How I approach AI SEO
I investigate how ChatGPT, Perplexity and Google's AI Mode choose their sources, and publish what I can verify in my research tracker. That research informs the work I do with Andy Chadwick and our team at Snippet Digital.
AI SEO, also called generative engine optimisation (GEO) or answer engine optimisation (AEO), is work to improve how a brand appears in AI answers. It covers whether you're mentioned, which pages get cited and whether the description is accurate. We can investigate and improve the inputs. We can't guarantee what an engine will say.
How AI answers choose brands
An AI answer can name a handful of brands and link a few sources. Being present in a search index doesn't mean your brand will make it into that answer.
I look at what the model believes about you from training, what it reads when it searches, and what the rest of the web says about you. Rewriting your own pages addresses only part of that picture.
What the model believes about you
Baked in at training. Which brands it already associates with your category, before it looks anything up.
What it reads when it searches
Fetched live per question. Whether your pages get retrieved, parsed cleanly, and survive the citation filter.
What the rest of the web says
Reviews, Reddit threads, comparisons and mentions. What sources beyond your own site say about your brand.
A mention and a citation are different outcomes. I investigate both.
Testing it on our own brands
On 21 August 2026 we probed ChatGPT, Gemini and Claude about our own brands, 195 samples in total. Relevance to AI SEO: 100%. Keyword Insights, the software I co-founded, sits fourth in the models' own associations for keyword clustering, ahead of SE Ranking and Moz. This site is the fourth most cited domain of the 301 that appear in AI answers across the categories we track. Those figures describe that sample and date.
My research tracker records how ChatGPT and Perplexity choose their sources, updated as the engines change. It separates what I can read directly from the network traffic from patterns in a small sample. That distinction carries through to client measurement, before the work and after it.
What the work involves
- 1
Structure content so models can lift it. Answers first, entities named consistently, service pages a machine can quote.
- 2
Keep the technical side solid. GPTBot, ClaudeBot and PerplexityBot need to fetch fast and parse clean. We audit crawler access, rendering and speed.
- 3
Say it where the models listen. Queries mapped to the platforms that feed answers: YouTube, Reddit, Quora. Reddit is our deepest bench, in the open and disclosed where disclosure is due.
- 4
Build the consensus. Owned content, contributed articles, interviews and partnerships, until sources beyond your site agree on your story. Models repeat agreement.
- 5
Track the answers. Citations, mentions and share of voice across ChatGPT, Perplexity and AI Overviews, reported monthly, including what didn't move.
Who it suits
Software, services and considered-purchase ecommerce, anywhere buyers ask AI for recommendations. Also brands that rank well but never get named, and brands AI describes wrongly. We're the wrong fit if you want a schema sprinkle or a guaranteed placement; nobody controls what a model says, and anyone promising that is guessing with your budget.
AI SEO services through Snippet Digital
Client work goes through Snippet Digital, the agency I founded with Andy Chadwick. We also built Keyword Insights, and this site hosts my free AI SEO tools and the research tracker above. My work has appeared in Moz, Search Engine Land, Search Engine Journal, Ahrefs, Semrush and the Financial Times, and Search Engine Journal voted me one of the top SEOs to follow. I'm directly involved on every engagement, working alongside the team.
What it costs
Engagements start at $5,000 a month, scoped by how many topics, platforms and markets we cover.
Questions we actually get
How is this different from traditional SEO?
Traditional SEO focuses on visibility in search results. AI SEO (GEO/AEO) also works on whether a brand is named and cited inside a generated answer, which depends on what models believe about you and what the wider web says.
Do I still need traditional SEO?
Yes. AI assistants retrieve from search indexes. Traditional SEO keeps you in those indexes. AI visibility is the work on top of it.
How do you measure success?
Citations of your pages in AI answers, mentions of your brand across the questions that matter, and share of voice against competitors, tracked monthly alongside AI referral traffic in your analytics.
Can you guarantee we'll be mentioned?
No. The models decide. At Snippet Digital, we measure what moved and what didn't, without promising a placement.
Put the research to work
Tell us what your business needs to understand or change about its visibility. Andy and I read every enquiry. We'll discuss whether the Snippet Digital team is the right fit.
Discuss AI SEO with Snippet