How to Measure AI Product Visibility: A Practical Testing Method

Traditional SEO has Google Search Console, keyword rank trackers, and analytics. GEO has prompt testing. There is no unified dashboard yet for AI product visibility — so you have to build your own testing workflow.

Here is a practical method for measuring whether your products appear in AI shopping answers.

The Core Principle: Test Buyer Prompts, Not Brand Searches

The mistake most merchants make is testing "does ChatGPT know about our brand?" That is the wrong question. The right question is: when a buyer asks an AI for help finding a product in your category, does your product appear?

This means testing buyer intent prompts, not brand searches.

Step 1: Build Your Prompt List

Write 10–15 buyer intent prompts for your category. Group them into three types:

Specific qualifier prompts (most important for GEO):

• "best vitamin C serum for oily sensitive skin under $50"
• "fragrance-free retinol alternative for beginners"
• "lightweight moisturizer for humid weather"

Comparison prompts:

• "serum vs oil for anti-aging — what do dermatologists recommend?"
• "niacinamide vs azelaic acid for post-acne marks"

Routine and use-case prompts:

• "what order do I use vitamin C, niacinamide, and retinol?"
• "how to build a skincare routine for dry sensitive skin on a budget"

Aim for prompts that sound like how a real shopper would ask a friend or Google — not how they would phrase a keyword search.

Step 2: Run the Prompts in Three AI Systems

Run each prompt in:

• ChatGPT (logged out, fresh session — avoid personalization)
• Perplexity (use the free version with web search enabled)
• Gemini (google.com/gemini — enables web search mode)

For each prompt, note:

• Which brands appeared by name
• Whether a product URL was linked
• What evidence the AI cited (reviews, price, ingredients, brand)
• Whether you appeared (brand name or URL) and in what position

Step 3: Build a Visibility Scorecard

After running all prompts, tabulate results:

| Prompt | Your brand cited? | Competitor cited? | Evidence cited | Action | |---|---|---|---|---| | "best C serum for oily skin" | No | Yes — Competitor X | Reviews, price, ingredients | Fix prompt fit + schema | | "fragrance-free moisturizer" | Yes (URL) | Yes — Competitor Y | Ingredients, claims | Add more evidence |

Step 4: Categorize the Gaps

Not retrieved — Your brand does not appear because the AI cannot find or does not trust your pages. Fix: crawlability audit, robots.txt check, schema validation.

Retrieved but not cited — The AI found your page but chose a competitor. Fix: comparison copy, evidence gaps, prompt fit issues.

Cited but wrong context — Your brand appears for the wrong use case or audience. Fix: product copy specificity, best-for blocks.

Cited correctly — Document what is working and apply the pattern to other pages.

Step 5: Re-test After Fixes

Run the same prompt list again after publishing GEO fixes. Compare:

• Did your brand appear more often?
• Did your citations move up in the answer?
• Did the evidence cited improve (now citing reviews instead of just price)?

This is a directional measure — don't expect overnight shifts. Track over 30–60 day intervals.

Automating This With GetPicked

GetPicked automates the prompt testing part of this workflow. For each category you specify, it runs structured buyer prompts against your store and reports how often your brand appears — with evidence cited and comparison to competitors. This replaces the manual spreadsheet with automated scoring you can track over time.

Run your first AI visibility test →