This session is a teardown of an agentic marketing team that was built and deployed in production. Drawing on a real GitHub repository and a pre-flight review process, Daphne Monro walks the audience through the architecture behind a working set of artificial intelligence (AI) agents designed to handle specific marketing tasks, including content review, campaign workflows, and quality assurance (QA) for tracking.

The talk covers the points of failure encountered along the way, the guardrails required before trusting agent output, and the measurable outcomes, from time saved and error rates to output consistency, alongside an honest account of what did not work. Rather than a hype pitch, this is a build log: attendees will leave with a concrete, reusable framework for scoping and structuring their first agentic marketing workflow.