Agentic Commerce Readiness Checklist for 2027

To be ready for AI shopping in 2027, fix your product data first, get Google Merchant Center approved and clean, then extend one master catalog to Microsoft, ChatGPT-related options and marketplaces. This checklist is in the order we recommend tackling it, with the reason each step sits where it does.

It is written for ecommerce managers who sell physical products. If you want the background first, read agentic commerce explained; the service behind this list is our Google Merchant Center agency.

1. Is your product data ready?

Product data comes first because every other channel reads it. Weak data makes every later step less effective.

• Every branded product has a valid GTIN and consistent brand.
• Titles start with what the product is and include key attributes.
• The attributes shoppers ask about are filled: size, color, material, compatibility.
• Price and stock update as often as they change.
• Shipping times, costs and return policies are defined.

Details in the product data AI shopping agents read.

2. Is Google Merchant Center in good shape?

Google Merchant Center is the base the other channels borrow from, so it should be clean before you copy it anywhere.

• Account verified, with no open warnings.
• Product issues reviewed and fixed.
• Free listings enabled.
• Shipping and returns configured to match the site.

3. Is Microsoft Merchant Center set up?

Microsoft is usually the quickest second channel, because it can reuse your Google catalog.

• Store created and website verified.
• Catalog imported from Google or uploaded, with scheduled updates.

See the Microsoft Merchant Center and Copilot setup guide.

4. What about ChatGPT and other assistants?

Here the rule is to enable what makes sense and make your pages readable. No one can promise placement in an assistant’s answers.

• Your platform’s options for AI discovery checked and enabled where they make sense.
• AI crawlers allowed to access product pages.
• Product structured data on every product page, matching the feed.

See how to get your products into ChatGPT Shopping.

5. Have you decided on marketplaces?

Marketplaces are a business decision first and a data task second.

• Decision made on Amazon, Walmart and TikTok Shop, based on your products.
• Catalog adapted to each marketplace from the master feed.

See TikTok Shop vs Amazon vs Walmart.

6. Is there one catalog behind everything?

One source of truth prevents the most common failure: each channel quietly holding a different version of your products.

• One source of truth for product data.
• One master feed, built to pass Google first.
• Channel outputs generated from it.

See one product feed for every AI channel.

7. Are you measuring the right things?

Measurement keeps the effort honest, because most AI-assisted shopping still ends on your site.

• Traffic and new customers from AI assistants tracked separately.
• Conversion of those visits measured on your site.

Why this matters: Discover in AI, buy on your site.

How ready are you, section by section?

Score each area honestly. The table shows what each level looks like, so you can see where to spend the next month.

In what order should you work through the list?

Follow this sequence if you are starting from scratch. It is the order that avoids rework.

1. Export the catalog and fix identifiers and brand consistency.
2. Fill the attributes shoppers ask about, product type by product type.
3. Rewrite titles to lead with the product type.
4. Align price, stock and shipping between the site and the feed.
5. Get Google Merchant Center clean and approved.
6. Add structured data to product pages and compare it to the feed.
7. Import the catalog into Microsoft Merchant Center on a schedule.
8. Decide which marketplaces fit your products and adapt the catalog.
9. Set up separate reporting for AI assistant traffic.

How does the list apply in a hypothetical store?

This is an illustration, not a client.

Imagine a store with a few thousand SKUs of home furnishings. Google Merchant Center is connected, but diagnostics show open product issues, and half the products have no material or dimensions. Microsoft has never been set up, and the marketplace question has never been discussed.

By the table, product data and Google Merchant Center are “not ready” and everything after them is blocked. The right first quarter is data and diagnostics, not new channels. Microsoft follows because it can reuse the cleaned catalog. Marketplaces wait for a decision based on the products themselves.

What mistakes slow stores down, and how do you spot them?

These patterns show up repeatedly when a store tries to do everything at once. Each has a symptom you can find in your own data.

• Starting channels before the feed is clean. Symptom: the same product issue appears in Google and Microsoft diagnostics.
• Keeping several catalogs. Symptom: the price of the same product differs between two channels.
• Treating structured data as a one-time job. Symptom: markup shows an old price after a promotion ends.
• Blocking crawlers by accident. Symptom: product pages are reachable by browsers but a robots or security setting turns bots away.
• No owner. Symptom: nobody can say who last changed the shipping settings.

Review this short list once a quarter. It takes less time than fixing the problems after they spread across channels.

When can you skip parts of this list?

You do not need every step.

• If you sell only a few products, a manual review of data and Google Merchant Center is enough.
• If you sell only on one marketplace, step 5 is your whole job and the rest is optional.
• If your products are custom or made to order, identifiers matter less and attributes matter more.

What should you ask your team or agency before starting?

• Who owns the product data, and where does the master version live?
• How many products currently have open issues in Google Merchant Center?
• Which fields do we fill by hand, and which come from a system?
• Can we see AI assistant traffic separately in analytics?
• What is our plan if a channel changes its requirements?

Where do you begin?

If you only do one thing this quarter, fix your Google Merchant Center feed. Compare the channels in Google vs Microsoft Merchant Center vs ChatGPT Shopping, and read ChatGPT vs Gemini vs Copilot to see how each assistant behaves today.

How does Avafa work?

Avafa is a consultancy based in Spain that works remotely. We audit the catalog against this list, fix the data, and build one master feed for each channel. We do not promise placement in any AI answer, and on marketplaces our work covers catalog and listing data only.

Start with our Google Merchant Center agency service, or see our Microsoft Merchant Center agency and ChatGPT Shopping agency pages for the other channels.

FAQs

How long does it take to get ready?

It depends on the state of your catalog. Stores with a clean Merchant Center feed are already most of the way there.

Do small stores need all of this?

Start with product data and Google Merchant Center. The rest can follow as you grow.

Why is it called a 2027 checklist?

It is a planning horizon: the steps are the ones a store should have in place by then. The order matters more than the label.

Can I skip Microsoft and go straight to ChatGPT?

You can, but a clean Google catalog is the usual base for both. Microsoft is often the cheapest next step because it reuses that catalog.

What is the single most common blocker?

Inconsistent product data, such as missing identifiers and blank attributes. Fixing it once helps every channel.

Do I need a developer?

For structured data and feed automation, often yes or at least some technical help. Much of the data cleanup can be done by the catalog owner.