Agent Rank Monitoring for E-Commerce

When a shopper asks AI what to buy, does it name your store?

ChatGPT, Claude, and Perplexity answer “what should I buy?” with a shortlist of two to five products. There is no page two. MentionCraft runs hundreds of simulated shopping sessions and shows you how often your store makes the list, and who takes your spot when it doesn’t.

Free scan · No credit card · Names the competitors winning your category

Simulated Session #214
Claude · Budget shopper

Best organic dog treats under $30 that ship to Canada?

Here are the strongest options I found:

1. Wilder Paws Co. Recommended
2. Barkfield Provisions Mentioned
3. Tundra Treat Co. Mentioned
Your store Not mentioned
Across 240 sessions this week You appear in 31% · Wilder Paws in 89%

We simulate shopping sessions across the agents your customers already use

ChatGPT / Claude / Perplexity / Gemini
Why This Is Different From SEO

There is no rank #11 in an agent’s answer

The Universal Commerce Protocol (UCP) is wiring agents directly into checkout, on Shopify and every other platform adopting it. When an AI shops on a customer’s behalf, it builds a shortlist, compares, and buys. You’re either in that shortlist or the sale happens without you ever knowing it existed.

Exclusion is total

A search result on page two still gets the occasional click. A product outside the agent’s shortlist doesn’t get bought at all. The whole consideration set is two to five products, and most merchants have no idea whether they’re in it.

One test proves nothing

Ask ChatGPT about your category twice and you’ll get two different answers. Responses shift with every run and every shopper, so a single manual check tells you almost nothing. It takes hundreds of runs before the numbers settle into something you can trust.

And it’s fixable

Agents recommend products they can read: complete data, policies a machine can parse, prices that mean what they say. Most of that is ordinary catalog hygiene, and hygiene is the kind of problem you can grade and work through.

How It Works

We do the shopping, hundreds of times

We shop your category the way real customers do, through the agents they actually use, and turn the results into numbers you can act on.

01

Simulate shopping sessions

We run your buy-intent queries on ChatGPT, Claude, and Perplexity. Each one repeats across shopper personas: budget-conscious, premium, repeat buyers, and regional variants like “ships to Canada.”

02

Measure your share of voice

We record every product and store mentioned in every session: who appeared, in what order, framed how. From that we compute your inclusion rate, first-mention rate, and share of voice, reported as ranges because the runs genuinely vary.

03

Fix what agents can’t read

Alongside the simulations, we grade your actual storefront: product schema, policy legibility, variant clarity, catalog completeness. Then we hand you a fix list, ordered by what’s most likely to move your numbers.

The Metrics

We don’t tell you your rank. We tell you your odds.

There’s no fixed leaderboard inside an AI agent. Every session is a fresh draw. So we measure the way brand trackers do: “You appear in 64% of simulated sessions for this query, up from 51% last month. Your top competitor appears in 89%.” Numbers you can watch move as you make changes.

  • Inclusion rate: do agents know you exist?
  • First-mention rate: are you the default answer?
  • Share of voice: what percent of the shelf do you own?
  • Persona spread: which shoppers see you, and which never will
Agent Visibility
Wildroot Botanicals
↑ 13 pts this month
64%
Inclusion
was 51%
22%
First mention
was 18%
17%
Share of voice
was 12%
Inclusion by Shopper Persona
Premium buyer 78%
Clean context 64%
Budget-conscious 31%

Insight: Premium shoppers see you. Budget shoppers don’t, because agents can’t verify your sale pricing.

The Free Scan

Your first scan names names

Enter your store URL. We run a batch of shopping sessions in your category and tell you exactly which competitors the agents recommend, and whether you’re one of them. You also get your Agent Readability Grade: how parseable your products, policies, and pricing are to a machine that’s deciding what to buy.

Agent Readability Grade
yourstore.com
C+
Product schema Complete
Shipping policy legibility Agents can’t parse it
Variant & price clarity Inconsistent
Return policy Parseable
UCP catalog readiness Not detected
Competitor Watch

Know the moment a competitor overtakes you

Agent recommendations shift as catalogs, reviews, and feeds change. We watch your tracked competitors on every scheduled run and email you when the lines cross, in either direction. Free accounts follow one competitor; paid plans follow as many as you like.

  • Competitors discovered automatically from the sessions themselves
  • Alerts only fire when a competitor actually passes you on a shared query
  • Trend lines built on rolling averages, so one weird run never panics you
Competitor Alert · This Morning

Barkfield Provisions just overtook you on “organic dog treats”

Barkfield Provisions 58%
Your store 52%

They added parseable shipping tiers to their policy page two weeks ago. Your readability grade flags the same gap. Fix it and watch the next four weeks.

Pricing

Start free. Pay when you want the trend line.

The scan is free and names your competitors. Paid plans keep the simulations running so you can see whether your fixes are working.

Free Scan $0

One-time simulation batch · named competitors · Agent Readability Grade · one competitor alert

Run My Free Scan
Monitor
$49 /month

Weekly simulations with trend lines, so you always know where you stand and whether it’s moving.

  • Weekly simulation runs
  • 10 buy-intent queries tracked
  • 3 competitors watched
  • Trend lines & overtake alerts
Start Monitoring
Most Popular
Growth
$149 /month

Daily runs and the full persona matrix, for merchants treating agent visibility like a channel.

  • Daily simulation runs
  • 30 queries · unlimited competitors
  • Full persona breakdown
  • Sentiment & framing analysis
Start Growth
Optimize
$299 /month

For merchants who want the fixing handled too. We draft the catalog changes your grade calls for and push each one after you approve it.

  • Everything in Growth
  • One-click feed & schema fixes
  • Pushed via your platform’s API, per-change approval
  • Priority simulation runs
Start Optimizing

Managing multiple stores? Agency plans include white-label reports and multi-store dashboards. Talk to us.

FAQ

Common questions

How is this different from SEO rank tracking?

Rank trackers measure your position on a results page. Agents don’t have a results page. They answer with a short list that changes from one session to the next. So we measure probability instead of position: how often you appear across hundreds of simulated sessions. It’s share-of-voice methodology from brand tracking, applied to agent commerce.

What is UCP, and why does it matter?

The Universal Commerce Protocol is the open standard that lets AI agents read a store’s catalog, check its policies, and complete a purchase without a human clicking through pages. As platforms adopt it, agents stop guessing from your marketing pages and start reading your catalog directly, which makes the quality of that catalog the thing that decides whether you get recommended. MentionCraft is built around UCP: the readability grade checks your store against it, and the simulations measure how agents actually treat you.

Why do you run the same query so many times?

Because agents answer differently every time, even with identical context. We repeat each query across every shopper persona until the numbers stabilize, then report ranges and rolling averages rather than a single figure that looks precise but isn’t. The spread between runs is real, and it’s often where the useful information lives.

Can you guarantee agents will recommend my store?

No, and anyone who promises that is guessing. What we can do is show you exactly where you stand, fix the things agents demonstrably can’t read on your storefront, and measure whether your inclusion rate moves. The improvements are real either way: better product data helps every channel, not just agents.

What does the readability grade actually check?

We crawl your storefront the way an agent does and grade what it can extract: Product schema completeness, offer and availability data, whether your shipping and return policies are parseable, variant and price consistency, and readiness for the Universal Commerce Protocol, the standard agents use to read and transact with stores. Every failing check comes with a specific fix.

Does this work with my platform?

Yes. Shopify, WooCommerce, BigCommerce, or fully custom: the scan and monitoring work from your public storefront, which is the same view an agent gets, so there’s nothing to install. You only connect your store if you upgrade to Optimize and want approved fixes pushed directly to your catalog.

Is agent shopping actually happening yet?

It’s early, and we won’t pretend otherwise. But the checkout rails are being built right now: UCP gives agents a standard way to read catalogs and complete purchases, the major platforms are adopting it, and shoppers already ask agents what to buy even when the agent can’t finish the sale. Merchants who sort out their agent visibility before the volume arrives will own the shelf when it does, the same way early SEO adopters did.

Find out who the agents recommend instead of you

One free scan: your inclusion rate, your readability grade, and the names of the competitors already on the shelf.

No credit card · Works with any storefront · Competitors named in your first report