Most of what shapes AI next year is sitting in a research paper today. Keeping up with them is a full-time job, and the abstracts don't help.

TLDR AI is the newsletter that does the translating. Each morning you get the handful of papers worth knowing about, curated by Anthropic and ex-Google engineers, with a plain-English summary of what they found and why it matters.

You come away smarter about AI research in the time it takes to finish your coffee.

Free, and delivered to 1.1M+ inboxes every morning.

Subscribe for free.

This week I'm writing from the other side of the expo floor. Obsero had its own stand at brightonSEO for the first time, and we spent Thursday and Friday talking AI search with the people who have lived and breathed organic growth for years.

brightonSEO is the biggest SEO conference going, which makes it the best place to meet the people fielding the most questions from the C-suite right now. SEOs have been at the heart of organic growth for the last 25 years, and they are the ones being asked whether their brand is the recommended answer in AI search.

Brands and agencies shouldn't put all of that responsibility on SEOs, but SEOs are the ones facing the biggest change in measurement and resourcing, so they were exactly who we wanted to learn from and talk to.

We went with two goals. The first was leads, and we got them. The second was to understand people's concerns about the GEO and AEO space, so we can answer them or help change them.

It's a new space, and in a new space you get people who want to be first at everything, which can come at the expense of quality. SEO was never easy either, with constant pressure and a lot of juggling of resource and stakeholders, but we had a good 20 years of knowing what works, what doesn't and what it takes.

Even core updates, as stressful as they were, brought big gains to the brands doing the work. I felt that personally on 3 December 2020 with Cazoo, when the third core update of the year sent us into the stratosphere in the depths of COVID. It's still the best feeling I've had in 20 years of SEO. Celebrating on my own, in my log cabin!

Honestly, I loved it. I enjoy talking about AI search and meeting like-minded people, although you have to stay conscious that you're selling something and some people just want to get on with their day and learn.

One thing stood out. As soon as I mentioned I'd spent 20 years in SEO, people were far more interested in what I had to say, because they were talking to someone who has faced the same problems they have and the conversation stopped feeling like a sales pitch.

I was also mindful that plenty of people go to Brighton to listen to the talks and meet other SEOs dealing with the same challenges, so I tried to respect that.

This was the first time we'd ever done a stand, which made it exciting. We picked a spot opposite the bar and just outside the auditorium so we'd catch the footfall as talks finished, and honestly I didn't know what to expect, how busy it would be or how people would respond to what we offer.

You're also wary of competitors coming over to ask questions. The VC-backed companies were downstairs with huge, glossy stands, front and centre as you walked in, although I didn't really venture down there. I went to see my old colleague Louise Lenihan at Ahrefs instead, who is always worth catching up with. She has been in the industry a long time, she's extremely insightful and she works at one of the best traditional SEO tool suites there is (I'm a customer too).

One of the best moments of the two days was my old mentor and friend Jon Earnshaw coming over to the stand. Jon was my boss from 2006 to 2013 and has been in the game for more than 25 years, and although we don't catch up anywhere near enough, it was fantastic to see him.

Our backboard carried a line we came up with a while back, "Your unfair advantage in AI search", and I think it works. It's an experienced team building a product that measures how effective your marketing activities and your product are, and then uses that data to recommend what the brand should do next to keep growing.

My football game was far too complicated and didn't really work in an enclosed space, so last week we pivoted to giving away a pair of Meta glasses each day, which went down very well. The lesson is that you can't assume people know there's a prize just because it's on the stand, and some people are a bit nervous about coming up to ask.

Before the next event we'll do more to promote the stand and what we're offering, and we'll probably have leaflets to hand out. The stand next to us, a smart niche product covering AI search and PR, did this really well. They ran a hook-a-duck game where a mark under each duck told you whether you'd won a bottle of prosecco, and they had a nice leaflet tied into Celebrity Traitors, which is very popular in the UK right now.

We were happy with where we were for the budget we had. Once the team grows we'll look at a bigger stand downstairs, but for now it's about doing more of what our neighbours did so well.

A talk is something we'd consider as well, although I'd want it to be more personal than another "What 200,000 query fan-outs taught me about X". Those talks have been done, so it would be more about my experience and what I've learned in 20 years. We'll see.

Thursday was definitely busier than Friday, so based on what I saw this year that's where the opportunity is. Friday is still worth doing, but Thursday is the day to meet people.

brightonSEO has been going for 15-odd years, which makes it hard for any talk to offer something new. Anecdotally, people I spoke to before going said a lot of talks are trying to sell you something, and many of them cover similar ground. Even so, I saw some good ones, including speakers who were clearly nervous about standing up in front of hundreds of people and still did really well. That's an achievement and something I've never done, so hats off.

The talk that stuck with me most was Tom Wells on base model bias. He looked at a software company and which attributes matter most when a model recommends a brand before any web retrieval happens. A customer never sees that layer, yet it shapes the final answer they get, which makes it central to what we're building. How training data shapes answers, and how to bring that into Obsero, is something I've been thinking about for a long time.

He ran a similar study for Lightspeed on where AI sends you when you ask where to shop (spoiler: the big guys), which is worth a read. He's ex-Peec and we chatted a bit at the stand. I've always found him really interesting, so I'm going to reach out to him, because it's an area I'd love to explore.

The other one I found really interesting was Frank van Dijk on surfacing the questions people are actually asking, on Reddit for example, so you can understand what your audience wants. It complements tools like Obsero, and he recommended Apify for doing your own scraping, which I use regularly for my own work. I reached out to tell him he'd done a great job.

The question we got asked most was simple: why are you different from the competition? From the conversations on the stand and what I've seen on LinkedIn, there can be a bit of negativity towards the category right now. Some people are firmly against prompt tracking, I assume because the answers are probabilistic and, unless you track regularly, there's a lot of noise.

If you bring the old SEO view of rankings, where everyone sees the same result, then prompt tracking is going to disappoint you, because that was never its purpose.

I think we answered the question well. We're tiny next to the other platforms on the floor, but Mike and I have 30 years of combined experience between us. We've seen the rise and fall of voice search, the move from desktop to mobile and Google updates like Panda and Penguin, so looking at GEO and building software to help brands win in this era felt like a natural next step.

The key is being clear with prospects and customers about which models you track. We use the free versions of ChatGPT, Gemini and Claude, and for brands whose ideal customers pay for a subscription we offer strategic guidance on logged-in experiences through our consultancy. That feels fair to me, because plenty of tools sell "what ChatGPT is saying about your brand", which holds true if you're looking at the free version.

What I'm building as CPO at Obsero goes well beyond simple prompt tracking as well, into areas like site crawls and content scoring, and now that I'm fully focused on Obsero there's a lot more to come.

What people said about rival tools

People using other trackers talked a lot about a lack of support. What surprised me more was how often trust in the data came up, and this wasn't the usual probabilistic versus deterministic debate either. A couple of people told us the data in their tool's interface didn't match what they exported for reporting, which I found extremely interesting.

Is GEO just SEO rebranded?

Some people don't like change and want things to stay the same. You have to be considerate of that, while making sure they know the world has moved on.

The example I always use is a huge betting firm I worked with back in 2009. The owner, an old-school businessman, thought the internet was a fad and that he should focus on the high street, because that's where the punters were. His digital team were stressing, because doing what he'd always done felt simpler to him even though expectations had changed, and in his case it was shareholders' expectations.

If you're an SEO who doesn't like change and wants things to stay as they are, I understand. You'll still need to answer to a CMO whose CEO is asking how the brand shows up in ChatGPT.

Cashing in on a trend

People are sceptical of anyone who looks like they're cashing in on an emerging trend, which is totally fair. What matters is what those people are saying and doing, their pedigree and experience, and why they're doing it.

I've been in SEO for over 20 years, as I liked to remind everyone on the day, and I saw a clear shift. I'm actually not very vocal, because I'd much prefer to build product, talk to customers and focus on growing Obsero.

There are some people you'll never win over, and that's fine. We'll always make sure we cater for SEO-focused teams, although we won't be right for everyone. What is clear is why we're doing this: we strongly believe the future of search is AI, and that has nothing to do with a payday or a bandwagon.

Offering both a platform and a consultancy makes us an ideal partner for brands that want to win in AI search. Agency partners are fantastic too, because they bring on more brands and we get to learn what's working across lots of different sectors.

Scroll through LinkedIn after the event and the same two themes keep coming up: scepticism about AI visibility tools and a lot of talk about measurement. I think there are a few reasons for that.

Going up against the big money

If you're up against companies that went first and have an astronomical amount of money behind them, you need to look at how you can better support the person in front of you. The behemoths are under pressure to grow, win new business and acquire customers as quickly as possible to hit their targets, which makes it much harder for them to look after customers as well as a small business can.

One of them is clearly heading for an IPO, and that brings a whole world of pressure with it. I've also seen first-hand how growth ambitions can distract a company enough that it drops the ball with existing clients.

Why people are sceptical

Transparency of data is a big part of it. There was an almost irrational willingness to believe rank trackers, even though the traditional SEO tools I worked with had clear inaccuracies. The difference was that Google Search Console sat alongside them as an official source. If you grouped your rankings into a topic or category, or whatever your rank tracker called them, you could see the same increase in Search Console and say, "Right, my efforts are paying off here."

The customer journey in AI search is different. Someone could spend hours in a conversation with ChatGPT, make their decision before leaving it, and then go direct to the website or type the brand into Google, so you never see the influence ChatGPT had on that purchase. That's a huge reason why this feels different and why people are anxious about the change.

The platforms themselves (Anthropic, Google or OpenAI) offer very little on the AI experience either. In my opinion the AI search reports in Google Search Console and Bing Webmaster Tools just aren't good enough, because they don't tell you whether anything is working.

Optimistic or anxious?

Is the industry more optimistic or more anxious than a year ago? Honestly, both. Everything at Brighton was AI, and whether you call it GEO or AI search, everyone has realised it's all AI now. There was still a lot of anxiety among the people we spoke to, and some are still panicking about the change.

Thanks for following the Obsero journey. Last week's poll told me you want to see more AI search insights and data, with half of you picking it, and I hear you. There's clearly interest in what we're building, and I know you want more of the data we see so it can help you in your roles. That will come.

As anyone who has started a business will know, you end up doing payroll, PR, marketing and product all at once. Research and studies are high on our agenda, so they will follow in time.

What I hope you take from this edition is that we were on the front foot at Brighton, talking to people about the space and what's working. There was a lot of interest in agent analytics, and when I demoed the platform people were especially keen on sentiment analysis, because they want to know whether an answer was positive or negative and which sources are shaping it. We're rolling that out this week, so watch out for it.

In April we'll do more to drive engagement on the stand, with prizes, games and packs to give away to anyone who comes to talk to us. We'll probably do another lunch as well, since Mike met some very interesting people at this one and came away with some good leads. Maybe a talk too, but we'll see.

The team on the stand

Mike and I were joined by:

We're hiring

We're going to be hiring across engineering and marketing, and the role at the top of my list right now is a content marketer who:

Send any recommendations my way, and if we hire them you'll get three free months of Obsero for your business.

That is it for this Sunday, until next time.

Have a good one.

Andy