Launching AI agents in the Fortune 500

On the latest episode of Talk to my agent, Max interviews Tara Roudi – Global Head of Agent Product Management at Decagon. They dig into what it takes to launch AI agents inside Fortune 500 companies, from setting the first goal to building the team that runs the agent long term.

Key insights include:

• Defining a North Star: Every AI rollout should start with one clear goal. A shared North Star keeps IT, CX, and business teams on the same roadmap, so they can sequence the work and focus on the 20% of use cases that drive 80% of the value instead of trying to boil the ocean.
• Governance and Change Management: For most enterprises, this is their first AI deployment at scale, and clearing internal security and governance reviews is often the biggest hurdle. Moving from rigid decision trees to AI agents also changes how teams test, so regression test sets and automated dashboards give legal and compliance the proof that sensitive scenarios end the right way.
• The Agent PM Role: Agent PMs act as guides through the transformation. They assess readiness, identify integration needs, and set best practices. They also teach customers to run the system themselves, taking on more of the work early and handing over control as the customer’s team gets up to speed.
• Building the Right Team: The strongest internal teams are small tiger teams that act as a center of excellence. They don’t need coding backgrounds. They need systems thinking, stakeholder management, and operational rigor, with Duet helping them handle the growing complexity without a large manual QA team.