Stripe has been one of my favorite B2B brand sites, and they just redesigned. The previous version ran for six years, which is forever on the web, and set a standard for considered design most brands are still chasing.

Katie Dill, Stripe's Head of Design, sat down with Y Combinator recently to walk through the new site. And watching it, I kept coming back to a question Patrick Collison apparently asked in a meeting during the redesign process: What is the point of a website, anyway?

It sounds simple…but it isn't. And the answer has changed in ways most brands haven't caught up to yet.

‍

‍

The old model: add sections, hope people scroll.

Stripe's previous site grew the way most do. Clear at launch, then layered over time. A new product here, a new audience there. By the end, you were getting glimpses of what Stripe did without ever getting the full picture. AI was one of their biggest growth areas, and it was showing up as a tile, and that wasn’t enough to create the signal Stripe needed in this new world.

Your website is where authority gets established, for humans and for AI.

A site is the owned space where authority is built. For customers, for prospects, and increasingly, for the AI systems forming opinions about your company before anyone visits a page.

“You’re demonstrating who you are, what you are doing, and why you do it. What colors you choose, what typography you choose, the details that you choose to care about and the ones you don’t, and then, of course, how you articulate what you’re doing and what you offer.” — Katie Dill, Head of Design, Stripe. Watch the full segment at 5:30

And when Stripe rebuilt theirs, they weren't just refreshing visuals. They were clarifying a story that the business had grown beyond.

The job has already been clarity, it’s just now, your audience has doubled from humans to humans + bots.

‍

‍

Specificity is what makes AI cite you.

The new Stripe site has a live counter showing what percentage of global GDP runs through their platform. It signals trust. It also signals authority to the AI systems deciding which sources to cite.

AI systems cite sources they consider authoritative. The signals they use to determine authority include the consistency and specificity of claims a brand makes about itself across the web. A live counter showing global GDP processed is a specific, verifiable claim. It's the kind of detail that trains the model to associate Stripe with scale and reliability.

Stripe built it to signal trust. The AEO benefit is a byproduct. That's how brand clarity and AEO credibility point in the same direction. For a deeper look at how AI rankings for commercial prompts actually work, and how specificity factors in, that post is a good place to start.

An AEO-ready website isn’t one with more content. It’s one with content that’s specific enough that an AI system can accurately describe what you do, who you serve, and why you’re different in a single sentence. 

Building for two audiences means rethinking what "visible" means.

This is where designing with AI changes the brief. When I work on a Webflow build with my team at Edgar Allan, we're thinking about two audiences at once. The human who lands on the page and the AI that reads the code underneath it.

The two audiences experience your site completely differently.

AI doesn't scroll. It doesn't trigger animations. It doesn't see the hero image or feel the weight of the typography. What Vivian Huang at Webflow discovered when she audited their own site is that content hidden behind scroll triggers and JavaScript reveals might not exist to a machine at all. If your most important message only appears after an interaction fires, the AI crawling your page may never see it.

Stripe solved this by leading with content present from the start and being more direct on the page. Katie Dill noted that this made the site feel more approachable for humans too. Clarity tends to do that.

The way we start every AEO-focused Webflow project at Edgar Allan: run the existing site through ChatGPT, Claude, and Perplexity and ask each one to describe what the brand does, who it serves, and how it's different. What comes back is what prospects are seeing before they ever visit the site. That gap between the brand's self-description and the AI's representation is the brief.

From there, the work is about closing the distance. Semantic HTML that communicates structure clearly. Content written for humans first, but structured so machines can extract it accurately. Schema markup, clean metadata, and a content hierarchy that doesn't depend on JavaScript to make sense. The Webflow AEO checklist we published covers the specific technical and content fixes that move the needle fastest.

Profound has become a core part of how we work. It surfaces which prompts a brand appears in, which ones it's missing, and gives us a measurable baseline to work from. AEO without measurement is just guessing.

Ask the question before every project, not after.

So, what is the point of this website? For this brand, at this moment, for these two audiences, human and machine, what does it need to say and how does it need to say it?

The brands getting this right aren't treating AEO as a separate workstream from brand and design. They're building sites where the story is clear enough that a machine can summarize it accurately, a human can feel it immediately, and an AI system defaults to citing it when someone asks a question this brand should own.

That's the whole task. And it starts with a question that Patrick Collison asked in a meeting that Stripe spent a year trying to answer. We've also written more on what happens to websites when AI shifts discovery upstream entirely, and it's worth reading alongside this: When Discovery Disappears.

Watch the full video here.