How to Check If AI Is Recommending Your Brand (And What to Do If It's Not)

There's a new channel sending customers to your competitors and you might not even know it exists.

More and more product and service research now happens in AI assistants like ChatGPT, Gemini, Claude, and Perplexity. People aren't just Googling anymore. They're asking AI "what's the best CRM for small teams" or "which running shoes are good for flat feet" and buying whatever comes back.

The question is: does AI recommend you?

Many businesses have no idea. They've never checked. They're optimizing for Google rankings while a growing chunk of their potential customers are getting recommendations from a completely different system, one that works on completely different rules.

Here's how to find out where you stand, what the results mean, and what to do about it.

Step 1: Run the Manual Test (5 Minutes)

Open all three major AI assistants. You need to check each one separately because they have different training data and give different recommendations.

• ChatGPT: chatgpt.com
• Gemini: gemini.google.com
• Claude: claude.ai

The Prompts to Use

Don't ask "do you know about [my brand]"; that's a recognition test, not a recommendation test. Instead, ask the prompts your customers would ask. These fall into three categories:

Discovery prompts (top of funnel): People just starting to explore your category.

• "What are the best [your category] brands?"
• "Top rated [your product type] in 2026"
• "Best [your category] for [your target audience]"

Consideration prompts (middle of funnel): People comparing options.

• "Alternatives to [well-known competitor] for [your category]"
• "[Your category] buying guide: what should I look for?"
• "Best [your product type] under [price point]"

Branded prompts (bottom of funnel): People who've already heard of you.

• "Is [your brand] worth buying?"
• "[Your brand] vs [competitor]: which is better?"
• "[Your brand] reviews: are they actually good?"

Run 2-3 prompts from each category across all three AI platforms. That's about 20 queries. It takes five minutes and you'll have a clear picture.

What to Write Down

For each query, note:

• Did AI mention your brand? (yes/no)
• What did it say about you? (positive, neutral, generic, or wrong?)
• Which competitors got mentioned instead?
• Did AI cite a source or just state things as fact?

You're building a map of your AI visibility. Many businesses are surprised by what they find: either they don't show up at all, or the AI says something generic and unhelpful about them while giving competitors specific, glowing recommendations.

Step 2: Understand Your Results

If You Showed Up in 0 Out of 9 Discovery/Consideration Queries

This is a common result for small and mid-size businesses. It usually means AI doesn't have enough content about your brand to form a recommendation.

The root cause is usually one of these:

• Your website content is too thin. Product descriptions under 100 words, no blog, no buying guides.
• You don't exist outside your own website. No press mentions, no review site presence, no Reddit discussions.
• AI crawlers can't read your pages. Your robots.txt blocks them, or your key text only appears after JavaScript runs.

If You Showed Up in Branded Queries Only

This means AI knows you exist but doesn't consider you a category recommendation. When someone asks about you by name, it can pull together a response. But when someone asks "best X for Y" without mentioning your name, you don't make the cut.

This usually means you have brand recognition but not topical authority. Your website talks about your products but doesn't establish expertise in the category. You're missing the content that positions you as a go-to recommendation.

If You Showed Up But the Information Was Wrong or Outdated

AI models are trained on historical data and may have outdated information about your brand. If ChatGPT says you're based in a city you left two years ago, or mentions a product line you discontinued, it's pulling from old sources.

The fix is fresh, current content on your own site. AI models increasingly use web search to supplement their training data. When your website clearly states current information, the model's web search can correct the outdated training data.

If a Competitor Showed Up Consistently and You Didn't

Pay attention to which competitors get cited. Look at their websites. You'll almost always find:

• Longer, more detailed product descriptions
• Blog content that answers buyer questions
• Press or review coverage on third-party sites
• Structured data on their pages

The competitor getting cited isn't necessarily better than you. They just have more content for AI to work with.

Step 3: Calculate Your Citation Rate

Count up your results. If you ran 9 queries across 3 platforms (27 total checks), and your brand appeared in 4 of them, your citation rate is about 15%.

Here's a rough way to interpret that number. These bands are a rule of thumb, not an industry benchmark:

A low number on your first check is not a failure. It's a starting point, and the fixes below are within reach of a small team.

Step 4: Fix the Gaps

Based on your results, here's what to prioritize.

If Your Citation Rate Is Under 5%: Fix Your Foundation

Your website doesn't have enough substance for AI to work with. Start with the basics:

Rewrite your product or service descriptions. Every key page needs at least 150 words of genuine, helpful content. Not marketing fluff: actual information about who this is for, what makes it different, and why someone should choose it. Write like you're explaining to a smart friend, not writing ad copy.

Add FAQ sections to your main pages. Think about the questions customers actually ask your sales team or customer support. Put those questions and honest answers directly on your product pages. A clear question followed by a direct answer is easy for AI systems to quote.

Start a blog with 3 posts. Write about:

1. A buying guide for your category ("How to choose the right [product type]")
2. A comparison post ("X vs Y: which is better for [use case]")
3. A problem-solution post ("How to fix [common problem your product solves]")

These three posts cover the main query types AI gets from buyers. They don't need to be long: 800-1,200 words each, focused and practical.

If Your Citation Rate Is 5-15%: Build Depth

AI knows you exist but doesn't have enough reason to recommend you over alternatives. You need more content signals.

Add structured data (schema markup) to your pages. If you're on Shopify, apps like JSON-LD for SEO make this easy. If you're on WordPress, plugins like Yoast or RankMath handle it. The key schemas to add:

• Product schema on product pages
• FAQPage schema on pages with Q&As
• Article schema on blog posts
• Organization schema on your homepage

Schema will not get you recommended on its own. Google no longer shows FAQ rich results, and its guidance says no special markup is needed for its AI features. What schema does is give machines your key facts in a standard format.

Publish content that positions you as a category expert. Not promotional content about your products; educational content about your space. If you sell coffee equipment, write about brewing methods, water temperature, grind sizes. If you sell running shoes, write about training plans, injury prevention, race prep. This builds topical authority, which makes you a more credible recommendation.

Get mentioned outside your own site. Reach out to bloggers who write "best of" roundups in your category. Answer questions on Reddit where your product type comes up naturally. Contribute guest posts to industry publications. Every third-party mention gives AI systems another independent source about you.

If Your Citation Rate Is 15-30%: Optimize and Expand

You're already showing up. Now it's about consistency and filling in the gaps where competitors still beat you.

Look at which specific queries you're missing. If you show up for branded queries but not discovery queries, you need more top-of-funnel content. If you show up for broad category queries but not specific use-case queries, you need more targeted pages.

Expand your content to cover more buyer intents. Every time you find a query where a competitor gets cited and you don't, that's a content gap. Write a page or blog post that directly addresses that query.

Keep your content fresh. Update your key pages whenever pricing, features, or FAQs change, so the facts AI systems pick up about you are current.

How to Automate This

The manual check gives you a snapshot, but AI recommendations change as models update. You should be monitoring this regularly.

XeoRank's AI visibility tracking runs the same kind of check for you. It asks ChatGPT, Gemini (grounded in Google Search) and Perplexity your buyer questions, and you see the results for each engine separately, next to an overall rate. See pricing for what each plan includes (Starter is $19 a month, Pro $49 per site a month).

For any website, the XeoRank free analysis checks your AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) signals, the structural factors that make it easier for AI to parse and cite your content.

The Window Is Open Right Now

Here's why this matters more than most people realize: AI recommendation patterns are forming now. Brands that get mentioned early may find it easier to stay in the answers as AI search grows.

Think about it like early SEO. The businesses that figured out Google in 2005 built domain authority that's still paying dividends twenty years later. AI visibility is at that same inflection point.

The difference is that AI visibility depends more on content than early SEO did. You don't need to understand link building algorithms. You do need AI crawlers to be able to reach your pages, and then you need genuinely helpful content about your products and your category. Our guide to ranking on ChatGPT walks through both.

The businesses doing this work now will be ahead of the ones waiting for it to become "proven" or "mainstream."

Check where you stand today. Fix the obvious gaps this week. Build from there.