Is AI recommending your competitor?
ChatGPT, Gemini, Claude and Perplexity now shape how buyers research products. If they think your shipping is slow or your returns are hard, they send customers to your competitors. We track your brand's visibility in AI answers every week, and fix what the assistants get wrong.
The shortlist is being written before buyers reach your site.
A buyer asks an assistant "which brand has the best returns policy for running shoes in Canada?" and gets two or three names back. If yours isn't one of them, you never get the click, the visit or the chance to compete.
That isn't a future trend. In a McKinsey survey of US consumers fielded in August 2025, half said they now intentionally use AI-powered search, and 44% of those users called it their primary and preferred source of insight when buying, ahead of traditional search at 31% (McKinsey, 2025 (opens in a new tab)).
Even on Google, AI summaries change behaviour. Pew Research Center found users clicked a traditional result on 8% of visits to pages with an AI summary, versus 15% on pages without one (Pew Research Center, 2025 (opens in a new tab)).
What assistants often get wrong
- Old policies. A shipping time or return window you changed a year ago, still repeated as fact.
- Other people's words. A handful of loud reviews or forum threads standing in for your whole reputation.
- Missing facts. No clear price range, delivery area or returns information they can find and trust.
- Competitor framing. Comparison pages written by a competitor, quoted as if they were neutral.
- Wrong category. Your brand filed under the wrong use case, so you never appear for the right questions.
McKinsey's analysis found a brand's own sites often make up only 5 to 10 percent of the sources AI search references. Most of what assistants "know" about you comes from elsewhere.
Same questions. Every engine. Every week.
AI answers change with model updates, new sources and even from one run to the next. One screenshot tells you very little. A consistent weekly panel tells you where you stand and whether it is moving.
Build the prompt set
We map 50 to 150 real buyer questions: category ("best…"), comparison ("X vs Y"), problem ("how do I…"), policy ("who has free returns…"), local and brand prompts.
Run it on every engine
Each week the set runs across ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews and Copilot, several times per prompt, from a Canadian or US location as agreed.
Score what comes back
Mentions, recommendations, position, the attributes attached to you (shipping, returns, price, quality) and which sources each answer leans on.
Act on the movement
A short weekly change log, a monthly deep dive, and a prioritised fix list. Agents flag drops; a strategist decides what to do.
Metrics that map to revenue, not vanity.
| Metric | What it tells you |
|---|---|
| Mention rate | Share of answers, per engine, that name your brand at all. |
| Recommendation rate | Share of answers where you are actively suggested, not just listed. |
| Share of voice | Your mentions relative to the competitors you choose to benchmark. |
| Attribute accuracy | Whether shipping, returns, pricing, availability and other facts are stated correctly. |
| Sentiment and reasons | The reasons engines give for recommending you, or a competitor instead. |
| Cited sources | The pages, reviews, directories and articles each engine relies on, so we know where to work. |
| AI referral traffic | Visits and conversions from AI assistants in your analytics, where referrers are passed. |
What the free AI visibility audit includes.
A fast, honest snapshot of how AI assistants describe you today, and the first fixes worth making. No obligation.
- A 25-prompt snapshot across ChatGPT, Gemini, Claude and Perplexity
- Side-by-side comparison with up to three competitors you name
- Accuracy check on shipping, returns, pricing and availability claims
- The top sources the engines cite for your category
- Technical readiness: AI crawler access, structured data and product, shipping and returns markup
- Five prioritised fixes, delivered in a short written report
Ongoing program deliverables
- Weekly visibility tracking and change log by engine
- Monthly report: share of voice, accuracy, sources and actions
- Answer-first pages and FAQs written for the questions you should own
- Organization, Product, MerchantReturnPolicy and shipping structured data
- Off-site citation plan: reviews, directories, roundups, communities and PR
- Listing and profile clean-up so your facts match everywhere
- AI referral tracking set up in your analytics
Three levers move AI visibility.
Your own site
Clear, answer-first pages that state who you serve, what you sell, delivery times, return terms and prices in plain text, plus the structured data that makes those facts machine-readable. Crawlable by search and AI bots.
Your product data
Accurate feeds and policies in Google Merchant Center, and product feeds where AI shopping surfaces accept them, so price, stock, shipping and returns are current wherever assistants look.
The rest of the web
Reviews, marketplaces, directories, comparison articles, communities and press. We find the sources the engines already trust in your category and earn a fair, accurate place in them.
Research note: in the original academic study that named Generative Engine Optimization, adding citations, quotations and statistics to source content improved its visibility in generated answers on a test benchmark (Aggarwal et al., 2023 (opens in a new tab)). Results vary by engine and category, which is why we measure every week instead of promising a number.
AI visibility questions.
What is Generative Engine Optimization (GEO)?
GEO is the work of getting your brand mentioned, recommended and described accurately in answers generated by AI assistants such as ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews and Microsoft Copilot. It is also called AEO (answer engine optimization) or AI search optimization.
How is GEO different from SEO?
SEO competes for a ranked link on a results page. GEO competes to be named inside the answer itself. They overlap heavily, because many assistants retrieve pages from a search index before they answer, so we run them together. GEO adds prompt-level tracking, attribute accuracy and a bigger focus on third-party sources.
Can you guarantee ChatGPT will recommend us?
No, and you should be wary of anyone who does. No one controls how these models answer. What we can do is measure your visibility consistently, fix inaccurate or missing information, and strengthen the sources the engines rely on, then show you the trend week by week.
Which AI assistants do you track?
By default ChatGPT, Google Gemini, Claude, Perplexity, Google AI Overviews and Microsoft Copilot. We can add others, such as Grok or Meta AI, if they matter in your market.
Why track weekly? Aren't AI answers random?
Answers do vary from run to run, which is exactly why a single check is misleading. We run each prompt several times per engine, on a fixed schedule and method, and report rates rather than one-off screenshots. Weekly is frequent enough to catch model updates and new sources, without overreacting to noise.
What if an assistant says our shipping is slow or our returns are hard?
We trace where the claim comes from (an old policy page, a review site, a marketplace listing, a forum thread), correct the source where possible, publish clear and current policy information with structured data, and track whether the answer changes in the following weeks.
Do we need an llms.txt file?
It's optional. llms.txt is a proposed standard, and Google says no special files or markup are required to appear in its AI features. We add one when it is cheap and helpful, but we focus first on crawlable, clear pages, structured data and third-party sources.
Does this work for B2B and service businesses, or only ecommerce?
Both. For ecommerce the focus is usually product facts, policies and reviews. For B2B and services it is usually category definitions, comparisons, case evidence and presence in the directories and publications buyers trust.
Find out what AI tells your customers about you.
Send us your website and up to three competitors. We'll show you where you are named, where a competitor is recommended instead, and the first five fixes to make.