In this post, I’m recapping the AEO strategy framework that Ethan Smith, CEO of Graphite, shared in a recent guide on how to implement AEO in your business.

The framework covers four stages:

1. Question Research

2. Question Tracking

3. Developing a strategy for each platform

4. Running experiments

This post covers stages 1 and 2. For stages 3 and 4, see AEO Strategy: Part 2.

Why this matters: AEO is still early enough that the businesses building systematic processes now will have a meaningful head start. The framework below is as close to a repeatable playbook as the space currently has.

Stage 1: Question Research

Question research is the AEO equivalent of keyword research. The goal is to identify and group the questions your customers are actually asking AI engines, so every optimization tactic you build is anchored to real demand.

This stage is the foundation. Without it, you’re guessing what to optimize for.

Why Question Research Is Harder Than Keyword Research

Unlike traditional search keywords, AI query data isn’t publicly available. The platforms don’t expose it (yet — that will likely change once they roll out advertising). So you can’t just open Ahrefs and pull a list.

There are three practical ways to build your question bank:

1. Extrapolate from your existing SEO keywords. Take your current keyword list and convert them into question format. If you rank for “AEO question tracking tool,” the question version is “What’s the best tool for tracking AEO questions?” This is the fastest starting point and requires no new research budget.

2. Mine your customer touchpoints. Your customers are already asking these questions — just not in a search bar. The places to look:

• Sales call recordings and transcripts

• Support tickets and help desk threads

• Customer reviews and feedback forms

• Community forums and discussions (Reddit, Slack groups, Discord servers)

Ethan specifically flagged community building as underrated here. AEO has a longer tail than traditional SEO, meaning there are more niche, conversational questions to capture. A community surfaces questions you’d never think to track on your own.

3. Purchase question data from tracking platforms. Tools like Semrush now aggregate AI search query data. This is the highest-signal source, but it comes at a cost. Worth it once you’ve exhausted the free methods.

Target: Compile 100-200 questions before moving to the next stage. That’s enough to identify patterns and build a meaningful tracking baseline.

Stage 2: Question Tracking

Once you have your question list, you need to know how AI engines are currently answering those questions — and specifically, who and what is showing up in the citations.

Choosing a Tracking Tool

The tracking tool market has exploded. Graphite maintains a comprehensive list of AEO tools that’s worth bookmarking. At last count, over 50 tools have emerged in this space.

The practical advice: most tools at this stage offer similar core functionality, so don’t overthink the selection. Pick the least expensive option that covers your target questions and the AI platforms most relevant to your audience (at minimum: ChatGPT, Perplexity, and Gemini).

What to Actually Track

There are two things that matter when you’re monitoring citations:

1. Who is showing up? Which brands, domains, and sources are being cited for your target questions?

2. What type of content is appearing? The format and source type of citations tells you what AI platforms are prioritizing.

Understanding Citation Source Types

Citation sources break into two buckets:

Within off-site citations, there are four distinct groups you’ll regularly see:

• Videos (YouTube, Vimeo)

• User-Generated Content / UGC (Reddit, Quora, forums, community sites)

• Affiliate and editorial sites (Dotdash, Meredith, Good Housekeeping)

• Blogs (Medium, Substack, independent publishers)

This breakdown matters because it directly shapes your content strategy. If UGC is dominating citations for your target questions, creating polished branded content alone won’t move the needle — you need a presence in those communities. If YouTube is showing up consistently, video becomes a higher-priority investment than it might be for traditional SEO.

The point of tracking isn’t just visibility — it’s intelligence. Understanding which citation types AI platforms are rewarding for your specific questions tells you exactly where to focus your content efforts.

Stages 3 and 4 (platform strategy and running experiments) are covered in AEO Strategy: Part 2.