← back
Building Agents (the hard parts!) - Rita Kozlov, Cloudflare
Takeaway
Build agents as four loosely-coupled components (client/AI/workflows/tools) and use durable, stateful compute primitives like Cloudflare Durable Objects to back them.
Summary
- Rita Kozlov (VP Product, Cloudflare Workers/Durable Objects) frames agents around four components: client (interface), AI (reasoning), workflows (execute actions), and tools (capabilities).
- Citations: 76% of developers use AI in dev process (was 44% a year ago); knowledge-worker AI usage already past Gartner's 2030 50% prediction.
- Business impact: 20% revenue lift on sales automation, 90% faster support response, 50-75% time savings broadly when using agents.
- Pitches Cloudflare's Durable Objects as the substrate for stateful per-user agents at the edge — ~20% of internet traffic flows through Cloudflare.
cloudflaredurable-objectsagents
Original description
AI workloads are rapidly shifting from AI being used for augmentation (co-pilots), to AI becoming responsible for full, end-to-end automation (agents). But building effective agents, and even more importantly, agent experiences that boost productivity requires many pieces. In this talk, we'll be covering the building blocks of agents, how to put them together, and what we've learned from top companies building agents along the way. About Rita Kozlov From the very beginning, Rita has been a key figure in the development of Cloudflare's developer platform. Their initial experience as a solutions engineer, helping early enterprise customers adopt the service, gave them firsthand insight into user needs. It was the power of Cloudflare Workers that truly resonated, inspiring a vision for a serverless future. For the past eight years, they have built out the platform which now spans products including storage, compute and AI, and is used by everyone from indie millions of indie developers to Fortune 500 companies. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter