InMobi’s Agent-Readiness Scan Exposes the E-Commerce Gap AI Shopping Agents Will Hit

AI shopping agents are beginning to move from product research to checkout, but most online stores may not be ready for an agent to complete the journey. On October 5, InMobi launched “Ready, Set, Agent,” a free storefront assessment, alongside research covering three million U.S. e-commerce merchants. Its findings offer a timely warning for retailers: product visibility in AI answers is only useful if an agent can also understand the offer and carry out an authorized transaction. InMobi’s announcement and findings make agentic commerce readiness a concrete operational question—not just a future-facing GEO talking point.

The headline needs a closer read

InMobi reports 36% of storefronts in its breakdown as “Superagent,” able to support discovery through payment; 15% as “Special Agent,” with cart and checkout capability but discovery or availability gaps; and 12% as “Field Agent,” where agents can read a catalog but the transaction path is incomplete. The remaining 37% are on platforms InMobi says it has not yet scanned. The company’s release frames the overall picture as nearly two-thirds not ready, but its own categories distinguish partial readiness from an unscanned population. Retailers should treat that 37% as unknown—not evidence that every one of those stores fails an agent checkout test.

That nuance matters. The useful takeaway is not a universal industry failure rate; it is that basic product discovery and end-to-end transacting are separate capabilities, and neither can be assumed from the other. InMobi says its free tool runs 80 live probes across 11 categories, shows the individual findings behind a score, and received more than 8,600 on-demand scans during beta. Those are vendor-reported results, so the assessment is a starting point for diagnosis, not an independent certification or proof of future sales.

Agentic commerce is shifting from concept to checkout

The timing is significant because major platforms are building the transaction layer. Google introduced the open Universal Commerce Protocol (UCP) with retail partners and described native checkout on Google surfaces as an initial use case; its stated goal is to let retailers participate in agent-driven shopping while retaining the merchant relationship. Google’s NRF announcement shows that product discovery and transaction infrastructure are being developed together.

Shopify has also moved beyond catalog search: in September it announced that eligible browser-based AI agents can inspect and update checkout, then submit a purchase with the buyer’s authorization through WebMCP tools. TechCrunch’s report on Shopify’s release describes how those tools connect product discovery, cart changes and checkout. Meanwhile, Stripe’s agentic-commerce documentation lists UCP and the Agentic Commerce Protocol (ACP) as seller integration options. Stripe’s technical guide makes clear that agents need a structured way to access catalogs and complete payment—not just a persuasive answer about a product.

For marketing leaders, this creates a new handoff point in the funnel. An AI assistant may recommend a product, but stale availability, missing attributes, inaccessible product pages or a broken cart handoff can stop the journey before a sale. GEO and SEO can help make products discoverable; operational readiness determines whether discovery can become an authorized order.

What retailers should test before scaling

Start with the product data an agent needs to make a reliable recommendation: accurate titles, variants, dimensions, compatibility, price, inventory, shipping and returns. Keep those details consistent between product pages, feeds and checkout. Confirm that approved AI agents can reach the relevant public pages and structured information, while maintaining appropriate controls over restricted content and customer data.

Then test the transaction path, not just whether a bot can read a product page. Can it select the right variant, build a cart, preserve the displayed price, apply relevant shipping or tax information, and hand off to a checkout that requires buyer authorization? Walk through declined payments, out-of-stock items and changed delivery options too. A readiness score can identify friction; it cannot substitute for end-to-end tests against the retailer’s own commerce stack.

Finally, instrument the journey. Separate AI referrals and agent-assisted sessions where possible, and track product views, cart creation, checkout completion, cancellation and contribution margin. Use a limited set of products and one controlled test before committing engineering resources across the catalog. Treat platform-reported traffic or scan counts as signals to investigate—not evidence that agentic traffic will convert at a particular rate.

A practical 30-day action plan

Week one: Run a storefront assessment, document the findings and verify the most important products manually. Week two: fix high-impact catalog gaps, inconsistent inventory, crawler access and broken product-to-cart paths. Week three: ask your commerce platform or payment provider which agent-facing integrations—such as UCP or ACP—your stack supports, and test authorization, order records and customer-service handoffs. Week four: launch a controlled pilot with a small product group, clear success metrics and human oversight; review failed handoffs as carefully as completed orders.

InMobi’s study is a useful prompt to audit the full commerce journey, but not a reason to assume every retailer has the same problem—or that a scan score guarantees performance. Businesses that make product information dependable, checkout testable and measurement explicit will be better positioned as AI discovery becomes transactional.

Want help evaluating AI search visibility and preparing your digital marketing operation for agent-driven commerce? Contact Real Internet Sales at 803-708-5514 or visit realinternetsales.com.