What is agent buyability?

Agent buyability is how easily a personal AI agent can reach an online store, understand its products, and complete a purchase for a shopper. A store with high agent buyability gets chosen and checked out by agents. A store with low buyability gets skipped.

ARC is the agent-buyability grade for online stores. This post defines the term and shows how to check yours.

Why agent buyability matters now

People are starting to hand shopping to personal agents. The agent reads product pages, compares options, picks a size, and tries to check out. If it gets stuck, it moves on to a store where it doesn't.

On 6 October 2026, Sierra and Meta announced the Personal Agent Protocol (PAP). It is an open standard for how personal agents interact with businesses. Partners named in the announcement include Genesys, Instinct, Rocket, Shopify, Stripe, and Walmart. Sierra says it plans to publish the v0.1 specification later in October 2026 (Sierra, "Introducing Personal Agent Protocol").

Be precise about what PAP covers. The announcement focuses on authentication, consumer permissions, and giving companies visibility into what agents do. Payments extensions are listed only as a possible future addition. PAP does not enable agent checkout today, and the spec is not out yet.

What PAP does signal is direction: big retailers and platforms are planning for agents as a customer channel. Your store needs to work for agents on the rails it has now.

In ARC's October 2026 Index snapshot, 52 of 83 scored stores (63%) had at least one of five ARC browser cart shoppers fail to add the tested product to the cart, and at 21 of 83 none of the five managed it. Index tests stop at add-to-cart, and nothing is bought (State of agent buyability: October 2026).

The three parts of agent buyability

ARC breaks agent buyability into three checks. Each one can fail on its own.

1. Reach. Can the agent get to your store at all? Bot protection, CDN rules, and robots settings can block agents before they see a single product. A blocked agent can't buy.

2. Understand. Does the agent get your products right? That means names, prices, variants, stock, shipping, and returns. If an AI model tells a shopper the wrong price or a size you don't carry, you lose the sale or earn a return.

3. Buy. Can the agent finish the job? This is where most of the friction lives: variant pickers, add-to-cart buttons, pop-ups, and checkout steps that a person handles easily but an agent doesn't.

In that same October 2026 snapshot, the most common stop was the add-to-cart step itself. It was the published stop reason at 39 of the 52 stores where a shopper failed. The size or colour picker stopped 8, a human-check 3, and a bot block at the door 2.

Free scan, no login: arcreport.ai. See your reach, understand, and buy checks in about a minute.

How agent buyability differs from AI visibility tracking

AI visibility tools track whether a chat assistant mentions your brand. That's useful.

Agent buyability asks the next question: once an agent decides to shop with you, can it actually get to checkout? A mention without a working cart is not a sale.

The two fit together. Keep your mention tracker. It just can't see your cart. ARC grades the part that comes after the mention: reach, understanding, and checkout.

	AI mention / visibility tracking	Agent buyability (ARC)
Question	Does AI mention my brand?	Can an agent find, understand, and buy here?
Unit	Share of voice in answers	Whether an agent can get to the cart and checkout, plus the step where it breaks
Output	Prompt and mention dashboard	Score, stuck step, and a specific fix

How to check your store's agent buyability

You can check it for free in three places:

• Free scan at arcreport.ai. Enter your homepage. You get a score across reach, understand, and buy, plus specific fixes. No login, no card.
• Your store hub at arcreport.ai/store. A bookmarkable page with your Reachability Matrix, your ARC Index score, agent-facts about your store, and early detection of PAP markers.
• The ARC Index. A free public leaderboard of stores ranked by agent buyability.

The diagnosis is free, and so is everything else. ARC is an open research project.

For how the tests work, including their limits, read How ARC tests whether agents can buy from your store and the full methodology.

FAQ

What is agent buyability, in one sentence? Agent buyability is how easily a personal AI agent can reach your store, understand your products, and complete a purchase on a shopper's behalf.

Is agent buyability the same as AEO or AI search visibility? No. AEO and visibility tools measure whether AI mentions your brand. Agent buyability measures how far an agent gets toward buying from you: finding the product, the cart, and checkout. They complement each other.

Does the Personal Agent Protocol let agents check out on my store? Not today. The v0.1 spec is due later in October 2026, and payments are listed only as a possible future extension. ARC watches for PAP markers as early detection and claims no PAP support.

What usually stops an agent from buying? Blocked access, wrong or unreadable product data, and checkout friction such as variant pickers and pop-ups. ARC shows the exact step where its tests stopped.

How much does it cost to check? Nothing. The scan, store hub, Index and datasets are free. There are no paid plans.

---

Check your store now. Free scan, no login: arcreport.ai. Then see your store hub: arcreport.ai/store.

ARC is independent and not affiliated with Meta, Sierra, or any agent vendor.