Six shifts from brightonSEO that will shape search in 2027

Ross Simmonds closed the conference with a keynote on building your brand into the memory of LLMs. His argument was that the bigger question isn’t what you tell an LLM about your brand. It’s what your audience tells it.

It was a thread that ran throughout the whole conference, with numerous talks featuring this theme. Jonathan Kiekbusch argued that being described as the cheap alternative can be worse than not appearing in an AI answer at all, while Tuhin Banik looked at how AI systems connect evidence to decide which businesses to trust.

Tom Wells tested how the sentiment and position of a brand being mentioned in an LLM response changes its effect. Laura Smith described a gap between human and machine memory: some brands are well known to people but invisible to AI, while others show up in AI answers but aren’t remembered by customers.

This matches what we’re seeing, and it’s a big part of why we’ve encouraged clients to invest more in off-page work this year. Digital PR and off-page marketing had gone under the radar for a while and probably hadn’t been prioritised enough by most businesses. AI platforms build their answers from what they read about you across the web, so your off-page presence needs to be better than ever. Our Head of Digital PR, Jack Hayes, made a similar point in June: the sentiment attached to your brand in published media shapes how AI perceives you.

Several speakers warned that AI visibility dashboards can mislead. Andrew Chadwick presented research showing that up to 75% of AI brand recommendations change when the person asking changes, which means generic tracking prompts can be inadequate. Tom Rudnai showed how changing a conversation early on can change the shortlist and recommendation that comes later. There’s no one-size-fits-all prompt either. Josh Blyskal gave some fascinating insights from Profound’s data, demonstrating how changeable AI search is – for example with ChatGPT fan out queries having doubled in August.

This level of regular change to the methodology of AI platforms means you could be creating strategies based on already-outdated results. That’s the thinking behind how we built our bespoke AI Visibility Reporting tool – measuring around topics and common themes, rather than a single headline number, broken down by LLM and against named competitors, down to the URLs cited for each prompt.

Reporting and budget came up again and again. Joseph Bauermeister made the case for bringing search insights into the boardroom, while Nikki Goodman looked at why SEO recommendations don’t get implemented even when they’re right. Harry Gadsby shared what his team learned from tracking hundreds of paid-media reports after they were sent: how many people open them, and what they actually look for.

The one-size-fits-all problem applies to reporting too: a report built for everyone ends up serving no one. Our Data & Analytics Manager, Katie New, covered this on the brightonSEO stage back in April with Nobody reads your reports (and this is how you fix it). Her argument was to start with the questions the reader needs answered and choose metrics last. Her framework is now our free Aligned Reporting Toolkit.

There’s a harder truth underneath this. Organic traffic will be lower than it was before AI search – that’s inevitable with so much information being available without needing to click through to a website. So we encourage clients to report on overall business performance, such as conversions and revenue, sentiment and visibility, rather than judging each channel by its clicks alone.

If you’re making that case to your own leadership team, our guide to explaining SEO/GEO to stakeholders covers the questions you’re most likely to be asked, from when you’ll see results to which KPIs to set. 

In a packed Organic Social session in the main auditorium, Ashley Liddell followed a buying journey through TikTok, Reddit, Instagram and YouTube before the customer ever searched for the brand. Susan Adeoye showed how TikTok’s own search data can shape a content strategy, while Carmen Dominguez explained how Reddit can support both SEO and reputation.

The industry has stopped treating Google as the most important part of search engine optimisation. “Search Everywhere” is the phrase we’ve used for this all year: treating traditional SEO, GEO and social search as connected disciplines rather than separate channels.

In January, our CEO, Tom Vaughton predicted that 2026 would be the year search marketers start meeting users where they actually are: in conversations, in visual discovery and in video. That’s why Social Search now sits alongside our technical SEO, GEO and digital PR services.

Agentic commerce has moved from speculation to practice. Malte Landwehr shared what tracking more than a million products across 250 million ChatGPT conversations reveals about which products get recommended, while Stijn Bergmans showed how agents fan a single query out into dozens of sub-queries

What connects them is simple: agents can only recommend what they can read and trust. When we tested 100 of the UK’s top ecommerce retailers in June, 88% of those we could test weren’t ready for the Universal Commerce Protocol. It’s why we built our Agentic Shopping Readiness Audit, which tests how well a website works for an AI agent trying to buy from it. The brands that earn a place in agentic shopping will be the ones that do that groundwork early. 

As more of the work is handed to AI, several speakers asked what that leaves for people to do. Michael Holloway argued that juniors learn more from investigating technical issues themselves than from running an automated crawl, because the tool’s first flag isn’t always the real problem. Tom Winter pointed out that AI can now carry out a growing share of SEO work, but deciding whether the result is good enough to trust is still a human call.

Varn’s Tom Vaughton addressed exactly that on the same stage a year ago, in his talk The SEO identity crisis. AI can carry out the work, but it can’t decide what matters most, and over-relying on it leads to sameness. It’s good to see that conversation growing. This is a good time for curious SEOs, because the people who want to know why something works will get the most out of these tools.

Which brings me back to where I started. Good marketing is still good marketing – and good SEOs (or GEOs…) are still good SEOs. Knowing your audience, judging what good looks like and earning your place are human skills, and AI makes them more valuable, not less.