errors. Engines sometimes describe a company with outdated pricing or features. Note inaccuracies about your own brand; they point to stale pages the model is learning from.
Turning competitor citations into a content plan
This is where the audit pays for itself. Sort what you found into four buckets.
Bucket 1: Third-party sources that name competitors but not you
If the same listicle, review site, or directory appears again and again in citations, pursue inclusion. That might mean submitting your product, pitching an update to the author, or earning a review. These pages feed AI answers directly.
Bucket 2: Questions where competitors win on their own site
When an engine cites a competitor's blog post or help article, read it. Note the format (step-by-step, table, definition up front), what it answers, and what it leaves out. Then publish a clearer, more complete answer on your own site. Lead with the direct answer in the first two sentences, use specific numbers and examples, and structure it with clear headings.
Bucket 3: Prompts where nobody is clearly winning
Answers that name generic or inconsistent brands signal open ground. These are often your fastest wins, because there's no entrenched source to displace.
Bucket 4: Prompts where you appear but are described poorly
Update the pages that shape that description: your homepage, product pages, and about page. State plainly who you serve, what you cost, and what you do best. Vague positioning produces vague mentions.
Build the calendar
Rank the opportunities by two factors: how closely the prompt matches buying intent, and how winnable the gap looks. Then schedule one piece of content per week, each targeting a specific prompt. If that cadence sounds heavy, it's exactly what an automated content calendar is meant to handle. AEO Juice's free 26-check AEO report is a good way to see where your site stands on the fundamentals before you plan.
What to track each month
Keep your scorecard short so you'll actually maintain it:
- Your mention rate across all prompts, and by engine
- Your AI share of voice versus your top three competitors
- Number of answers where you're the first brand named
- New cited sources that include you
- Inaccurate descriptions found and fixed
Improvement usually shows up first in citations and Perplexity results, since that engine reads the live web, and later in engines that rely more on training data.
FAQ
Can I see which competitors ChatGPT recommends for free?
Yes. Write realistic buyer prompts, run each one several times in fresh conversations, and log the brands named. It takes about an hour for a solid first pass and requires only a spreadsheet.
How many prompts do I need for a reliable picture?
Start with 15–20 prompts across category, problem, and constraint types. Run each at least twice per engine. More prompts improve reliability, but consistency over time matters more than a huge first sample.
Why do AI answers change each time I ask?
These models generate responses probabilistically, and some engines pull in live search results that shift. That's why repeated runs and monthly re-sampling are essential. Treat any single answer as one data point.
What is AI share of voice?
It's the proportion of brand mentions you earn compared with competitors across a fixed set of prompts. If 100 total brand mentions appear and 20 are yours, your share of voice is 20%.
Do paid trackers replace manual checking?
Not entirely. Trackers automate scale and trend reporting, while manual sampling shows the full wording and reasoning. A quick manual spot-check is a good way to confirm a tool's data matches reality.
How long until content changes affect AI recommendations?
It varies by engine. Engines that search the live web can reflect new content within days or weeks. Others change more slowly. Plan on tracking for at least two to three months before judging results.
The short version
Find out who's being named in your place by running real buyer prompts across ChatGPT, Claude, and Perplexity, logging brands and cited sources, and repeating monthly. Use a paid tracker when you need automation and trend lines. Then act on what you find: get into the sources that already name your rivals, answer the questions they answer better, and claim the prompts nobody owns yet.
Want a head start? Run the free 26-check AEO report at aeojuice.com to see where your site stands, and squeeze your way into the answer from there.