To check if ChatGPT recommends your competitor over you, open ChatGPT and ask the same questions your customers ask before buying. Something like "What's the best [your category] for [your audience]?" Run it 3-5 times. Note who shows up, who doesn't, and where you land in the list. That's your baseline.
Most ecommerce founders have never done this. They track Google rankings religiously but have zero visibility into what AI systems say about their brand. That's a problem because 61% of consumers have now used AI tools like ChatGPT for online shopping research (Capital One Shopping, 2026). And here's the part that stings: if your competitor shows up in those answers and you don't, you just lost a sale you never knew existed.
This guide gives you the exact process to audit your AI visibility against competitors. Not theory. Actual prompts to copy, a scoring spreadsheet framework, and a fix playbook for when the results look bad.
Why You Need to Check ChatGPT Recommendations Right Now
AI-powered product discovery isn't coming. It's here. ChatGPT has over 800 million weekly active users, and it drives 87.4% of all AI referral traffic (Superlines, 2026). When someone asks "what's the best protein powder for runners" or "top Shopify apps for inventory management," ChatGPT gives a list. Your customers are reading that list.
The scary part? AI recommendations and Google rankings are almost completely disconnected. BrightEdge found that 88% of URLs cited by AI systems don't rank in Google's top 10. You could be #1 on Google and invisible to ChatGPT. Or ranked nowhere on Google but recommended by every AI platform.
I think most store owners are going to be shocked when they first run this audit. The brands AI recommends often aren't the ones with the biggest ad budgets or the best SEO. They're the ones with the strongest web presence in the places AI models actually learn from: Reddit, YouTube, review sites, and editorial coverage.
For a deeper look at how ChatGPT decides which products to recommend, we broke down the full mechanics in a separate guide.
Step 1: Build Your Prompt Library (15-25 Prompts)
Random testing gives you random results. You need a structured prompt library that mirrors real customer queries. Aim for 15-25 prompts across three categories.
| Prompt Category | What It Tests | Example Prompt |
|---|---|---|
| Discovery | Whether you appear in category-level queries | "What are the best [your category] brands in 2026?" |
| Comparison | How you stack up head-to-head against named competitors | "[Your brand] vs [Competitor]: which is better for [use case]?" |
| Problem-Solving | Whether you surface when users describe a pain point | "I need a [product type] that solves [specific problem]. What do you recommend?" |
| Audience-Specific | Whether AI associates you with your target customer | "Best [your category] for [specific audience or niche]?" |
| Price-Sensitive | Whether you appear in budget-conscious searches | "Affordable alternatives to [expensive competitor]?" |
Use the language your customers actually use. Not your internal product terminology. If you sell "premium athleisure" but customers search for "nice workout clothes," test with "nice workout clothes."
Quick note: don't include your own brand name in discovery prompts. The whole point is to see if ChatGPT brings you up organically. If you type your name into the prompt, you're testing recall, not recommendation.
Step 2: Run Each Prompt 3-5 Times Across Multiple AI Platforms
This is where most guides get it wrong. They tell you to run a prompt once and call it done. AI responses are probabilistic. Ask ChatGPT the same question 10 times and you'll get 10 different responses.
Run each prompt at least 3-5 times. A brand that appears in 1 out of 5 runs is in a very different position than one that appears in 5 out of 5. Frequency matters more than a single snapshot.
And don't stop at ChatGPT. Test across all four major platforms:
| AI Platform | Why It Matters | Key Difference |
|---|---|---|
| ChatGPT | 87.4% of AI referral traffic (Superlines) | Weights training data heavily; Shopify integration live |
| Perplexity | Growing fast for product research | Real-time web search; cites sources explicitly |
| Google Gemini | Integrated into Google Search (AI Overviews) | Google's own index; favors structured data |
| Claude | Popular with professionals and researchers | Tends to be more cautious and balanced in recommendations |
A brand that dominates ChatGPT might be completely absent on Perplexity. Honestly, I've seen cases where a mid-tier competitor shows up on every platform except ChatGPT, and vice versa. Every platform you skip is a blind spot.
Step 3: Score Every Response
Raw responses are useless unless you quantify them. For each prompt run, track four things:
- Mention status: Were you mentioned at all? Yes or no. This is the binary that matters most.
- Position: Were you #1 in the list? #3? Last? Or not mentioned at all?
- Sentiment: Was the mention positive, neutral, or subtly negative? ("Brand X is good but expensive" counts as mixed.)
- Competitor count: How many competitors appeared alongside you? Who was first?
Here's the scoring framework I'd recommend:
| Metric | How to Score | What It Tells You |
|---|---|---|
| Mention Rate | % of runs where your brand appeared (e.g., 3 out of 5 = 60%) | Overall AI visibility for that query |
| Average Position | Average list rank across all runs (1st = 1, not mentioned = N/A) | How strongly AI associates you with the query |
| Competitor Gap | Your mention rate minus top competitor's mention rate | How far behind (or ahead) you are |
| Sentiment Score | +1 positive, 0 neutral, -1 negative per mention | Whether AI is helping or hurting your brand perception |
A simple spreadsheet works fine here. Rows are prompts, columns are platforms, cells contain your mention rate and position. After 20 prompts across 4 platforms, you'll have a clear picture of exactly where you stand.
Want to skip the manual work? True Margin's AI Authority Checker automates this analysis and gives you a scored breakdown of your AI visibility across platforms.
What Your Results Actually Mean
Once you've run the audit, your results will fall into one of four scenarios. Each one requires a different response.
Scenario 1: You show up consistently, competitors don't. You're winning. Maintain your presence by continuing what's working (content, community engagement, review generation). Don't get complacent. Competitors who invest in GEO can close this gap in 3-6 months.
Scenario 2: Both you and competitors show up. This is the competitive zone. Look at position and sentiment. If competitors consistently appear first, they have stronger AI signals. Focus on the specific sources where they outmatch you (more on that below).
Scenario 3: Competitors show up, you don't. This is the most common result for brands that haven't invested in Generative Engine Optimization. You're invisible to AI-driven shoppers. The fix playbook below is designed for exactly this situation.
Scenario 4: Neither you nor competitors show up. Your category might be too niche for AI to have strong opinions yet. This is actually an opportunity. The first brand that builds AI visibility in an underserved category often dominates it for a long time.
Why ChatGPT Recommends Your Competitor (the Real Reasons)
If the audit reveals that ChatGPT recommends your competitor over you, here's what's actually happening under the hood. It's not magic. It's data.
They have more mentions on AI training sources. ChatGPT pattern-matches against everything it learned during training. If your competitor gets talked about more frequently on Reddit, YouTube, review sites, and industry publications, the model associates them more strongly with your category. YouTube alone accounts for 39.2% of AI citation sources (BrightEdge), and Reddit has signed over $130M in AI training data deals.
Their content is clearer. If your website uses vague marketing language ("we empower brands to thrive") while your competitor uses specific, structured, category-relevant copy ("inventory management software for Shopify stores under $5M/year"), AI will prefer them. Every time. AI needs to understand exactly what you do to recommend you.
Their entity signals are more consistent. Name, positioning, and key facts should be identical everywhere: website, LinkedIn, G2, Crunchbase, press releases. Inconsistency reduces AI confidence. If your homepage says one thing and your LinkedIn says something slightly different, AI treats that as a trust penalty.
Most brands that lose the AI visibility game lose it on the first point. They simply aren't present where AI models learn. For more on the mechanics, read our breakdown of how ChatGPT handles Shopify product recommendations.
The Fix Playbook: How to Overtake Competitors in AI Recommendations
Here's the honest truth: you can't fix this overnight. Paid ads have negligible impact on AI recommendations (BrightEdge). You earn AI visibility. But the brands that start now have a massive advantage because most of your competitors haven't started yet.
Priority 1: Build your presence on AI training sources. Reddit and YouTube are disproportionately weighted. Create genuine, helpful content on these platforms. Not promotional spam. Real answers to real questions in your niche. One viral Reddit thread where someone organically recommends your product is worth more for AI visibility than 50 blog posts.
Priority 2: Fix your own site content. Every product page and landing page should clearly state what you do, who it's for, and why someone should choose you over alternatives. Structured data (FAQ schema, Product schema, Article schema) makes it easier for AI to extract and reference your information.
Priority 3: Earn third-party mentions. PR coverage, podcast appearances, expert quotes in articles, reviews on independent sites. Every authoritative mention is a signal AI models learn from. Focus on publications and platforms your target customer actually reads.
Priority 4: Create comparison content. If customers are asking ChatGPT to compare you and competitors, create the definitive comparison on your own site. Structured, fair, data-driven. AI models are more likely to cite content that directly addresses the comparison query.
Priority 5: Track monthly. Set up a recurring monthly audit using the prompt library you built in Step 1. AI models update frequently. Competitor visibility can shift fast. The brands that track consistently catch changes before they become permanent disadvantages.
Check your AI visibility score right now
True Margin's AI Authority Checker scans your brand across the signals that AI systems use to generate recommendations. Free, instant, no signup required.
Open AI Authority Checker →How to Set Up Ongoing Monitoring
A one-time audit is a start. But AI visibility is a moving target. Here's the cadence that works:
- Monthly full audit: Run your complete 15-25 prompt library across all four platforms. Update your scoring spreadsheet. Compare to last month.
- Bi-weekly spot check: Pick your 5 highest-intent prompts (the ones closest to a purchase decision) and run them quick. If something shifted, trigger a full audit early.
- Quarterly competitor deep-dive: Add 2-3 new competitors you've noticed gaining traction. Expand your prompt library with new customer language you've picked up from support tickets or reviews.
The stores that track this consistently tend to see improvement within 3-6 months of starting their GEO efforts. The ones that check once, panic, and forget about it don't. For a broader view of what goes into an AI visibility score and why it matters, we cover the full framework in a separate guide.
Common Mistakes When Checking ChatGPT Competitor Recommendations
After helping brands run these audits, I've seen the same mistakes over and over:
Testing once and drawing conclusions. AI responses are non-deterministic. A single run tells you almost nothing. You need 3-5 runs per prompt minimum to see the pattern. Some brands appear in 60% of runs and think they're invisible because they caught a bad sample.
Only testing on ChatGPT. Each AI platform draws from different data sources and weights signals differently. Perplexity does real-time web search. Gemini leans on Google's index. Claude weights editorial content differently. Testing on one platform gives you one-quarter of the picture at best.
Including your brand name in discovery prompts. If your prompt says "Is [my brand] better than [competitor]?" you're testing recall (can AI recognize you?), not discovery (does AI recommend you organically?). Both matter, but they're different tests. Keep them separate.
Ignoring sentiment. Being mentioned isn't always good. If ChatGPT says "Brand X is popular but has reliability issues," that mention is worse than not being mentioned at all. Track the tone, not just the presence.
Not documenting results. If you don't record your baseline, you can't measure improvement. A spreadsheet takes 10 minutes to set up. Use it.
Frequently Asked Questions
How do I check if ChatGPT recommends my competitor?
Open ChatGPT and ask the category-level questions your customers would ask, like "What are the best [your category] brands?" Run at least 15-25 prompts across discovery, comparison, and problem-solving categories. Run each prompt 3-5 times because AI responses are probabilistic. Track whether your brand appears, your position in the list, and how you're described relative to competitors.
Why does ChatGPT recommend my competitor instead of me?
ChatGPT recommends brands it has the strongest confidence in based on training data, web presence, and entity consistency. If your competitor has more mentions on Reddit, YouTube, review sites, and industry publications, the model associates them more strongly with your category. Vague marketing copy on your site also hurts because AI prefers clear, structured, category-specific language.
How often should I check ChatGPT for competitor recommendations?
Run a full prompt audit at least once per month. AI models update their knowledge regularly, and competitor visibility can shift quickly. Between full audits, do a quick spot-check every two weeks with your top 5 highest-intent prompts to catch sudden changes.
Should I test on ChatGPT only or other AI platforms too?
Test across ChatGPT, Claude, Perplexity, and Gemini at minimum. Each AI model draws from different data sources and weights signals differently. A brand that dominates ChatGPT recommendations might be completely absent on Perplexity. Every platform you skip is a blind spot where competitors could be winning.
Can I pay to get recommended by ChatGPT over competitors?
No. AI recommendations are almost entirely organic. Paid ads have negligible influence on which brands AI systems recommend. The signals that matter are brand mentions across the web, third-party reviews, editorial coverage, Reddit and YouTube presence, and content depth on your own site. You earn AI visibility, you don't buy it.
What is a good AI visibility score for an ecommerce brand?
Appearing in over 50% of relevant prompts across multiple AI platforms puts you in strong territory. If you appear in fewer than 20% of category prompts, competitors are likely capturing the majority of AI-driven discovery traffic. True Margin's free AI Authority Checker gives you a baseline score to track over time.