Dose #197: Agentic Commerce - How AI Will Discover (Or Miss) Your Subscription

Matt here with your weekly Subscription Prescription 💊

This week I sat down with my business partner David Bradley to talk about something I haven't heard anyone else in the subscription space address directly. Not churn. Not LTV. Not onboarding flows. AI discoverability. Specifically: how do you get your subscription offer to show up when an AI agent is shopping on someone's behalf?

Check out the full interview with David and hear everything he has to say on AI discoverability and agents for subscriptions. Listen in or watch on your favorite platform:

Most brands are starting to think about AI search. Will my products show up in ChatGPT? Will an AI agent recommend me over a competitor?

That's the right question. But most people are stopping at the product level. They're not thinking about the subscription offer specifically.

Here's the shift that's happening. Last October, Google introduced a new structured data property for physical goods sold on subscription at a subscription price. That is the first time Google has formally acknowledged that subscribe and save is a distinct commerce event worth surfacing in search results. That is a massive signal.

Google controls a large share of how products get discovered online. When Google adds a new data property for something, the rest of the industry follows. And AI agents are already building on top of these same structured data standards.

If you are not structuring your subscription data now, you are going to fall behind.

Your Explainer Page Is Your AI Front Door

AI agents are not patient. If you ask one to go research a thousand products, it will take a random sample and tell you it's done. You have to push it to do the full job.

When an AI agent scans your website, one of the first things it looks for is a page that clearly explains your subscription program. What it is. What the benefits are. What changes a subscriber can make. 

This is your explainer page. And if you don't have one, your subscription offer is invisible to AI.

But here's the part most brands miss: the name on that page has to match exactly how the subscription is labeled everywhere else on your site. If your product page says "Auto Ship" and your explainer page says "Subscribe and Save," you are sending a confused signal. AI picks up on consistency. Inconsistency slows it down and weakens the recommendation.

Takeaway: Build a dedicated subscription explainer page if you don't have one. Name it exactly what you call the program on your product pages. Treat it as table stakes.

The Data Flip You Need to Understand

Here's the insight from this conversation that I keep thinking about.

For years, brands have been capturing customer data through quizzes, keyword searches, surveys, and product filters. You collect their intent, and you use it to serve them the right product.

That model is being flipped.

In agentic commerce, AI already knows the customer. It knows their dog's name, breed, age, and what bag size they ordered last time. It's not waiting for the customer to tell it anything. It's shopping on their behalf using context it has already built.

Your job as a brand is no longer to capture customer data. Your job is to give the AI the tools it needs to match your product to the customer it already knows.

That means your product pages need to do more work than they have before. Specifically, they need to clearly express:

Who this product is for. What size, variant, or formulation fits what kind of customer. What subscription frequency makes sense given how the product is used. What the benefits of subscribing are at the product level, not just on a general program page.

Chewy and PetSmart are doing this better than almost anyone right now. When you search for dog food in ChatGPT, they show up consistently because their product data is structured to answer the AI's question before it even has to ask. 

Takeaway: Audit your product pages. Are the subscription benefits explained at the product level? Does the data connect customer type, variant, frequency, and offer clearly? If an AI agent is trying to match your product to a specific customer profile, can it do that without clicking through five pages?

Negative Context Is Just as Valuable

This was one of my favorite moments from the workshop. A brand owner in the audience asked: could I use my bad reviews to build negative context for my products?

The answer is yes. Absolutely.

Think about how negative keywords work in paid search. You tell the algorithm what your product is not for so it stops showing up in the wrong searches. The same logic applies to AI discoverability.

If your protein powder consistently gets bad reviews from people who tried to bake with it, that tells you something. You can add context to your product page that says this product is not intended for cooking or baking. That negative signal helps AI route around mismatched use cases and surface your product to the right person.

On the positive side, your best customer reviews are a goldmine of structured data waiting to be extracted. Who leaves five-star reviews? What are they doing with the product? At what frequency are they subscribing? What life situation or goal connects them to your product?

That data can be organized and embedded into your product page structure. It helps AI connect the dots between your best customer and the offer that fits them perfectly.

Takeaway: Read your reviews with AI discoverability in mind. Use positive reviews to build customer profiles. Use negative reviews to add honest context about who the product is not for. Both improve how accurately your offer gets recommended.

One More Thing: Your Catalog Structure

This one is simple but underused.

Look at how Chewy and PetSmart organize their product URLs. You will see structures like chewy.com/autoship/small-breed/senior. That is long-tail SEO built into the catalog itself.

In the past, building this kind of URL structure was labor-intensive and hard to scale. AI can help you do it now. You can work with your AI to build out collection pages and category structures on Shopify or WooCommerce that make it easier for agents to navigate your store.

Think of it like a well-organized library. If the books are just stacked in a pile, the librarian has to dig. If they're categorized by genre, age group, and topic, they find the right book in seconds.

AI agents work the same way. Give them a catalog that's organized, labeled, and structured, and they will find your subscription offer faster and recommend it more confidently.

Bottom Line

AI search is not a future problem. It is a right now problem.

The brands that get discovered in agentic commerce are going to be the ones with clean explainer pages, structured product data, clear frequency and variant correlations, and catalog organization that lets AI navigate quickly.

This is not a massive technical overhaul. Start with your explainer page. Audit one product page. Mine your reviews for customer profiles. Add one layer of structure to your catalog.

Do that this week. You will already be ahead of most brands in your space.

Until next Tuesday, that’s your Subscription Prescription. 💊

 - Matt Holman 🩺

The Subscription Doc