WPP
AI agents for advertising campaign teams
Automated workflows, an agentic chat assistant and the evaluation harnesses behind them, built for campaign managers on WPP Media's Open Intelligence.
Grounded AI agents and automated workflows in daily use, with output quality measured on every release.
WPP
WPP is one of the world's largest marketing and communications companies. Open Intelligence is WPP Media's AI and data product: the intelligence layer behind WPP Open, the company's agentic marketing platform, which uses predictive models and AI agents to help brands plan, buy and optimise campaigns.
Campaign managers were spending much of their day on repetitive planning, reporting and data-wrangling across many systems. WPP wanted AI agents and automation that could take that work on reliably: answers grounded in real data, quality that could be measured, and releases that didn't regress.
What we did
We built automated workflows and an agentic chat assistant for campaign managers, delivered as Python and FastAPI services on Google Cloud and integrating LangChain, Google Vertex AI, BigQuery and OpenAI models.
To keep the agents trustworthy we built knowledge graphs on Neo4j and PostgreSQL that ground responses in campaign and client data, and shipped evaluation harnesses that measure LLM and agent output quality and catch regressions before every release.
We led backlog prioritisation and roadmap planning across client and delivery teams, deciding what shipped and in what order, and set up GitLab CI/CD pipelines so releases went out safely and often.
Results
- Agentic chat and automated workflows in daily use by campaign managers
- Output quality measured on every release
- Roadmap and backlog led across client and delivery teams
Work
More work
Book a call
Have a workflow you'd like to automate?
Tell us where your team loses the most time. If AI isn't the right fit, we'll say so.