How to Build a Human Agent: Turn an Archive Into Expertise on Demand

Point AI at your own files, and it'll still hand you generic answers. It's never seen how the market is moving, and it doesn't know what the people on your team have already figured out, so it fills the gaps with whatever's on the web.

Piers Fawkes shows us how he builds what he calls a Human Agent. It takes someone's published work, chats, meeting transcripts, and shared docs, and turns all of it into an agent that researches and answers in their voice, citing its sources as it goes.

A person's expertise becomes something the whole team can reach on demand, and every answer traces back to real, credible knowledge.

In this session, he builds one live from start to finish. He probes a real archive, runs a short AI interview to capture what documents don't, takes the agent live, and then asks it something only that person would know so you can watch it answer in their voice. You leave ready to build your own.

You'll learn...

How to turn an archive into a working agent, step by step and on screen, so you can see how a body of work becomes something a whole team can query.

Where the outside knowledge comes from. An agent that's grounded only in internal files stays generic, so you'll see why your market and a person's expertise are crucial.

The five decisions this raises for any team, from which roles to start with to who owns the agent and how staff consent works.

This session's for marketers, strategists, and team leads who want the knowledge their team has built to be something everyone can access in seconds. Register now →