You already know your competitors. You know whose trucks are on which sites, who lowballed the last big tender, whose welds come back for rework. That knowledge took twenty years to earn, and it is still worth something. But there is a new competitive scoreboard you probably have not looked at: when a buyer asks ChatGPT, Perplexity, or Google's AI Overviews who to call in your category, somebody gets named. If you have never checked who, you are competing blind in the one arena where shortlists are now made.
This is competitor analysis for the AI era: less about their price list, more about their citations. The goal is simple to say and very doable: find out exactly which competitors the engines recommend and why, then take those recommendations back.
Quick answer: AI-era competitor analysis means running the real buyer queries in ChatGPT, Perplexity, and Google's AI Overviews to see which companies get recommended in your category, then reverse-engineering why: which pages get cited, which directories the engines lean on, and what content depth the winners have. From there you close the gaps with structured data, question-format buyer guides, and consistent listings, because AI recommendations are earned through what machines can read and verify, not bought.
Why does competitor analysis look different in the AI era?
Because the shortlist is now assembled before anyone calls you. The old competitive game was won at quote time: sharp price, good rep, fast turnaround. The new game is won earlier, when a purchaser or engineer types the problem into a search box or an AI assistant and gets back two or three names. Those engines do not show ten blue links and let the buyer judge. They answer. If the answer is consistently your competitor, you lose quotes you never knew existed.
The other change: this scoreboard is completely visible. You cannot see a competitor's margins or their customer list, but you can see exactly what the AI engines say about your category, which of their pages get cited, and which sources the engines trust. Competitor analysis used to be guesswork and rumours from suppliers. Now the most decisive part of it is sitting in public, waiting for someone to look.
How do you find out which competitors AI engines recommend?
You ask the engines the way a buyer would, and you keep notes like an estimator. The process takes an afternoon:
- Build the query list. Ten to twenty questions your real buyers ask: "best [your product] for [application]", "who makes [product] in Canada", "[product] supplier near [region]", "[your product] vs [alternative]". Pull them from your inbound calls and quote requests, not from a keyword tool.
- Run them in three engines. ChatGPT, Perplexity, and Google with AI Overviews. Run each query more than once; answers vary, and the names that keep reappearing are the real incumbents.
- Record who gets named and who gets cited. Perplexity in particular shows its sources. Note which competitor pages, directories, and review profiles the engines lean on. That citation list is your target map.
- Ask about yourself directly. "What do you know about [your company]?" If the answer is thin, wrong, or confuses you with someone else, the engines lack clean information about you, and that is fixable.
If you would rather have this done properly, with the queries your buyers actually use and a prioritized fix list, that is exactly what our AI visibility audit delivers.
Why do the engines recommend your competitor and not you?
Almost never because their product is better. The engines cannot judge weld quality; they judge what they can read and verify. When we audit Canadian categories, the recommended companies share the same traits, and none of them are secret:
- Their site answers questions. Twenty pages of fitment, sizing, and comparison content beats your three-page site every time an engine needs something to cite.
- Machines can read them. Structured data, clean headings, an AI-crawler-friendly setup. Our full breakdown of why AI recommends your competitor and not you walks through each signal.
- Third parties corroborate them. Reviews naming specific products, directory listings, association memberships, a trade article or two.
- Their identity is consistent. Same name, same description, same details everywhere. Engines do not recommend what they cannot verify.
That is the whole moat, and it is shallower than it looks. Most Canadian categories in 2026 still have no deliberate AI-search leader, just an accidental one whose website happened to be readable.
| Dimension | Old-school competitor analysis | AI-era competitor analysis |
|---|---|---|
| What you study | Price lists, trade show booths, rumour | AI answers, citations, content depth, listings |
| Where the intel lives | Private: hard to get, often stale | Public: run the queries any afternoon |
| What winning looks like | Sharper quote at the table | Named on the shortlist before quoting starts |
| How fast it changes | Slowly, with reputation | In weeks, as engines index new content |
| Who can win | Usually the biggest player | Whoever publishes the most citable answers |
Want to see exactly which competitors AI engines recommend in your category?
We'll run the real ChatGPT, Perplexity, and Google AI Overview queries your buyers use, name the companies being recommended instead of you, show you which pages earn the citations, and hand you the plan to take them back.
Book Free AuditHow do you take the citations back from a competitor?
You out-publish and out-structure them, source by source. The engines have no loyalty; they cite whoever makes the answer easiest and most defensible. The reclaim sequence we run:
- Fix your own readability first. Structured data, question-format headings, llms.txt, and an open door for AI crawlers. This is one-time work and it is the prerequisite for everything else. Our AEO services page covers what this involves.
- Target their cited pages one by one. If their fitment chart gets cited, publish a better one: more models, clearer table, honest limitations. Engines prefer completeness and freshness, and most incumbent pages are neither.
- Claim the third-party sources. Get listed and complete in every directory and association the engines cited. If Perplexity leans on an industry directory you are absent from, that absence is a hole in your hull.
- Build review evidence. Steady reviews that mention specific products and applications give the engines the corroboration they need to add your name.
- Keep publishing. Depth wins. A category is taken one answered question at a time, and the library compounds while a competitor's static site ages.
Nobody honest promises you will own every answer; the engines vary and the ground shifts. What compounds is share: more queries where your name appears, more citations pointing at your pages, quarter after quarter.
How do you track competitor visibility over time?
The same way you track anything that matters in the shop: same measurements, same schedule, written down. Re-run your query set monthly, log which companies get named per query, and watch three numbers: how often you appear, how often each competitor appears, and which sources are being cited. Movement is slow week to week and unmistakable quarter to quarter. Our guide to measuring AI visibility covers the mechanics.
When we run this for clients, it lands on the weekly scorecard next to answered-call rate, speed to lead, and quotes sent, so visibility connects to the numbers an owner actually runs the business on. No vanity dashboards, just whether the machines name you more this quarter than last, and what that did to the quote log.
Frequently asked questions about AI-era competitor analysis
What is AI-era competitor analysis?
It is the practice of running the real queries your buyers ask in ChatGPT, Perplexity, and Google's AI Overviews to see which companies get recommended in your category, then studying why: which pages get cited, which directories the engines trust, and what content the winners publish. It replaces guesswork about competitors with a public, repeatable scoreboard.
How do I see which competitors ChatGPT recommends in my category?
Ask it the way a buyer would, without naming your company: best product for a given application, who makes it in Canada, supplier recommendations for your region. Run each query several times because answers vary, and note which names keep reappearing. Then repeat in Perplexity, which shows its sources, so you can see exactly which pages and directories earn the citations.
Why does AI recommend my competitor instead of my company?
Almost always because their online presence is easier for machines to read and verify, not because their product is better. The engines reward crawlable sites, question-format content with direct answers, structured data, consistent business details, and third-party proof like reviews and directory listings. A competitor with a deeper, better-structured site wins the citation even if you win on quality.
Can a competitor pay to be recommended by ChatGPT or Perplexity?
No. There is no ad product that buys a recommendation inside ChatGPT, Claude, Perplexity, or Google AI Overview answers today. Recommendations are assembled from what the engines can read and corroborate. That is good news for established companies: the position is earned, so a well-funded newcomer cannot simply outspend your reputation, but it also means you cannot shortcut the work.
How long does it take to win citations back from a competitor?
Specific citations can move within weeks: once a better, fresher, well-structured page on a specific question is indexed, engines often start citing it quickly, especially where the incumbent content is thin. Becoming a default recommendation across a category typically takes 6 to 12 months of consistent publishing, structured data, and third-party proof. The signals compound, which is why starting early matters.
How often should we re-run the competitor queries?
Monthly is the practical rhythm for an established company: often enough to catch movement and connect it to the work you shipped, rare enough that trends are real rather than noise. Keep the query set identical between runs, log who gets named per query, and review it quarterly against your quote and lead numbers.
Does AI-era competitor analysis matter for B2B manufacturers, or just consumer businesses?
It matters at least as much in B2B. Engineers, purchasers, and dealers increasingly ask AI assistants to shortlist suppliers and compare equipment, and the engines name specific companies in response. B2B categories also tend to have thinner content than consumer ones, which means the citations are less defended and a deliberate effort can take visible share faster.
Related reading
- Slow Season Marketing for Canadian Manufacturers and Trades
- Trade Shows vs Digital: Where Canadian Manufacturers Should Invest Now
- How Established Canadian Companies Choose a Marketing Agency in 2026
The bottom line on competitor analysis in the AI era
You did not lose your competitive instincts; the scoreboard moved. The competitor winning AI recommendations in your category today is usually not the best shop, just the most readable one, and everything that earned them that position is public, imitable, and improvable. Run the queries, map the citations, close the gaps, and keep publishing. The companies that do this now are building a position that gets harder to displace every quarter.
If you want the scoreboard read for you, with a prioritized plan to take the citations back, book a free fit call with AlphaPixels or start with the AI visibility audit. We will show you exactly who the machines recommend today, and what it takes for that answer to be you.