Prismatic logo
Prismatic Labs

Securing Integrations in the AI Era: 5 Lessons from 30 Years in Security

Every integration is a promise that data goes where it's supposed to and nowhere else. Keeping that promise takes isolated runtimes, logs that don't store customer data, secrets that stay secret, and authentication that works when other pieces fail. Add in AI coding agents that need credentials to do anything useful, and the list grows.

Live workshopDecember 3, 20262:00 PM ET

In this workshop

Buzz's 30 years in security span GRC consulting and live incident response, so he's seen both sides: how to prevent a breach and what to do when one does happen. He'll take us through the five things that matter most when connecting systems and Taylor Reece will join in with the developer's view of what those things look like in practice.

Then it's your turn. Bring Buzz and Taylor the questions you've been sitting on about securing integrations, AI agents, and the systems they connect.

Brian Munz, our Developer Advocate, will keep things moving. And we'll use Prismatic as an example where it helps, but you can apply what you'll learn to whatever you're building.

What you'll learn

  • Runtime isolation: Keeping one customer's data from ever touching another's, and what that scary vuln-scan finding means.
  • Logging in production: Why the logs that save you in dev become a liability in prod, and what to disable.
  • AI, privacy and secrets: What happens to your data when it passes through a model, what security reviewers are now asking about AI, and why a key pasted into a chat is a key you rotate.
  • Shared responsibility: What your cloud provider covers, what your platform covers, and what's still on you.
  • Authentication and layered security: Building auth so one failure doesn't open up everything behind it.

Speaker

Buzz Hillestad
Buzz Hillestad
Information Security Officer
Taylor Reece
Taylor Reece
Developer Experience Engineer
Brian Munz
Brian Munz
Developer Advocate