Last Thursday, 22nd of October, CodeNode at Liverpool Street was full of marketers who have moved past asking whether to try AI. The question most of them are working on now is a harder one, about getting agents to work inside real teams with the data and sign-off processes those teams actually have.
We asked speakers to talk about what they had built and what they had learned from it, including the parts that went wrong. That made for a more candid day than industry events usually are, and a more useful one if you’re trying to make progress in your own team.
Who was in the room
Practitioners from Gartner, Reckitt, Moody’s, HelloFresh, Miro, SAP, Pearson, Hiscox, Finastra, Eaton, ManyPets and Charles Tyrwhitt spoke during the day, along with several of the vendors whose platforms those teams use. You could hear a brand describe a problem in one session and then see how a platform approaches it in the next.
Most of the audience worked in marketing leadership, marketing ops or martech, and there was a good number of CRM and lifecycle people too. Having the people who set AI strategy in the same room as the people who make the systems work suited a programme built around what teams have actually put live.
The Data Layer
This track dealt with the part of agentic marketing that most teams find hardest to get right. An agent can only work with the context it can reach, and what gets in the way is usually fairly mundane. It might be messy CRM fields or permissions nobody had mapped, and quite often it’s two systems disagreeing about who a customer was.
Nina Bennett leads global marketing operations at Eaton, and her session, Marketing, Re-engineered for the Agentic Era: Process, Data and Governance, went straight at this. Governance came up again with Len van Hoogenhuijze of CRMT Digital, whose session Taming the Agentic Jungle proposed a control-plane approach to governing AI-powered campaign execution. When several agents are running campaigns at once, someone has to be able to see what each one is doing and step in when needed, and that depends on the data and permissions underneath being in order.
Dora Kondeva, RVP Marketing at Snowflake, focused on value. Her session, The AI Blueprint: Turning Agentic AI Ambition into Measurable Value, dealt with a problem many teams run into after a promising pilot, which is showing a return clearly enough for the project to be scaled up.
Agents at Work
This track was about what agents do once they’re live in campaign planning, execution and optimisation, which is the work that fills most of a marketing team’s week.
Kshitij Suhag, formerly Global Head of Omnichannel & On Demand at Reckitt, told the story of one agent in Killing the Promo Calendar: How We Built and Scaled a Signal-Based Revenue Growth Agent. As the title suggests, the idea is to move away from a fixed promotional schedule and let an agent respond to signals as they come in. Patrick Muntzinger, Product Tribe Lead at HelloFresh, presented with Hightouch on how HelloFresh is building AI-powered marketing beyond the traditional customer journey, and Andrei Badoiu brought the Hiscox perspective to Marketing Activation Re-imagined for an Agentic World.
The Agentic Team
Carlos Doughty, LXA CEO, on building the agentic marketing org.
If agents take on more of the execution, what will a marketing team look like in two years? And what should someone early in their career be learning now? Those were the questions behind this track.
Carlos Doughty’s session, Building the Agentic Marketing Org, treated humans and agents as one workforce. Once agents are part of the team, managers end up managing different things, and the way work passes from one person or agent to the next changes along with the skills worth investing in.
Fani Kelesidou from Moody’s offered a way to think about those skills in The Four Marketing Archetype Roles of the Agentic Era. Her roles of strategist, creator, orchestrator and optimiser give teams a framework for working out where their people add most value when agents handle more of the day-to-day work.
Omair Izhar from Miro and Luis Clark from Monday.com both looked at handing their own work over. Omair’s session was called I Built a MOPs Agent. Then I Gave It My Job, and Luis covered How to Give 30% of Your Job to AI. Harjeet Singh of Finastra asked whether people act as a safeguard or a bottleneck in Human in the Loop, or Human in the Way in an Agentic World, which is a question that gets harder to answer the more work agents take on.
Content at Agent Speed
Content was looked at from two directions. On the production side, agents can now create, test and personalise content at a volume no team could manage by hand. Jonathan Woolf from Intent HQ spoke on The Art and Science of Agentic Hyper-personalisation, and Nick Mason, CEO and co-founder of Turtl, discussed what agents mean for B2B marketing in Agentic AI & the Box of Light: The Paradigm Shift in B2B Marketing. Producing content at that scale puts real pressure on brand guidelines and approval processes, most of which were designed for a time when people wrote everything themselves.
The other direction was discovery. Buyers are increasingly asking AI assistants for recommendations instead of searching, and Alex De Fursac Gash, VP Analyst at Gartner, summed up the risk in his title, Invisible to the Agent, Invisible to the Customer: Brand Discovery in the Agentic Era. Pearson’s Mrunal Bhagat put a similar question to the room with When an AI Agent Recommends a Product, Will It Choose Your Brand?, and Arthur Mstoyan from Storyblok spoke on The Importance of Content Confidence in the Age of AI Visibility. Many teams are only just starting to think about how tools like ChatGPT, Claude and Gemini find and interpret their content, and what that means for the way they write it.
Thank you
Thank you to every speaker for talking openly about what’s working and what isn’t. That takes more nerve than presenting a polished success story, and it’s what made the day worth the trip. Thank you also to our sponsors, who made it possible: Funnel, Hightouch, Upp.ai, Hut 3, SAP, Similarweb, Storyblok, Optimizely, Intent HQ, Tealium, Snowflake, Turtl, Monday.com, CRMT Digital, Customer.io, Prismic and Treasure AI.
Finally, thanks to everyone who came…
If you missed it
The summit grew out of our Certification in AI Marketing, and the frameworks behind a lot of the day's conversations live there. If your team is working out how agents fit into your 2027 plan, it's a good place to start.
You can also read our Blueprint for Architecting Agentic Marketing at Enterprise Scale, or join us at the next AntiCon.