Dose #200: How AI Search Actually Works, and What Subscription Brands Need to Do Now

Matt here with your weekly Subscription Prescription 💊

We’re diving back into AI search. Every brand I talk to is worried about improving this, so it’s worth another deep dive. Let’s talk about AI context, building out site content, and how conflicting information will kill the sale.

This week’s newsletter is also a podcast, so you can listen or read at your convenience:

Special Offer: Free AI Search Audit

Before we dive in, I want to make you an offer.

We're running free AI search audits for subscription brands right now. All we need is your website and a couple of customer personas. We run it through AI, surface what it finds — and what it misses — and give you a prioritized list of improvements you can actually act on.

Grab your free audit here: thesubscriptiondoc.com/audit

Now, onto this week's topic — because it connects directly to that.

A couple of weeks ago, I had my business partner David Bradley on the podcast to break down the technical side of AI search. I just got back from ChargeX, and this was all anyone was talking about. So I want to go deeper on the practical side — because most of what you need to do is simpler than you think.

Here's the stat that frames everything: 23% of Americans used AI to research products in 2025. That number is climbing.

If your subscription brand isn't showing up when AI does that research, you're invisible to a growing segment of buyers. Here's what to do about it.

The game has shifted from personalization to context

For years, the brand personalization playbook looked like this: run a targeted ad for a specific audience, drive traffic to a landing page that mirrors the ad, and let Meta's algorithm find the right people.

That model put the burden of personalization on the brand.

AI search flips it entirely.

AI already holds context about the user. It knows where they live, what they've been searching for, and what their goals are. When someone asks AI to recommend a protein powder for long hikes, AI goes and finds the best match based on everything it already knows about that person.

Your job is no longer to reach the right person with the right message. Your job is to make sure the right context exists on your site so AI can find it and surface your product.

That is a fundamental shift. Brands that understand it early will have a real advantage.

AI reads your whole website, but most humans don't

Here is a quirk worth understanding.

About 80 to 90% of site visitors never scroll past the fold. Maybe 5 to 10% make it to the bottom of a page. AI goes all the way down.

This means you have room to build context deeper in your pages without cluttering the experience for human visitors. Detailed FAQs, product use cases, shipping specifics, subscription program details — these can live further down and AI will still find them.

A few things to prioritize right now:

Your subscribe and save page. If you don't have one, build it this week. If you have one, audit it. It should explain your program clearly, lead with perks and benefits, include FAQs, and be consistent with how you refer to the program everywhere on your site. If your program is called Auto Ship, call it Auto Ship on every page, every URL, every mention. Consistency is not just a brand thing. It is an AI thing.

Shipping and delivery info. AI is not crawling through your checkout. If your shipping details only appear there, AI does not know them. Make it explicit on your product pages, your subscription explainer page, and your FAQs. If you ship frozen, say so. If you ship next day, say so. Be specific.

Product use cases — both positive and negative. Tell AI who your product is for AND who it is not for. If your supplement is designed for adults 30 and older and you have a version for younger customers, say that and link to it. This specific context helps AI route the right buyer to the right product instead of moving on to a competitor.

Conflicting information costs you the sale

AI is eager to get people answers. But when it finds inconsistencies across your site, one of two things happens: your brand gets ranked lower, or AI surfaces the conflict directly to the user.

Neither is good.

This is more common than you think. Help docs that reference old shipping windows. A product page that calls the program "subscribe and save" while the URL says "autoship." FAQs that contradict your subscription explainer page.

The fix is not complicated, but it takes attention. Use AI to audit your own site. Give it two or three customer personas and ask it to find everything it can about your brand. Ask what questions a potential subscriber might have that your site does not answer. You will spot the gaps quickly.

One technical note worth passing to your dev team: HTML is crawlable by AI. JavaScript is much harder. If your subscription frequency options or pricing only appear inside a JavaScript modal, AI may never see that information. Get it into the HTML wherever possible.

Where to start

This week: Build or audit your subscribe and save explainer page. Make shipping info explicit across your key pages. Run an AI audit using two or three customer personas.

This month: Review your product pages. Add specific use case context. Fix any conflicting information you find.

Over the next couple of months: Think about URL structure. Chewy and PetSmart are already building crawlable patterns like chewy.com/autoship/pet-food. Structured URLs make it easier for AI to index what you sell. Worth the investment.

The brands that win in AI search are not going to be the ones with the biggest ad budgets. They are going to be the ones with the richest, most consistent context on their websites.

You own that content. Build it.

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

 - Matt Holman 🩺

The Subscription Doc