Applied AI is live. Every week, The Delta's team shows you what we're actually building with AI - no slides, no theory, just a screen share and honest commentary.
Episode 1: Building Agent Teams
Running a single AI agent on a complex task is like asking one person to do a ten-person job. It gets slow, it makes mistakes, and it doesn't scale.
At The Delta, with our work building agent teams for ventures in our ecosystem, we have explored various ways of building out agent teams.
A group of specialised Claude agents working in parallel, each handling a defined part of the task, coordinated by an orchestrator that brings the outputs together.
In this first episode of Applied AI, Wynand Viljoen and Louis Buys will show you some examples of how we are building agent teams for ourselves and our ventures. You'll see the architecture and hear our learnings of what works and what doesn't.
This is not a presentation about AI. It's a demo of AI in production, inside a real company.
Founders and operators who are past the "should we use AI?" question and working on the "how, exactly?" one.
Format: 45 minutes. Live demo + Q&A. Recorded and published to YouTube.
Wynand Viljoen - AI Strategist at The Delta
Wynand leads applied AI development at The Delta, building the internal tools, workflows, and agent systems that run the company's operations and ventures. He hosts The Delta Applied webinar every week, showing what the team actually built and not what they planned to build.
Louis Buys - CEO and Co-founder of The Delta
Louis is the CEO and Co-founder of The Delta. He is a serial entrepreneur and has built 20+ companies across multiple industries. He joins for Episode 1 to explain the multi-agent system he has built on {aperclip, to support his role as CEO.
The Delta - The Home for Entrepreneurs
The Delta is a venture ecosystem, campus, and AI-native operator based in Berlin. We build companies, run AI systems, and work with founders to bring real applied AI into their businesses. Applied AI is our weekly live demo of what that actually looks like.