An Indian online store should check its actual channel eligibility before buying an “AI checkout readiness” project. It should also avoid the opposite mistake: assuming there is nothing available in India because an older announcement described a US-only launch.
The current programmes differ. Product discovery is not the same as checkout inside an assistant. The merchant's country is not necessarily the buyer's country. And an open protocol does not mean every customer can use every checkout experience.
As checked on 7 November 2026, Shopify's Meta channel documentation includes India among its supported buyer countries for direct checkout. Its Google channel guidance names the US, Canada and Australia. Its ChatGPT route sends shoppers to the merchant's own checkout and requires selling to US customers. Those distinctions should change what a supplier proposes to build.
What do people mean by AI checkout?
Separate three jobs: finding a product, managing a purchase session and authorising payment. A shopping assistant can perform the first without performing the other two. A customer may discover your product in a conversation and still complete an ordinary checkout on your website.
Shopify's agentic storefronts overview distinguishes referral shopping from Shopify-powered direct checkout on supported channels. It also explains that the available paths depend on the channel and store setup.
Ask a prospective supplier to describe the complete customer journey. Where does the shopper choose a variant? Where are shipping charges calculated? Where is payment authorised? Where does the confirmed order appear? “Your store will support agents” is too vague to answer any of those questions.
That distinction matters commercially. A product-data improvement and a custom transaction endpoint have different scope, operating costs and failure cases. Do not commission the second when the supported channel already provides it, or when the first is the actual blocker.
Which documented paths can an Indian store evaluate now?
The table summarises the specific sources below, not every AI shopping product in the market. Availability and requirements can change, so recheck the live channel documentation and your store's admin before acting.
| Documented path | Buyer-country requirement relevant here | Where purchase completes |
|---|---|---|
| Shopify's ChatGPT channel | Store must sell to US customers; merchant may be elsewhere | Merchant's online store checkout |
| Shopify's Google AI Mode and Gemini direct checkout | US, Canada or Australia; merchant may be elsewhere | Supported direct checkout, subject to rollout and eligibility |
| Shopify's Meta direct checkout | India is included among supported buyer countries | Supported Meta checkout surfaces, subject to eligibility |
| OpenAI's direct merchant-feed programme | Merchant page currently describes shopping for US users | Merchant-owned website or app |
Shopify's ChatGPT guidance says eligible products can be discovered through Shopify Catalog. Checkout takes place in the merchant store, in an in-app browser or a new browser tab. An Indian merchant selling to US customers is a different case from a merchant selling only domestically.
Its Google AI Mode and Gemini guidance names the US, Canada and Australia and describes a gradual rollout. Do not interpret a supported country as proof that a particular shop or shopper already has access. The guidance also requires eligible products and a valid Merchant Center setup.
Most importantly for a domestic Indian store, Shopify's Meta channel guidance explicitly includes India. It requires eligible catalogue products, applicable terms and store policies, with additional channel setup for shoppable ads. This is a reason to inspect your eligibility, not a guarantee that every Indian store, product or payment method is supported.
Does ChatGPT still require a custom in-chat checkout build?
Do not start from the original Instant Checkout announcement. OpenAI's current merchant page says it is moving away from a standalone Instant Checkout experience towards product discovery and merchant-owned checkout. It describes shopping as currently live for US users and offers an application route for product feeds.
That page still describes the Agentic Commerce Protocol as infrastructure for its commerce work. It does not justify saying the entire protocol is dead, nor does it mean an Indian store should build transaction endpoints without a confirmed use case.
For a Shopify merchant, first check the supported channel and catalogue path. For another commerce platform, establish what feed access or partner integration is actually available to the business. Do not buy an integration merely because someone can demonstrate a protocol against a sample store.
The acceptance condition should be concrete: an eligible product is discoverable in the intended experience, the shopper can reach the correct purchase path, and the resulting order can be handled by your team. A successful technical handshake is a narrower achievement.
Does an open protocol mean you should build your own integration?
No. Google's UCP overview describes an open standard and integration options, while also linking to a programme waitlist. Those are different layers: a specification can be public while a particular commercial experience has onboarding or rollout restrictions.
Equally, do not turn Google's waitlist into a claim that all developer access across every platform is closed. Shopify's developer announcement from June 2026 says its agentic commerce tools opened to the wider developer ecosystem. Evaluate the merchant platform, developer tools and buyer-facing channel separately.
A custom integration can make sense when the supported path exists and your requirements are not already covered by the platform. Ask what capability you need, which live experience will consume it, and who will maintain compatibility when that experience changes.
A project with no named supported destination is still exploration. That can be useful, but its budget and success criteria should say so rather than promise production orders.
What does UPI Reserve Pay establish about availability?
It establishes a payment capability to investigate, not universal access to every AI shopping channel. Razorpay's UPI Reserve Pay explanation, published in March 2026, describes a user-authorised reserved amount and multiple debits within the agreed boundary. The page includes an early-access route and describes demonstrations and pilots.
The earlier Razorpay, NPCI and OpenAI announcement described a pilot in October 2025. A dated announcement is evidence of that initiative; it is not, by itself, a current merchant eligibility list or a promise of general availability today.
Ask your payment provider about access for your business, supported flows, buyer requirements, authorisation, revocation, settlement and refunds. Do not assume that a channel listing India automatically means it supports your preferred UPI flow.
Keep the customer-authorisation boundary explicit. A shopping assistant selecting a product does not itself establish permission to debit the customer. Use the supported payment arrangement rather than asking a general-purpose agent to handle payment credentials or improvise consent.
What work is useful even before a new checkout integration?
Start with accurate product data. The product a shopper selects should match the variant, image, price, currency, availability and delivery promise in the actual checkout.
Google's Merchant Center product specification identifies problems such as missing variant information, incorrect identifiers and conflicting website and feed data. Fixing those issues supports existing shopping distribution as well as any future AI channel that consumes your catalogue. Do not invent a barcode or claim a variant is in stock simply to fill a required field.
Inspect several real product cases: a variant with different pricing, an unavailable item and a product with delivery restrictions. A clean home page says little about whether the catalogue describes those accurately.
Where relevant, use Product structured data generated from the same maintained information as the visible page. That supports applicable Google product features; it does not grant access to a separate checkout programme. Our structured-data guide explains why vocabulary validity, feature eligibility and actual search appearance are separate checks.
What should you test in the customer journey?
Test the route customers will actually use. For a referral channel, that means the handoff to your store, including any in-app browser. For direct checkout, test the channel's supported experience rather than assuming it behaves like your normal website.
Shopify's Meta documentation notes unsupported features including custom pixels and product bundles in direct checkout. Review those restrictions against your store. A working normal checkout does not prove that every extension or tracking arrangement carries across unchanged.
Use a defined set of cases:
- Selecting the correct product and variant, with an accurate final total.
- Applying the shipping and payment options supported for that buyer.
- Handling abandoned, declined and delayed payment outcomes.
- Creating the intended order once and making its status clear to fulfilment staff.
- Processing a cancellation or refund through the supported channel workflow.
Where payment verification is required, use the provider's supported confirmation process rather than trusting only a browser success screen. Keep accepted orders, paid orders and fulfilled orders distinct. Cash-on-delivery and other delayed-payment flows need their own agreed state transitions.
These are proposed acceptance checks, not claims that every listed scenario is available on every channel. Choose cases according to the actual supported features and use a controlled test process before exposing changes to customers.
How should you judge the results?
Keep discovery, checkout starts, completed orders, payment status and refunds separate. A product appearing in an assistant's answer is not a sale, and an order attributed to an AI channel does not prove that a custom integration caused incremental demand.
Check what the platform records for each channel and reconcile it with your order system. Where direct checkout does not support your usual tracking extension, agree how reporting will work before treating missing browser events as missing sales.
Do not forecast a percentage of revenue from AI channels without evidence relevant to your store. Run a limited evaluation, record the setup and product changes, and compare completed periods while accounting for promotions, stock and seasonal demand.
For Shopify development, start with the capabilities and limitations already supplied by the platform. For e-commerce development on another stack, confirm the destination and integration requirements before commissioning custom code.
What should you avoid spending money on?
Avoid a generic readiness package that cannot name your eligible channel, buyer country and purchase path. Avoid rebuilding the storefront solely because someone says agents require a particular framework. And avoid publishing old country restrictions as if they apply to every programme.
Equally, do not activate a channel casually because it appears in an admin screen. Review its terms, supported features, catalogue, measurement and fulfilment implications. Country eligibility is the beginning of the assessment, not the end.
The practical next step is small: identify one supported opportunity, inspect what your current platform already provides, and fix the catalogue or checkout gaps that prevent a reliable customer journey. Then decide whether any additional engineering has a measurable purpose.
Ask us to review your store's actual eligibility and checkout path if you are considering this investment. The answer may be a configuration change and better product data rather than a new protocol implementation.
Bharat Vasireddy
Founder & Digital Marketing Strategist8+ years of experience in performance marketing and web development. I help businesses scale through data-driven SEO, Google Ads, and high-performance websites.