I launched a new product two months ago. How do I know when AI assistants start recommending it?

There’s no single notification when an AI assistant begins recommending your product—and a referral visit only tells you someone clicked through, not exactly what they asked. Track recommendations, citations, and resulting customers separately.

Start with a simple weekly check:

1. Write down 10–20 realistic buyer questions, such as “What’s a good [category] for [use case]?” Include non-branded questions and comparisons with competitors.
2. Run the same questions in the assistants your customers use, in fresh chats, once a week. Record whether your product is mentioned, recommended, linked or cited—and which competitors appear. Answers can vary, so treat this as a trend, not a definitive ranking.
3. Track visits and conversions from AI sources in your analytics. Look for referrals from ChatGPT, Perplexity, Claude, Gemini and Copilot, then compare sign-ups or purchases—not just visits. OpenAI says ChatGPT Search referral links include utm_source=chatgpt.com; other sources may be less consistently identifiable. (help.openai.com)
4. Check webmaster reporting. Bing Webmaster Tools’ AI Performance report shows pages cited in supported AI answers and visibility trends. Google’s Search Console reporting covers generative AI features in Google Search; that’s useful for search visibility, though it’s not a complete view of every assistant. (bing.com)
5. Ask new customers how they found you. Add “AI assistant” as an option to your signup or purchase survey; analytics can miss visits when referral details aren’t passed along.

Since you launched two months ago, establish a baseline now. A useful weekly snapshot is: tested prompts, recommendation rate, citations, AI-attributed visits, and conversions. The clearest sign that assistants are recommending you is repeated appearances in relevant buyer prompts; the clearest sign it’s paying off is those appearances leading to qualified visits and customers.

You won't get an automatic alert from the AI assistants themselves, so you have to check on purpose. The most reliable way is a prompt set you run on a schedule, either by hand or with a tracking tool.

Manual checks (free, but noisy)
Ask the assistants the questions your buyers would ask, such as "best [category] for [use case]," and see whether your product appears. AI answers vary from one run to the next, so a single manual query won't tell you how often you appear, in what context, or how you stack up against competitors over time. Run the same prompts repeatedly, across several locations if relevant, and log the results in a spreadsheet.

Dedicated tracking tools (automated)
These tools run your prompts on a schedule and record mentions. Dedicated trackers run many prompts on a schedule and average the results, which turns a one-off impression into a metric you can actually monitor. Some also send alerts. For example, Beamtrace says it can send real-time alerts when your brand appears in new AI-generated answers, new topics, or competitive scenarios. Pricing varies. Otterly AI starts at $29/month with a 7-day trial, and most dedicated trackers offer a 14-day free trial with no card. Semrush, Peec AI, and Ahrefs Brand Radar are other common options.

What to look for

• Distinguish mentions from citations. Track mentions separately from citations, because the two often differ. A mention means the assistant named you. A citation means it linked to your site.
• Read the answers, not just the counts. A tool's count is only meaningful once you've checked what the answer actually says about you.
• Track name variations, such as the product name, company name, and common misspellings.
• Watch for inaccurate descriptions, not just absence. Some tools can show how each assistant describes your product and what sources it relied on.

Set expectations on timing
Two months is early. AI systems update on their own schedules, and some tools refresh monthly, so your monitoring should follow those cycles. Expect changes to show up slowly. Also, recommendations often depend on what third-party sites say about you, so reviews, comparison articles, and listings help.

A simple setup

1. Write 20 to 25 buyer-style prompts.
2. Run a baseline check now and record the results.
3. Re-run the same prompts monthly, or weekly if you use a tool.
4. Track mentions, citations, and how you are described, and note which competitors appear alongside you.

Separately, check your website analytics for referral traffic from ChatGPT, Perplexity, and similar sources. That won't show every recommendation, but it's a useful signal that assistants are sending people to you.

Knowing whether AI assistants are recommending your product requires a shift from traditional SEO thinking to "Answer Engine Optimization" (AEO). Because AI responses are non-deterministic (they change from run to run), a single manual search isn't enough to give you a reliable answer.

Here is how to track your brand’s AI visibility effectively.

1. The "Do-It-Yourself" Method (For Initial Diagnostics)

If you are just starting, you can run a manual diagnostic to get a baseline. Do not search for your brand name directly (which is biased); instead, search for the problems your product solves.

• Create a set of "Buyer Prompts": Draft 10–20 questions your ideal customer would actually ask.
  • Category questions: "What are the best [your product type] under $[price]?"
  • Use-case/Problem questions: "How do I solve [specific problem your product fixes]?"
  • Comparison questions: "Compare [Your Brand] with [Competitor]."
• Use Fresh Sessions: Use private/incognito windows for every query to prevent your personal search history or past conversations from influencing the results.
• Record in a Spreadsheet: Track:
  • Mentioned? (Yes/No)
  • Recommended? (Did it suggest you as a solution?)
  • Cited? (Did it link to your website?)
  • Who else appeared? (Which competitors are "winning" the prompt?)

2. The Professional Approach (Ongoing Monitoring)

Because AI results vary over time, manual tracking will eventually become unsustainable and statistically unreliable. To get a true trend line, consider dedicated tools that automate this process by querying multiple models (ChatGPT, Claude, Gemini, Perplexity) consistently.

• AI Visibility Tools: Tools like Botric, MentionBird, Peec AI, or Semrush’s AI toolkit are designed specifically for this. They systematically track thousands of prompts, providing you with a "Visibility Score" and alerts when your status changes.
• What to Look For:
  • Citation Sources: The most valuable insight these tools provide is which websites the AI is citing when it mentions you (or your competitors). This tells you exactly where you need to build more authority to improve your recommendations.
  • Negative Framing: See if the AI mentions you but includes caveats or mentions competitor strengths over yours.

3. Immediate Technical Checklist

Before investing in tools, ensure your website is actually "readable" by AI:

• Check robots.txt: Ensure you haven't blocked AI crawlers (like GPTBot, ClaudeBot, or PerplexityBot) from accessing your site.
• Optimize Structured Data: AI models rely heavily on the structured data (schema) in your page code—price, stock status, ratings, and clear product descriptions. If you aren't sending this data, the AI may skip you because it lacks the confidence to recommend you.
• Monitor "How did you hear about us?": Add this field to your signup or checkout form. If people start selecting "AI" or "ChatGPT," you have a qualitative business signal that confirms your visibility efforts are paying off commercially.

Summary Recommendation: Start by manually testing 10 buyer prompts in a spreadsheet this week to see where you stand. If you find you are being missed or misrepresented, look into an AI visibility tool to help identify the "source sites" you need to target to earn more influence.

The best way is to treat AI recommendations as something you measure repeatedly, not something you can reliably detect from a single answer. Responses vary by platform, prompt, and sometimes session, so establish a baseline and watch for a trend.

1. Create a short list of buyer questions—around 15–25 is a useful starting point. Include discovery prompts (“What are the best options for…?”), comparison prompts (“How does [product] compare with…?”), and specific use cases your product serves. ¹⁶
2. Check the main AI assistants your customers use, such as ChatGPT, Gemini, Perplexity, and Google AI Overviews. Run the same prompts periodically, preferably in logged-out or fresh sessions to reduce personalization effects. ¹¹⁰
3. Log the results: date, assistant, prompt, whether your product was mentioned, its position, how it was described, competitors named, and any cited sources. Repeat prompts more than once when possible, since answers can differ between runs. ⁵¹¹
4. Track a few simple metrics over time:
   • Mention rate: share of tested prompts that mention your product.
   • Recommendation rate: share that put it forward as a choice, ideally among the first few options.
   • Share of voice: your mentions compared with competitors.
   • Citations: whether assistants cite your site or other sources about your product. ²⁶¹¹

A spreadsheet is enough to get started. If manual checks become a burden, AI-visibility monitoring tools can rerun prompts and compare mentions, citations, and competitors across supported platforms. ¹¹²

Also check your website analytics for visits and conversions from AI-related sources, and Google Search Console for AI-feature search traffic. These can show whether visibility is bringing people to your site, but they won’t tell you every time an assistant mentioned your product. ³

Since you launched two months ago, start tracking now and keep the prompt set and testing schedule consistent. That will give you a useful baseline; a one-off answer is only a snapshot, not proof of a lasting recommendation trend.