READINESS · CANONICAL DEFINITION

What is AI Commerce Readiness?

AI Commerce Readiness is whether AI can find your brand and sell it from your own product data, whether the experiences you own are right for the shopper in front of them, and whether the people who buy your category buy you and keep buying you.

Definition

AI Commerce Readiness (n.)

the discipline of making a brand discoverable by AI, its products usable by AI shopping systems, its owned experiences relevant to permissioned shoppers, and the brand preferred by the people who buy its category.

Shoppers increasingly ask AI assistants what to buy. Assistants name a few brands, are beginning to complete purchases, and leave a brand very little of what it is used to seeing. AI Commerce Readiness is everything that has to be true for a brand to win in that environment, of which AI visibility is only the first part.

TERM MAINTAINED BY HERM · FIRST PUBLISHED JUNE 22, 2026 · CITE: herm.io/for-brands/readiness

Become AI Commerce Ready Read the methodology

Three things changed. No existing discipline covers all three.

Discovery moved. Buying decisions now happen inside conversations a brand's analytics never see. A shopper asks an assistant what to buy for sensitive skin; three brands come back. Absent from the answer means never in the running, and no dashboard records the loss.

Product information has to be machine-usable. A page built for a human to browse is not the same artefact as a catalogue a shopping system can work with. Specs, price, stock and a destination that actually ships to the shopper now have to be readable by software, not just legible to a person. A brand that is visible but not usable gets mentioned and skipped.

Relevance and purchase happen in different places. The answer, the signed-in experience and the purchase are increasingly three different surfaces, and no single measurement spans them. That is why readiness needs separate evidence for each dimension rather than one number stretched across all four. It is also why a verified purchase, on its own, still does not say which answer produced it.

Search-era disciplines cover the first shift at best. AI Commerce Readiness names the whole problem: not "are we mentioned," but "can the machine that's replacing the shop assistant find us and sell us, are our own surfaces right for the shopper it sends, and do the buyers in our category actually buy us."

Readiness is a ladder. Each rung has a test.

Four rungs, one destination. Every brand has a position across all four. Each rung below is defined by the question it answers, the evidence it is read from, and the thing it must not be read past.

Rung 01 · Be Visible

When buyers ask AI about this category, does your brand enter consideration, and how is it represented?

A brand is visible when AI models know it and recommend it in answer to the questions its buyers actually ask, and when it can see, in the observed answers themselves, which brands were named, what the models said about them and which sources they cited.

Evidence: Observed model responses, cited sources and recurring answer patterns.
Does not mean: Observed outputs are not hidden model reasoning, and they do not prove a later purchase.

Brand Visibility Live

Rung 02 · Be Sellable

Can an AI shopping system use your product information to satisfy a real buying need and provide a valid path to purchase?

A brand is sellable when an AI shopping system can use its product information to satisfy a real buying need: quote specs, price and stock from a live, agent-readable catalogue, and route the shopper to a destination verified to ship to them.
Herm reads this rung from the brand's connected product information, put through controlled Shopping Tests. Connecting the feed is required for the full Product Readiness assessment.

Evidence: Connected catalogue data and controlled Shopping Tests.
Does not mean: Controlled tests are not the live consumer experience of third-party AI assistants.

Product Readiness Live

Rung 03 · Be Relevant

Can your owned signed-in experience make a better choice for this connected shopper without taking custody of their profile?

A brand is relevant when the signed-in experiences it owns make a better choice for a shopper who chose to connect. In an eligible session the brand sends Herm the options it could already show, along with that shopper's connection context; Herm returns a ranking or a selection, and the profile behind the decision stays inside Herm.
This rung scores the surfaces the brand controls. It does not measure, and Herm does not influence, what a frontier AI assistant recommends to an individual shopper.

Evidence: Eligible options, permissioned connection context and Herm's internal decision process.
Does not mean: The brand receives the decision, not the underlying shopper profile.

Customer Intelligence Early access

Rung 04 · Be Preferred

Among people who buy your category, do they choose your brand, and do they come back?

A brand is preferred when the shoppers who buy its category buy it, and keep buying it. Two things move that number: reaching category buyers who do not yet buy the brand, and rewarding loyalty a shopper has already shown across retailers, not just in the brand's own store.
A receipt confirms that a purchase happened. It does not establish which AI answer caused it; the cross-surface identifier that join would need is not generally available today.

Evidence: Consented purchase outcomes and, where Offers is used, covered offer and revenue outcomes.
Does not mean: Receipt verification confirms an outcome; it does not establish answer-level causality.

Offers Live

Readiness is not a mention count.

A generation of tools now tracks whether AI models mention your brand. Useful, and incomplete by three dimensions. A mention count tells you the machine said your name; it does not tell you whether the machine could quote your price, confirm your stock, or send a shopper somewhere that ships to her, whether the experience waiting for her there was right for her, or whether the shoppers who buy your category are buying you at all.

Visibility gets you mentioned. Readiness gets you mentioned, sellable, relevant, and preferred.

A progression, not a purchasing sequence.

The four rungs are conceptually cumulative, diagnostically independent and commercially modular. Those are three different statements, and collapsing them is how a framework turns into a queue.

Conceptual progression

Read together, the four rungs describe an increasingly complete AI commerce capability. A brand that is visible, sellable, relevant and preferred is ready in a way a brand that is only visible is not.

Diagnostic independence

Read separately, they are four different measurements of four different things. A brand can be strong on one dimension and weak on another, and most are. There is no single rung a brand "is on".

Commercial modularity

Products do not have to be bought in rung order. A team can start with Product Readiness without buying Brand Visibility first, because the problem it solves is the problem that team owns.

Technical prerequisites

Some products do need something to work with: a product feed, an eligible signed-in experience, a permissioned shopper connection, a supported market. Those are implementation requirements for that product. They are not a requirement to have bought the ones before it.

The rungs work together. They are not a mandatory purchasing sequence. Start with the problem your team owns.

Different evidence for different jobs.

Each rung uses evidence appropriate to the job.

Rung Evidence What it does not mean

Be Visible Brand Visibility Observed model responses, cited sources and recurring answer patterns. Observed outputs are not hidden model reasoning, and they do not prove a later purchase.

Be Sellable Product Readiness Connected catalogue data and controlled Shopping Tests. Controlled tests are not the live consumer experience of third-party AI assistants.

Be Relevant Customer Intelligence Eligible options, permissioned connection context and Herm's internal decision process. The brand receives the decision, not the underlying shopper profile.

Be Preferred Offers Consented purchase outcomes and, where Offers is used, covered offer and revenue outcomes. Receipt verification confirms an outcome; it does not establish answer-level causality.

Receipts are how Be Preferred is read. Reading the other three off receipts would overstate all three, and would make the one measurement that genuinely uses them sound like the whole product.

Four dimensions, scored separately.

The AI Commerce Readiness Score reports each dimension on its own, with plain-language findings under each: "visible in 2 of 5 models," "product feed required." Brand Visibility inputs come from a fixed five-model panel (ChatGPT, Claude, Gemini, Perplexity and Grok) asked a sample of questions designed around your brand and the buyers it is trying to reach, personalization-free, run on demand when you request your score. Product Readiness requires the brand to provide its product feed so Herm can assess whether AI can use current product data correctly; Be Relevant uses the Customer Intelligence connection once the brand enables it, and Be Preferred is measured from Herm's consented receipt panel where receipt panel coverage is sufficient, currently Turkey, with fuller measurement once Offers is connected.

Where Herm does not have enough measured dimensions to support a meaningful overall Readiness Score, the result is reported as a partial readiness assessment showing the dimensions that were measured, rather than as a headline score. An unavailable dimension is one Herm does not yet have the evidence to score; it is not a failed one.

The public AI Visibility Index currently ranks 3,734 brands. A brand's visibility result comes from questions designed around that brand, so it is not graded against the Index, which puts one shared question set to a whole sector.

Become AI Commerce Ready Read the full methodology →

◍ herm · readiness dimensions

VELORA · skincare · partial readiness assessment

01 be visible 52 ▲6 · visible in 2 of 5 models

02 be sellable unavailable · product feed required

03 be relevant not yet eligible · connection not enabled

04 be preferred unavailable · panel coverage building

Unavailable does not mean failed. It means Herm does not yet have the required evidence to score that dimension.

illustrative example · 5-model panel · herm.io

Where should your team start?

Four dimensions, four teams that already exist. Pick the entry point that matches the problem you own. There is no required first rung.

Brand / Growth

Start with Be Visible

When You need to understand whether AI systems include and represent your brand.

Brand Visibility Live

Ecommerce / Product

Start with Be Sellable

When You need to know whether AI shopping systems can actually use your product information.

Product Readiness Live

Digital / Customer Experience

Start with Be Relevant

When You want your signed-in experience to choose better among eligible options without importing shopper profiles.

Customer Intelligence Early access

Revenue / Loyalty

Start with Be Preferred

When You want to understand category preference, repeat purchase and reward-driven outcomes.

Offers Live

The destination

Scale on AI.

Beyond the four rungs sits the destination: distributing structured product and commercial experiences into AI-mediated commerce as those surfaces open, from five frontier assistants today to the wider ecosystem of AI shopping apps and agents. It is where the framework points, not a fifth rung, and not something a brand can buy today.

See what is planned →

AI Distribution Roadmap

How this is measured

The methodology is public.

Everything above states what readiness means. The methodology states how it is measured, and where the measurement stops.

Measurement boundaries: what each rung does and does not assert
Inputs: what Herm reads, and which inputs need a brand-controlled connection
Availability: how an unavailable dimension is reported, and why that is not a failure
Interpretation: what a result means, and what it cannot be used for
Limitations: model variance, and the limits of receipt verification
Comparability: why Index and Brand Visibility results are never compared
Versioning: every material change, dated, so an older result stays interpretable

Read the methodology

This page defines the ladder. The suites climb it. Three of the four rungs are live today and Customer Intelligence is early access; the destination wears its true state.

See all five suites →

Questions

What is AI Commerce Readiness?

AI Commerce Readiness is the discipline of making a brand discoverable by AI, its products usable by AI shopping systems, its owned experiences relevant to permissioned shoppers, and the brand preferred by the people who buy its category: everything that has to be true for a brand to win when shoppers ask AI assistants what to buy. It is structured as four rungs: Be Visible, Be Sellable, Be Relevant, Be Preferred.

How is AI Commerce Readiness different from AI visibility?

AI visibility measures whether AI models mention a brand; AI Commerce Readiness is the wider discipline that visibility belongs to. Visibility gets a brand mentioned. Readiness gets it mentioned, sellable, relevant, and preferred. Visibility is one dimension of four.

Do the four rungs have to be bought in order?

No. The rungs are conceptually cumulative but commercially modular: a brand can start with the product that addresses the problem its team owns, and does not have to buy Brand Visibility before Product Readiness. Some products do have technical prerequisites of their own (a product feed, an eligible signed-in experience, a permissioned shopper connection, a supported market), but those are implementation requirements for that product, not a requirement to have bought the earlier ones.

How is AI Commerce Readiness measured?

Each of the four dimensions is measured separately, with plain-language findings under each. Brand Visibility inputs come from a fixed five-model panel (ChatGPT, Claude, Gemini, Perplexity and Grok) asked a sample of questions designed around the brand, run on demand when the brand requests its score. Product Readiness requires the brand's product feed and is read through controlled Shopping Tests. Be Relevant uses the Customer Intelligence connection once the brand enables it, and Be Preferred is measured from Herm's consented receipt panel where receipt panel coverage is sufficient, currently Turkey, with fuller measurement once Offers is connected. The public AI Visibility Index currently ranks 3,734 brands. The full scoring rules are documented in the Readiness methodology.

What happens if a dimension cannot be measured?

Which AI platforms does readiness cover?

Today, the five frontier surfaces where AI shopping conversations happen: ChatGPT, Claude, Gemini, Perplexity and Grok. AI Commerce Readiness as a discipline is broader than those five surfaces and is designed to extend to the wider ecosystem of AI shopping apps and agents as that market develops.

Who coined the term "AI Commerce Readiness"?

Herm introduced the term AI Commerce Readiness and first published this canonical definition on June 22, 2026. Herm maintains the definition, the four-rung framework and the public methodology at herm.io: the framework is published openly, and the scoring methodology is documented at herm.io/for-brands/readiness/methodology.

Who in an organisation is responsible for AI Commerce Readiness?

Typically four teams share it: brand or growth marketing (Be Visible), e-commerce or product (Be Sellable), digital or customer experience (Be Relevant), and the revenue or loyalty owner (Be Preferred), convened by the CMO or digital leader. Each dimension is reported separately so each owner has a result of their own.

You've read the definition. Get the diagnosis.

See the four dimensions applied to your own brand: what was measured, what was not, and which problem is worth solving first.

Become AI Commerce Ready Read the methodology