Factor 04 · 15% of the AI Commerce Score
AI Trust
Confidence™
A human can feel your store is trustworthy in a second. An AI agent cannot feel anything. It can only verify what it is able to parse. AI Trust Confidence measures whether your store proves its trustworthiness in machine-readable form.
What it measures
Trust the machine can verify
15%
AI Trust Confidence™
One of the heaviest factors in the AI Commerce Score. It gates recommendation: AI rarely recommends a store it cannot confirm is safe to buy from.
AI Trust Confidence™
AI Trust Confidence is the degree to which an AI shopping agent can independently verify that your store is legitimate, reliable, and safe to transact with, using structured, machine-readable signals rather than visual cues.
Recommending a store is a risk for an AI agent. If it sends a buyer to a store that turns out to be a scam, ships nothing, or refuses returns, the agent looks unreliable. So before it recommends you, it tries to confirm you are trustworthy. The problem is that the trust cues you built for humans are mostly invisible to it.
The core gap
Humans trust what they see. AI trusts what it can parse.
Your store can radiate trust to a human visitor and still register as unverifiable to an AI agent. These are two different trust languages, and most stores only speak the first one.
▶ What earns a human's trust
✓Polished design and professional photos
✓A wall of five-star reviews on the page
✓Recognisable payment and trust badges
✓A confident, familiar brand voice
🤖 What an AI agent can verify
✗Reviews rendered as images carry no data
✗Badges in a footer prove nothing parseable
✗Return policy hidden in a popup is unreadable
✗No structured business identity to confirm
Same store. One looks trustworthy, the other cannot be verified.
What carries the weight
The signals that build AI Trust Confidence
These are the machine-readable signals AI agents look for when verifying a store. The bars show their relative weight in our scoring. The exact rubric lives in the methodology, but the order rarely surprises anyone: proof of real customers and a clear refund path matter most.
Relative weight in AI Trust Confidence
AggregateRating schema
Highest
Return policy in HTML
Highest
Organization identity schema
High
Consistent structured pricing
High
Contact details in page source
Medium
Shipping terms stated clearly
Medium
Third-party authority (Trustpilot, press)
Medium
Weights are directional and illustrative. The binding values are defined in the AI Commerce Score methodology.
Where stores land
The verification spectrum
AI Trust Confidence is not on or off. A store sits somewhere on a spectrum from fully unverifiable to fully confirmed. Most stores cluster in the lower two bands without realising it.
Unverifiable
no machine-readable trust
Partially verified
some signals present
Confirmed
trust fully parseable
A store in the red band can have thousands of happy customers and still be skipped, because the agent has no way to confirm any of it.
What kills it
The trust signals AI silently misses
✗
Reviews that exist only as a widget
A review app that paints stars onto the page with JavaScript shows nothing to an agent reading the HTML. The rating exists for humans and is invisible to the machine. AggregateRating schema fixes this.
✗
Return policy behind a click
If the only way to read your refund terms is to open a modal or follow a JavaScript link, the agent cannot verify the policy exists. A plain crawlable return policy page is one of the fastest trust wins.
✗
No structured business identity
Without Organization schema stating who you are, the agent cannot confirm a real entity stands behind the store. A logo in the footer is a picture, not an identity it can check.
✗
Pricing that does not agree with itself
When the schema price and the displayed price disagree, or prices shift between pages without reason, the agent treats the store as unreliable and lowers its confidence in everything else it read.
How to raise it
Four changes that prove trust to AI
1
Turn on AggregateRating schema in your review app
Judge.me, Yotpo, Loox and most review apps can output AggregateRating with a real ratingValue and reviewCount. Enable it in the app settings so your customer proof becomes data, not just pixels.
2
Publish your return policy as crawlable text
Create a dedicated return policy page in plain HTML, state the window and conditions in clear sentences, and link to it from every product page. Avoid burying it in a modal or an iframe.
3
Add Organization schema with your real identity
Include legal name, logo URL, contact point, and social profiles in Organization JSON-LD. This gives the agent a verifiable entity to attach trust to and ties your store to your wider presence.
4
Make pricing agree everywhere
Ensure the price in your schema matches the price shown to shoppers, on every page. Consistency is itself a trust signal, and inconsistency quietly drags down confidence across the whole store.
The fastest win: AggregateRating schema plus a crawlable return policy together cover the two heaviest trust signals. For most stores they can be live in a single afternoon and move AI Trust Confidence more than any redesign would.
FAQ
AI Trust Confidence: common questions
What is AI Trust Confidence?▾
AI Trust Confidence measures whether an AI shopping agent can verify that a store is trustworthy using machine-readable signals such as AggregateRating schema, a crawlable return policy, business identity markup, and consistent pricing. A human can feel a store is trustworthy from its design, but AI can only trust what it can parse.
What signals build AI Trust Confidence?▾
The strongest signals are AggregateRating schema backed by a real review platform and a return policy in crawlable HTML. Organization identity schema and consistent structured pricing carry high weight as well. Contact details, shipping terms, and third-party authority such as Trustpilot or press add further confidence.
Why do trustworthy stores still score low?▾
Because their trust signals are built for human eyes, not machines. Review widgets render as images, return policies sit inside modals, and business identity lives in a footer logo rather than structured data. The store is trustworthy, but AI cannot verify it, so it recommends a competitor it can confirm instead.
How is this different from Recommendation Confidence?▾
AI Trust Confidence asks whether the agent can verify you are safe to buy from. Recommendation Confidence asks how strongly the agent will endorse you once it does. Trust is a gate that lets you into the recommendation set. Recommendation Confidence then decides how prominently you appear inside it.
See your AI Trust Confidence score
The free AI Commerce Score breaks out your AI Trust Confidence and shows exactly which trust signals AI cannot currently verify.
Free. No credit card. Results in 30 seconds.