Making AI Agents with Julian Harris

The promise of AI agents is to accelerate your business with fewer people, cheaper and faster. But will they really? I help you make better decisions by sharing my direct experience with agents and agent-making tools so you don't have to.

Launched 2 years ago

Stop Tuning the Model. Start Tuning the Harness.

“Harness” a new buzzword. I explain why you’re already using one and why it’s important to understand the other options available now.

May 18 • Julian Harris

JP Morgan Wired the Orchestrator. Sakana Trained One.

Two flavours of multi-agent orchestration shipped in the same news cycle. Only one of them is what you're already building.

May 10 • Julian Harris

Your AI's Biggest Problem Is Permission

Enterprise AI's scarce resource just flipped, and the platform vendors saw it before you did

Apr 12 • Julian Harris

Your AI Agent's Failures Might Be More Valuable Than Its Successes

IKEA studied what its chatbot couldn't do, and turned into a whole new business line.

Apr 5 • Julian Harris

Your Agent Needs a Bank Account

162 projects are racing to build financial rails for software that doesn't have a credit score yet

Mar 29 • Julian Harris

Your Software's Next Users Won't Have Eyes

The plumbing for an agent-native internet is being laid right now, one error page at a time

Mar 15 • Julian Harris

We're Rebuilding the Entire Tech Stack for Users That Aren't Human

The quiet revolution isn't smarter models. It's building the entire digital world again, but for non-humans.

Mar 8 • Julian Harris

Multi-agent orchestration: When your single agent simply isn't enough

Edit: Ooops this was entirely authored by AI.