How We Turned AI Search into a Real Growth Channel at Webflow

From <2% CMS market share to owning ~60% of AI answers in the category.

Most teams still talk about “being late to AI.”

But the reality I’m seeing is more basic than that:

It’s not that you’re late.
It’s that models can’t actually use your content.

At Webflow, we didn’t change our product.
We changed how models read our site.

And that’s produced some wild outcomes:

• Webflow now owns ~60% of AI answers in the CMS category… with <2% CMS market share

• ChatGPT traffic converts ~6x higher than non-brand Google (24% vs 4% CVR)

• LLM-attributed signups are up ~129% YoY and customers are up ~8x YoY (Q3)

• Visit-to-signup for LLM-attributed traffic is ~50% higher than our overall average

Same product. Same market.
Different structure and systems.

This post is a walkthrough of one of the highest-leverage levers inside that system: using schema and structured content as an AEO (Answer Engine Optimization) primitive.

At the end, I’ll link to a much deeper playbook I published with Aakash Gupta where we go into the full AEO setup.

---

From “SEO side quest” to AEO as a real channel

For a long time, AI search lived on the edges of the growth conversation.
Interesting, but not material.

That changed when we did two things:

1. We treated AEO as a first-class channel (with forecasts, dashboards, and owners)

2. We made a deliberate bet on structure, not just “more content”

Once we did that, we started seeing:

• LLM-attributed signups sharply up and to the right

• LLM customers compounding

• AI-driven segments that behaved very differently from traditional search segments

If you strip it all down, a big part of the lift comes from one question:

“Can models reliably understand what we do, who we’re for, and how to map us to the right intent?”

That’s where schema comes in.

---

Why schema is such a big lever for AEO

Most marketing teams think of schema as an SEO checkbox.

In an AI-search world, schema is a way to make your site machine-readable:

• It turns pages into explicit facts and entities: questions, answers, products, prices, organizations, authors, etc.

• It increases the odds you’re cited in systems like ChatGPT, Perplexity, Claude, and Gemini

• It creates a technical moat while competitors stay “fuzzy” and under-specified

One focused schema rollout at Webflow produced:

• FAQ schema added to 6 core feature pages

• 300+ new AI citations in ~90 days

• +24% SEO impressions on those pages

We didn’t rewrite the product.
We didn’t even rewrite all of the content.

We made the content that already existed far more legible to both search engines and LLMs.

---

The collaboration with Aakash (and why he’s one of the key voices here)

I’ve known Aakash for a while.

We first joined forces at Affirm during a wild growth chapter — him a powerhouse in the product org, me leading Growth & PMM. It was obvious even then that he’d become one of the most important voices in tech on AI, product, and growth.

When we started to see AI search become a measurable acquisition channel at Webflow, it made a ton of sense to collaborate on a public playbook.

The result is a full AEO guide we published together that digs into:

• How we measure AI search as a channel

• How schema, content, and experimentation fit together

• What’s actually worked vs. what’s just hype

I’ll link it at the end.

---

Why FAQ schema (and friends) punch above their weight

There are a lot of schema types, but a few have been especially impactful for us in an AEO context:

FAQ schema

• Maps real user questions to clear, structured answers

• Gives models a clean Q&A graph to reuse

• Was responsible for that 300+ citations in ~90 days on 6 feature pages example

Article schema

• Clarifies authorship, topical authority, and entities

• Helps connect deeper content (guides, explainers) into the knowledge graph

HowTo schema

• Works extremely well for “how do I…” or “step-by-step” intent

• Aligns nicely with onboarding, implementation, and migration content

When you pair these with truly good content (this is where partners like Graphite matter a lot), you get both:

• Human legibility (people actually read it and convert)

• Machine legibility (systems can reuse it confidently)

One without the other doesn’t move the numbers nearly as much.

---

A 3-hour weekly schema workflow that compounds

Schema can sound like a heavy lift, so here’s one slice of the program that’s surprisingly manageable and high impact.

As one part of the broader AEO motion, we run a simple weekly loop that looks like this:

1. Mine ~50 real questions with Gumloop
   We pull from Reddit, search data, and internal logs to find the actual language people use.

2. Use AirOps to generate + refine FAQ page content and JSON-LD schema
   AirOps helps us produce both the FAQ content and the corresponding schema in a structured workflow. Humans still edit and approve.

3. Deploy the updated FAQ pages + schema with AirOps
   We push changes into our stack, keeping everything versioned and consistent.

4. Validate via Google’s Rich Results Test
   To make sure our schema is valid and actually eligible for rich results.

5. Track impact in Profound + our own data stack
   We track pre/post citations in Profound, and conversion with our own analytics and warehouse data.

Partners like Graphite help ensure the underlying content deserves to rank — schema can’t save thin or low-intent pages.

This loop alone doesn’t explain every number I shared at the top.
But it does compound quickly when you run it every week for a quarter.

---

If you’re starting from scratch, do this first

If all of this feels overwhelming and you don’t have the luxury of a big team, here’s where I’d start:

1. Pick your top 5 pages by intent, not just traffic
   (Think: your best “money pages” where you actually convert.)

2. Add real FAQs based on real questions
   Pull from support, sales calls, Reddit, search logs. Use that language.

3. Add FAQ schema
   Use JSON-LD, validate it with Rich Results Test.

4. Track citations and signups for 90 days
   Use a tool like Profound plus your own analytics to see what’s actually moving.

If you do just that consistently, you’ll be ahead of most of the market.

---

The full playbook with Aakash

This post is just one zoomed-in slice of the larger AEO picture.

In the full guide with Aakash Gupta, we cover:

• How to think about AEO as a channel (not a gimmick)

• The broader system beyond schema: content, experimentation, entity strategy, and measurement

• More detail on how we’re tracking AI search performance and attribution

👉 Read the full AEO playbook here