If you caught Dr Pete Meyers’ keynote at Brighton SEO this year, you’ll know it was one of those sessions that makes you want to go home and rethink your entire content strategy. If you missed it, we’ve put together the highlights so you can get up to speed.
Search has shifted more in the last three years than in the previous twenty. The way we think about keywords hasn’t kept up.
A Quick Look at How We Got Here
Pete opened with a walk through Google’s history, and it told a clear story. From plain blue links in 2000, to knowledge panels, featured snippets and People Also Ask boxes, every update has pushed search closer to understanding what you actually mean rather than what you literally typed.
AI hasn’t changed the direction, but it has just floored the accelerator.
By April 2026, Google’s AI Mode handles queries like “Why can’t I book a direct flight from Chicago to Gatwick without a layover in Madrid?” and returns a detailed, sourced answer. That’s a long way from typing “cheap London flights” and hoping for the best.
The Long Tail Just Got Infinite
Most SEOs know the keyword curve well enough: fat head terms with big volume and big competition at the top, chunky middle terms in the middle, and specific long tail phrases trailing off at the bottom. That tail always had an end, as there were only so many queries being typed into a search box.
AI removes that limit. When someone uses an AI assistant, they don’t just type keywords. Instead, they have a conversation, including full sentences and follow-ups. They ask things they’d never have bothered searching before because a search engine couldn’t handle the nuance. The query space is now bottomless, and the tail just keeps going.
That makes the old approach of targeting a fixed list of phrases increasingly difficult when your audience is generating new query variations constantly, many with no search volume data at all.
Search is Hybrid Now
It’s not all AI and it’s not all traditional organic. It’s both, running side by side.
Google’s Web Guide feature shows this well. Search for “how to properly brew tea” and you get traditional organic results, a Web Guide AI summary, and topic clusters grouping results into subtopics like “Comprehensive Brewing Guides,” “Beginner’s How-To Guides,” and “Avoiding Common Brewing Mistakes.”
Those clusters matter as they tell you what Google’s AI layer considers the important angles of a subject, based on meaning and intent rather than keyword matching.
Ranking for a phrase is still valuable, but there’s now a second game running alongside it: being the kind of source an AI reaches for when assembling a response. That requires a different approach, and a solid SEO strategy that accounts for both.
Query Fan-Out: The Concept That Changes How You Do Research
Query fan-out describes how a single topic branches into multiple related queries across different intent types. Pete used “how to brew a cup of tea” as his example and showed how it spreads in ten directions:
- Semantic – the same question in different words. “What’s the best way to make a hot cup of tea?” means the same thing but is a completely different phrase.
- Entity – the same topic with a specific brand or type attached. “How do I make a traditional cup of Earl Grey?” or “Are Adagio Teas easy to brew?”
- Follow-up – what someone naturally asks once they have the basics. “Should I add milk before or after the tea brews?”
- Anticipate – what they’ll need next. “Should I get an electric kettle with temperature settings?”
- Attribute – a very specific detail. “Does the shape of the tea leaf affect brewing time?” Never makes a keyword list, but people are asking it.
- Factual – a number or a date. “How long has tea been drunk in China?”
- Tutorial – step-by-step. “How to make a perfect cup of Matcha.”
- Insight – opinion or context. “Is it wrong to add milk to green tea?”
- Compare – weighing up options. “Electric kettle vs stovetop for tea?”
- Transact – ready to buy. “Best budget electric kettles for tea.”
Quite a spread from one seed topic. A brand in that space shouldn’t just be optimising for “brewing tea.” They should be building content across all of these angles, because AI systems pull from this entire landscape when they put a response together. You want to be in the pool.
Thinking in Layers
Pete’s practical framework for research works across three tiers:
2 to 4 words – your seed term. Short, broad.
6 to 8 words – mid-length queries covering different fan-out types.
10 to 12 words – long, specific queries targeting very particular intent.
Most keyword research has focused on that middle layer and stopped there. Mapping the full picture and building content that covers it properly is where the advantage now sits. Getting the technical foundations right matters here too. Content that can’t be properly crawled and indexed won’t make it into any response, AI or otherwise.
How Your Brand Shows Up in AI Answers
Pete tested brand visibility across 300 Gemini responses using three prompt types. The results were interesting:
Brand prompts (explicitly naming the brand): 100% presence. Expected.
Soft-brand prompts (describing the category without naming the brand): 78% presence.
Non-brand prompts (generic topical questions with no brand signals): 53% presence.
That 53% is the one worth thinking about. Even on a completely generic question in your space, there’s a reasonable chance an AI mentions your brand, but only if you’ve built real authority and your content genuinely covers those kinds of questions. It doesn’t happen by accident.
Your content strategy needs to reach beyond product pages. Covering the wider conversation in your industry is what earns you a place in AI responses on non-brand queries. Being broadly useful is now a visibility strategy.
What to Actually Take Away
Pete’s talk was indepth, but here are the key takeaways you can implement:
- Treat keyword research as topic mapping, not list-building.
- Think about the full fan-out of questions around your subject areas and build content that covers them.
- Work across all three query length layers, not just the middle.
- Plan for brand, soft-brand and non-brand coverage as three separate content opportunities.
- Check what AI systems are already saying about your brand and industry. Moz’s Prompt Suggestions feature, shown in the talk, is one way to do that.
The long tail isn’t dead – instead, it’s gone infinite. For anyone willing to do the work, that’s very good news.
If you want to talk through how this applies to your brand (whether that’s SEO, email marketing or just figuring out where to start) get in touch with the team.
Based on Dr. Pete Meyers’ keynote “The Infinite Tail: Keyword Research for AI” at Brighton SEO, April 2026. Pete is at peter@moz.com and @drpete on LinkedIn.
