How to Measure AI Visibility for Ecommerce: Tools, Metrics and the AI Readiness Score
AI visibility tells you whether assistants like ChatGPT and Gemini mention and recommend your products. This guide covers the free reports and paid tools that measure it, plus the product data metrics that explain why the numbers move.
By Alberto Barberis, Founder and CEO, AndromedAI · Updated October 2026
AI visibility, defined
AI visibility is how often and how favorably AI assistants such as ChatGPT, Gemini, Google AI Mode, Perplexity or Copilot mention, cite or recommend a brand and its products. Ecommerce teams measure it through GA4 AI referral traffic and Google's own AI reports, plus prompt tracking for share of voice. They improve it by fixing the product data those assistants read.
AI visibility in 30 seconds
AI visibility is how often, and how well, AI assistants mention and recommend your brand and products. You measure it with four free platform reports plus prompt tracking. You improve it by fixing product data.
- 01
Ask an assistant for a product and it usually names three to five. AI visibility is the share of those answers that include your brand and your products.
- 02
Four free reports each show a slice of it: GA4 AI referral traffic, the Search Console generative AI report, Merchant Center AI performance insights and Shopify's Agentic Storefronts page.
- 03
Prompt tracking tools add competitor share of voice across the big assistants. Answers change almost every run, so read them as samples and never as rankings.
- 04
Visibility tools measure the symptom. Product data readiness measures the cause, and the cause is the part you can fix.
- 05
A setup that holds up puts visibility and traffic next to per-SKU readiness, reviewed once a month.
AI search visibility: symptoms vs causes
Every AI visibility metric is an outcome. For an online store, the inputs that decide it are product attributes, keyword coverage, answers to shopper intents and clean metadata.
growth in AI traffic to US retail sites in Q1 2026, year over year
The channel is compounding fastSource: Adobe via Decryptbetter conversion for AI traffic than non-AI traffic in March 2026, after converting 38% worse a year earlier
AI visitors now arrive ready to buySource: Adobe via Decryptchance that ChatGPT, Claude or Google's AI return the same brand list twice for the same prompt
Single prompt checks tell you very littleSource: SparkToro and Gumshoe via Search Engine LandAI referrals are still a low single-digit share of traffic for most stores, yet they averaged 770.7 million visits a month worldwide between June 2025 and May 2026, up 117.4% year over year (Similarweb).
What AI visibility tools actually measure
Most AI visibility tools send a fixed list of prompts to AI engines every day and record whether your brand shows up, where, and how it's described. That gives you AI brand visibility (mention rate) and share of voice in AI, with citations and sentiment on top. It tells you if you're in the answer.
It doesn't tell you why.
Say a competitor's trail shoe appears in 64% of answers about "waterproof trail running shoes for wide feet" and yours in 8%. Brand awareness is rarely the reason. When we open the losing product page, it often doesn't mention width or the membrane at all, and the feed has an empty material field under a title like "Trail Runner II Black". The winning answer looks more like this:
ChatGPTLook at what the reply leans on. Width, membrane, lug depth, price. Each one is an attribute somebody wrote down.
Symptom metrics
Mentions, citations, share of voice, AI referral sessions, revenue. They record decisions AI engines already made and wobble from run to run.
Cause metrics
How complete the attributes are, which keywords and intents the page covers, how specific the titles are, whether PDP and feed agree. Deterministic, measurable per SKU, fixable this week.
For an online store, the fix for being left out of AI answers almost always lives in the product catalog, so measure the catalog too.
Google seems to agree. Its Merchant Center AI report includes an attribute completeness score that flags products missing the specifications shoppers search for, such as color, style or material (Google Merchant Center Help). Shopify's agentic page suggests improvements to your product data (Shopify changelog). That's the logic behind Agentic Commerce Optimization: search engines and shopping ads read the same product data that AI agents do.
How to measure AI visibility for your brand: the four reports to set up
Start with four free reports: GA4 for AI referral traffic, Search Console for AI impressions, Merchant Center for AI shopping share of voice and Shopify's Agentic Storefronts page for AI channel sales.
| Report | What it measures | Surfaces | Availability (Oct 2026) | Main limit |
|---|---|---|---|---|
| Google Analytics 4, AI channel | Sessions and revenue from AI assistants, plus key events | ChatGPT, Perplexity, Gemini, Copilot, Claude, others | All GA4 properties, requires setup | Misses clicks that arrive without a referrer |
| Search Console, generative AI performance | Impressions of your URLs inside AI answers | AI Overviews, AI Mode, AI features in Discover | Launched June 2026 for a subset of UK sites, wider rollout to follow | Impressions only, no clicks, no queries |
| Merchant Center, AI performance insights | Share of voice, funnel stage, product terms, attribute completeness | AI Mode, AI Overviews, Gemini app | US, Canada, Australia, India, New Zealand since September 2026 | English queries only, organic only, benchmarked against your chosen competitors |
| Shopify, Agentic Storefronts | Results by AI channel, queries your products appear for | ChatGPT, Copilot, Shop plus other AI channels via Shopify Catalog | Shopify merchants, admin.shopify.com/agentic | Only channels connected through Shopify Catalog |
Many teams buy a visibility tool first. We'd do it the other way round. The free reports are first party and tied to revenue, and they show whether AI traffic in your market justifies a paid tool.
Merchant Center AI performance insights
Google announced the report on May 27, 2026. Rollout started in September in Australia, Canada, India, New Zealand and the US (The Keyword). It has four modules (Google Merchant Center Help). Share of voice compares you with similar brands across Google's AI surfaces (AI Mode, AI Overviews, the Gemini app). Shopping funnel performance splits results into discovery, evaluation and purchase. Product term insights covers the terms in play. Product attribute insights carries the attribute completeness score, listing products that lack the specifications shoppers ask about.
We look at the attribute module first. It's a list you can hand to a catalog team on Monday.
It covers organic visibility only, English queries only, and it benchmarks against the competitor set you define (The Keyword), so a soft competitor set flatters you. Brands in the UK, Ireland, the Netherlands and Italy don't have it yet. If you also sell in the US, use its attribute module as a proxy, since one product feed usually serves every country.
Search Console generative AI performance reports
Google added dedicated generative AI performance reports to Search Console in June 2026 (Google Search Central). They count impressions of your URLs in AI Overviews and AI Mode, plus Discover's AI features, split by page, country, date or device. No clicks, no CTR, no queries (Evolv). On Shopify, filter pages on /products/ for PDPs and /collections/ for category pages. Then compare before and after.
Shopify Agentic Storefronts
Since May 2026 Agentic Storefronts has its own admin page. Merchants see results across AI platforms such as ChatGPT or Copilot, the queries their products appear for, and suggestions to improve product data (Shopify changelog).
Only GA4 works on every platform and ties AI traffic to revenue. Setup takes about 10 minutes.
How to track ChatGPT and AI referral traffic in GA4
Create a custom channel group in GA4 with an "AI assistants" channel that matches AI referrer sources by regex, and place it above Referral. Custom channel groups apply retroactively, so your history shows up right away.
By default GA4 files AI visits under Referral, next to affiliates and old blog links. AI referral traffic needs its own channel so you can compare it with Organic Search.
- 1
Open channel groups
In GA4 go to Admin, Data display, Channel groups. Standard properties can have 2 custom channel groups plus the default one; 360 properties can have 5.
- 2
Copy the default group
Click Copy to create new on the Default Channel Group and name it "Channels with AI".
- 3
Add an AI assistants channel
Click Add new channel, name it "AI assistants" and set the condition Source, matches regex, with the pattern below.
- 4
Move it above Referral
Click Reorder and drag "AI assistants" above Referral. GA4 assigns traffic to the first matching channel, so a Referral rule on top swallows every AI visit.
- 5
Save and set as primary
Save the group. Custom channel groups apply to your reports retroactively, so past AI visits get reclassified too.
- 6
Build the report
In Acquisition, Traffic acquisition, switch the dimension to your new channel group. Add revenue and key events next to engagement rate, then save a comparison of AI assistants vs Organic Search.
The regex
Use a pattern that matches the source names AI assistants pass, for example chatgpt|openai|perplexity|gemini|bard|copilot|claude|grok|deepseek|mistral. The main referrer hostnames are chatgpt.com (plus the legacy chat.openai.com), perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai. ChatGPT also appends utm_source=chatgpt.com to citation links (Terminus).
Every quarter, sort your Referral report by sessions and look for assistant names you haven't seen before.
The menu path, the limits, the ordering rule and the retroactive behavior are all documented in Google Analytics Help.
What GA4 will miss
Treat the AI channel as a floor. The real number is higher.
- Clicks without a referrer. Links opened from mobile apps or in-app browsers can arrive with no referrer data and land in Direct or (not set) (SEO Works).
- AI Overviews and AI Mode. Clicks from Google's AI features arrive as google / organic, so they stay inside Organic Search. Search Console is the place for those.
- Influence without a click. Someone reads a ChatGPT answer, then searches your brand. That visit looks like branded search or direct.
- Agent purchases. Purchases completed inside an AI surface show up in your commerce platform's channel reports, such as Shopify's, and never become GA4 sessions.
Best AI visibility tools for ecommerce (2026)
Profound, Peec AI, Semrush's AI Visibility Toolkit and Ahrefs Brand Radar all track how often AI engines mention and cite your brand. They differ on engine coverage and SKU tracking, and on price. None of them rewrites your product data, which is where a readiness tool comes in.
| Profound | Peec AI | Semrush AI Visibility Toolkit | Ahrefs Brand Radar | AndromedAI | |
|---|---|---|---|---|---|
| Category | Dedicated AI search tracking platform | Dedicated AI search tracking platform | Add-on to an SEO suite | Index inside an SEO suite | Agentic Commerce Optimization platform |
| Engines (as listed by vendor) | ChatGPT, Perplexity, Claude, Gemini, Copilot, DeepSeek, AI Overviews | ChatGPT, AI Mode, AI Overviews, Copilot, Gemini, more on Enterprise | ChatGPT and AI Mode named | AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini, Copilot | Optimizes the data all of them read |
| Ecommerce features | ChatGPT Shopping visibility at SKU level, agent analytics | Catalog upload from Shopify or CSV, SKU-level visibility, win rate | Prompt research, AI search site audit | Estimated impressions weighted by search volume | AI Readiness Score per SKU, rewrites, attributes, PLPs, publishing |
| Pricing (public) | Demo, no public price | Plans by prompt volume, from 50 prompts | $99 per month, 25 tracked prompts | From EUR 47 per month for custom prompts | Free AI Readiness Audit |
Sources: Profound, Peec AI, Semrush, Ahrefs.
How to choose
- Leadership wants competitive context. A dedicated platform such as Profound or Peec AI gives you the brand view.
- You already pay for Semrush or Ahrefs. Start with their AI modules. It's the cheapest test.
- You run a large catalog. Look for SKU-level tracking and pair it with something that fixes products. A dashboard showing 4,000 invisible SKUs doesn't write 4,000 product pages.
One honest caveat about our own column: AndromedAI isn't a visibility tracker. If what you need is a weekly mention chart, buy one of the other four.
Where AI readiness fits
Visibility tools
Mentions, citations, share of voice. Best for benchmarking and for spotting the intents where competitors win.
Platform reports
GA4, Search Console, Merchant Center, Shopify. Free and first party, tied to revenue, limited by market.
AI readiness
Scores each product page on the data AI engines need, then fixes it. Turns a vague gap into a list of SKUs.
For the content and authority side, see GEO for ecommerce, and for what changes compared with classic search work, GEO vs SEO for online stores. For ChatGPT specifically, read how to get your products recommended by ChatGPT.
The AI Readiness Score and an AI visibility dashboard template
The AI Readiness Score measures the cause: how complete and intent-rich each product page is, across four dimensions. Put it next to your visibility and traffic metrics in one monthly dashboard.
The four dimensions
| Dimension | What it reads | What it checks | Example gap |
|---|---|---|---|
| Product Data Completeness | Attributes, metafields, category, tags | Are the attributes an agent needs to match a request present? | material and size missing on a linen shirt, fit never stated |
| Keyword Coverage | Title, description, bullets, FAQ | Does the page use the words shoppers search and ask with? | Page says "natural top", shoppers ask for "men's linen shirt" |
| Customer Intent Match | Description, bullets, FAQ | Does the page answer who it's for, when to use it, how it compares? | Nothing says "breathable for hot, humid weather" |
| Shopping Metadata | Page title, H1, meta description | Are titles and metadata specific and intent-rich? | Title is a SKU code plus a color |
Each dimension maps to something an AI engine does: it filters on attributes, matches words, justifies a pick with use cases and reads titles first. In Merchant Center that data lives in fields such as title, description, gtin, color, material, size and product_detail (Google Merchant Center Help).
Completeness is where we find the most damage. Often the data exists, just in the wrong place. Color sits only in a variant option called "Sage", or material lives in a custom Shopify metafield your feed app never maps, so the PDP looks fine while Merchant Center gets an empty field. Check the feed too.
title
BeforeTrail Runner II Black
AfterKestrel Trail WP Wide Men's Waterproof Trail Running Shoe, 2E Width, Black
description
BeforeOur best trail shoe yet. Built for adventure.
AfterWaterproof trail running shoe in a true 2E wide fit for runners with wide feet. A breathable membrane keeps water out on wet trails, and 5 mm lugs grip mud and loose gravel.
attributes
Beforecolor: black
Aftercolor: black; material: recycled mesh with waterproof membrane; size_system: US; width: 2E wide; terrain: mud, wet trails
Crawler access sits outside the score, but nothing above works without it. Keep OAI-SearchBot allowed in robots.txt so ChatGPT search can read your pages. GPTBot only governs model training (OpenAI), and blocking it doesn't hide you from ChatGPT search. Check your CDN as well, since a bot rule there can block crawlers your robots.txt allows. Sort this out before you spend time on an llms.txt file for your store.
Score
Which products are not ready, and why?
AI Readiness Score per SKU across the four dimensions
Prioritize
Which gaps cost the most?
High-demand products with low scores, and intents where competitors win
Fix
What changes on the page and in the feed?
Attributes, titles, descriptions, bullets, FAQ, use cases, conversational attributes, category pages
Publish
Is the fix live everywhere AI reads?
Store, PIM, Merchant Center, other feeds, consistently
Measure
Did visibility and revenue follow?
GA4, Search Console, Merchant Center, prompt tracking
Re-score after every catalog update and every quarter.
Dashboard template
We keep it to one page in Looker Studio or a spreadsheet, refreshed monthly.
| Metric | Type | Source | Cadence | Read it as |
|---|---|---|---|---|
| AI assistants sessions and revenue, with key events | Traffic | GA4 custom channel group | Weekly, reviewed monthly | Revenue impact, trending up |
| AI impressions on PDPs and PLPs | Visibility | Search Console generative AI report | Monthly | Which pages Google's AI surfaces |
| AI shopping share of voice by funnel stage | Visibility | Merchant Center AI performance insights | Monthly | Position against competitors (eligible markets) |
| AI channel sales and top queries | Revenue | Shopify Agentic Storefronts | Monthly | Which channels and queries sell |
| Mention rate and share of voice by intent | Visibility | Prompt tracking tool | 4-week rolling average | Direction only, never a single day |
| AI Readiness Score by category | Cause | AI Readiness audit | Monthly and after each update | What to fix next |
Read it bottom up. Readiness moves first. Impressions and share of voice follow, and revenue comes last. No movement after two to three months often points to a crawling or feed approval problem, so check those before rewriting again, then revisit your intents.
AI visibility with AndromedAI
AndromedAI measures AI readiness for every product, fixes the gaps and publishes the improved pages to your store and feeds. It doesn't replace your visibility tool or your analytics. It's the step that turns what they find into better product data.
How the platform maps to the work
| Step | AndromedAI | What it does |
|---|---|---|
| Measure the cause | AI Readiness Audit | Scores product pages on Product Data Completeness, Keyword Coverage, Customer Intent Match and Shopping Metadata, and shows which products to fix first |
| Fix product pages | Optimizer | Rewrites titles, descriptions, bullets, FAQ; extracts attributes; adds use cases and intents |
| Create missing pages | Creator | Creates complete product pages from brand or supplier data, in 12 languages |
| Win category prompts | Category page optimizer | Builds category pages around real demand, for questions such as "best linen shirts for summer" |
| Publish everywhere | Integrations | Publishes to Shopify, Google Merchant Center, Salesforce Commerce Cloud, Adobe Commerce, Shopware, Akeneo, Plytix and WooCommerce |
The Brand Kit keeps every page in your voice. Approval workflows decide what goes live, with auto-approval above an AI Checker score of 4.0, so people review only the borderline pages. AndromedAI also generates Merchant Center conversational attributes. More than 500 catalogs have been optimized on the platform.
Your visibility tool finds the intents where you lose. The AI Readiness Score finds the SKUs behind the gap. AndromedAI fixes them and publishes, and GA4 confirms whether it worked.
clicks from AI chats
Bomboogieto first sales from ChatGPT, with -95% catalog generation costs
Matassasales and +160% organic traffic, 483 hours saved
SemprefarmaciaFAQ
Set up an AI channel in GA4 for AI referral traffic and revenue. Check the Search Console generative AI report and, where available, Merchant Center AI performance insights. Add prompt tracking if you need share of voice in other assistants, and score your product pages for AI readiness to find the causes.
In GA4, go to Admin, Data display, Channel groups and copy the default group. Add a channel called AI assistants with the condition Source matches regex, using a pattern such as chatgpt
Profound and Peec AI are dedicated platforms, and both list SKU-level product tracking. Semrush's AI Visibility Toolkit and Ahrefs Brand Radar add AI tracking to SEO suites they already sell. Pair any of them with a tool that scores and fixes product pages.
It's the percentage of brand mentions in AI answers, for a set of prompts, that belong to your brand rather than competitors. Merchant Center reports it for Google's AI surfaces. Prompt tracking tools calculate it for other engines.
AI answers are probabilistic. A 2026 SparkToro and Gumshoe study found under a 1 in 100 chance of getting the same brand list twice for the same prompt. Read visibility as a percentage across many prompts over several weeks.
No. Visibility measures outcomes, such as being mentioned. Readiness measures inputs: whether each product page carries the attributes and keywords AI engines need, answers shopper intents and has clean metadata. Readiness is measurable per SKU and directly fixable.
Partly. Since June 2026, the Search Console generative AI performance reports show impressions in AI Overviews and AI Mode, plus Discover AI features, by page, country, date or device. They don't show clicks or queries. Access started with a subset of UK sites, with a wider rollout to follow.
Glossary
- AI visibility
- How often and how favorably AI assistants mention, cite or recommend a brand or product
- AI referral traffic
- Visits to a website from links inside AI assistants such as ChatGPT, Perplexity or Copilot
- Share of voice in AI
- A brand's share of all brand mentions in AI answers for a defined set of prompts
- Prompt tracking
- Running a fixed set of prompts on AI engines on a schedule to record mentions and positions, plus cited sources
- Custom channel group
- A GA4 setting that groups traffic sources into channels you define, such as AI assistants
- AI performance insights
- Merchant Center reports on how products perform in Google's AI shopping surfaces, including share of voice and attribute completeness
- AI Readiness Score
- AndromedAI's score of how ready a product page is to be recommended by AI engines, across four dimensions
- OAI-SearchBot
- OpenAI's crawler for ChatGPT search, which OpenAI recommends allowing in robots.txt
Keep reading
A closer comparison of the trackers in this guide
Guidellms.txt for Ecommerce: Does Your Store Need One?What the file does and what evidence exists for stores
GuideGEO vs SEO: What's Different for EcommerceHow optimizing for AI answers differs from classic search
GuideGEO for Ecommerce: The Complete Guide to Generative Engine Optimization for Online StoresThe content and authority side of AI visibility
Sources (18)
- Google Search Central: Generative AI performance reports
- Evolv: Search Console AI reporting
- Merchant Center Help: AI shopping insights
- The Keyword: Merchant Center AI insights
- Shopify changelog: Agentic Storefronts
- Decrypt: Adobe AI retail traffic, Q1 2026
- CMOtech UK: Adobe holiday AI traffic
- Similarweb: AI referral traffic
- Search Engine Land: SparkToro AI consistency study
- Google Analytics Help: Custom channel groups
- Terminus: Tracking AI referral traffic
- SEO Works: ChatGPT traffic in GA4
- OpenAI: Crawlers
- Peec AI: Pricing
- Profound
- Semrush: AI Visibility Toolkit
- Ahrefs: Brand Radar
- Merchant Center Help: Product data specification
Featured in
