What I took home from two heated days on the seafront; where talks about Claude, fan‑out queries and hybrid SERPs gave me fuel for thought …
This time around, BrightonSEO was more action- and people-packed than ever. It started with long, winding queues to get into the Brighton Centre on both 30th April and 1st May. The lines stretched 100-people-deep, all along the seafront.
The draw of the event had to be down to AI, right? The AI question is fuelling interest in Search and making noise in all areas of eCommerce and Marketing. AI, Agents and Automated workflows are biting every sector and job title, from Executive to C-Suite, as we all clamour to leverage shiny new tools and approaches.
As the conference got underway, my instinct was confirmed. Across every talk – no matter the focus: Technical SEO, content SEO and Paid – mentions of AI ran rife. It was the common thread everyone picked at in the corridors between sessions, too.
With the changing shape of Search in mind, here’s what stood out to me from Brighton, as an agency SEO looking to stay on top of my game – and the techniques I’ll input going forward.
Key lessons & reframings from BrightonSEO 2026
- Tom Winter said during his talk ‘AI Systems Win by Doing What Humans Should Do, but at Impossible Scale’: “When bots get it wrong, we say they’re rubbish. When people get it wrong, we call it experience.”
That line landed because it reframed hallucination as a human problem too. We know the models mess up (even OpenAI admits this is a feature of LLMs, not a bug). So, the question is how teams can embed accountability into their AI workflows.
Tom made it clear AI systems win when they replicate human problem‑solving at scale. So, our task – as SEOs, as humans – is to design the right problems. This looks like: automating the repeatable (deployments, schema injection, canonical checks) and investing our precious time in building knowledge banks and designing prompts, implementing guardrails and performing robust checks.
- There was a clear winner of the LLM’s – and that was a bot named Claude. Claude – and specifically Claude Code – was noticeably cited above all other competitor models, praised for how it thinks deeply, writes code and can even deploy via SSH.
That detail mattered because it signalled a shift: we are picking sides and settling into the new ways of working. ‘Good in = good out’ also extends to the tools we choose.
- Dr Pete Meyers closed the conference with a clear, practical thesis: keyword research has stepped away from isolated long‑tail phrases and become more about managing networks of related ‘fan‑out’ queries. Linked to this is the shape of Search. It’s switching to a hybrid space – a mix of organic links, AI overviews and structured topic clusters.
A customer might start off asking a simple question – yet the path the Search function leads them down is multi-pronged. E.g. gain deeper understanding around one (or more) product attributes, make a detailed comparison, watch a ‘how to’ video tutorial… and SEOs need to be prepared.
A good starting point when looking to assess this in action is to move from single‑keyword tracking to prompt clustering and intent network analysis. For example, start with 2–4 seed words, expand to 10–12 word prompts, then measure how content performs across those clusters. These should be brand, ‘soft‑brand’ (i.e. where your brand still has relevance) and non‑brand prompts.
“I’m worried about the mad rush toward LLMs and the assumption that stand-alone chat is our likely future. I want SEOs to be curious about how all of this impacts organic search, especially as Google develops hybrid models like Web Guide.” – Dr Pete Meyers
What this means in actuality. Our next steps
BrightonSEO sparked debate and new ideas – so what will I be doing differently post-event? Below are the methods we’re testing and rolling out for clients.
- Pre and post launch checks become AI‑aware. We’re merging Nicolas Montañares’ refined technical checklist into our pre/post launch workflow and adding agent‑simulation tests to validate discoverability after launch.
- Prompt libraries replace static keyword lists. Each client will have a versioned prompt library mapped to user problems. Every prompt includes expected intent, canonical page and verification steps for hallucination risk.
- Hybrid KPIs and AI citation tracking. Dashboards will blend organic metrics with AI citation signals: AI overviews, Web Guide clusters and platform citations (YouTube, Reddit). We’ll measure brand lift and conversion alongside agent visibility.
- Safe AI deployments. Any AI‑driven code or content deployment must pass automated tests and a human review gate before going live. Claude Code speeds delivery; our guardrails keep it safe.
- Channel strategy for citation sources. We’re investing in platform content that feeds AI models: video transcripts, forum answers and structured data on third‑party platforms.
“I’ve been as guilty as anyone of riding the 20-year wave down the SEO funnel from traffic to conversions. I’ve come to believe that focusing exclusively on either the top *or* bottom of the funnel is flawed. Content and brands have to be present across the entire journey to really be effective.” – Dr Pete Meyers
Google is now anticipating and trying to cater for searchers’ entire journey. Natural language (including LLMs and GenAI) slots into the picture. But we should heed Meyers’ warning – agentic chat is just one part of where Search is going.
Final thoughts
Last years’ Brighton was most definitely the year AI adoption went mainstream for digital marketing. A huge chunk of the 2025 talks started their gambit with ‘let’s talk GEO’ or ‘AOI’ or ‘GEOAIEO’… almost a full rendition of Old McDonald had a Farm – EIEIO.
Whereas this April was all about the rise of automated workflows and agents. In my 15+ years in SEO, the last three of those have been defined by pivotal technology change, strategies to keep up and the famous line – is SEO dead? But with Claude tapping our shoulders, eager to help, I feel that the case for SEO in 2026 and beyond is more exciting and stronger than ever. For one thing, we can do more – measure more. And the SERPs remain an ever-changing playground for us to run riot, trialling and testing new processes.
Read more about AI Search
- Generated. Not Found: Building brands that matter in an Agentic world. Download the full report
- What are the eternal truths of brand discovery in AI?
- How to design a brand for machine-mediated discovery
- Organic visibility in the age of AI with StrategiQ COO James Bavington: Watch the webinar
A day well spent on the Brighton seafront, soaking up SEO learnings.