What We Learned at Our 2nd Annual AI Summit

Last year we brought clients and friends together to take stock of where AI was heading. On 11 June, we did it again, this time with a lot more testing, building, auditing and future-gazing behind us.

This year’s Summit looked at how consumer AI use is changing, what we’ve been building to speed up the boring jobs and sharpen the strategic ones, what we now know about AI Search, and where agentic search goes next. We also pitted Claude Code and Codex against each other over lunch to see who could build the best tap-to-flap game. Claude won.

State of AI: Where Are We Really?

Our CEO Sam opened the day with a look at how the chat window has become the new front door for how people find, ask and decide. It’s reshaping what “search” even means, and where brands need to show up.

Propellernet Labs

Sophie (Head of AI & Strategy), Jim (Creative AI Technologist) and Dan (AI Search Director) took us inside Propellernet Labs, our home for the AI tools we’ve built for our team and clients. They demoed a handful of what’s live, from tools that take the grind out of everyday jobs like coverage tracking, through to more involved builds like our LLM audit tool for SEO teams. All built in-house. Some do things that simply weren’t possible before, others just make the repetitive stuff faster.

Dan also gave everyone a first look at Ebb, one of our newer builds. It tracks how brands, and their competitors, show up in AI search results across the likes of ChatGPT, Gemini, Claude and Perplexity. So rather than guessing, you can see whether AI is recommending you, ignoring you, or quietly pointing people to a rival instead. Think of it as a brand tracker, but for AI answers. It means we can show clients exactly where they’re winning ground, and where there’s work to do.

AI & Search: The GEO Session

Sophie and Dan shared what a year of testing, auditing and learning across client accounts has taught us. Sophie walked through her ten core beliefs about AI search, and Dan brought findings from a structured programme of tests: how LLMs handle schema markup and JavaScript content, and how fan-out queries vary across different models and users. There was even a pig farm involved.

Digital PR & Brand Building in the Age of AI

Sophie and Steve (Head of Digital PR) explored how Digital PR has grown from a link-building tactic into a real influence layer across both traditional and generative search. They looked at how PR now has to work across three timelines at once: immediate search impact, mid-term brand visibility, and the longer-term influence on what AI models come to know about a brand. Plus what the agentic era means for Digital PR specifically.

The Open Claw Diaries

Jim shared the agentic site auditing prototype he’s been building. He demoed his OpenClaw multi-agent system, showing the audience what his little robot gang can do, and how it’s inspired the simpler agents we’ve built internally to help the team do better work.

Agentic AI: The Future

Ben, Founder of Rise AI, closed the day with a sector-by-sector look at what fully autonomous AI agents mean for e-commerce, insurance and travel. What the signals are telling us, and what brands should be thinking about now to be ready.

Where We Go From Here

Two Summits in, and the conversation has moved from “where is AI heading” to “what are we actually doing about it.” We’d love to keep that going with clients and friends who want to test, build and learn alongside us. If that’s you, get in touch, or explore more of our insights.