Organized by

Learn how AI agents are built for production, on your own use case
One day for people who already build AI agents. Bring a real process from your company. Your team builds it as a voice agent on the platform behind InPost's voice agent, tests every change with simulated conversations, and demos it live by 3 p.m.
After the prototype
What separates a good demo from an AI agent in production
- An AI agent changes often. Its functions and code live in Git, like any other code, so every change is versioned and can be reviewed.
- A change can break a conversation that worked before. Simulated conversations, run after each change, show what broke before customers see it.
- An AI agent calls live systems. Mocked APIs let you build and test it without touching them.
- PRO is one day on the Agent Development Life Cycle (ADLC): you build, test, and release AI agents the way engineering teams ship software.
Why come
What you take back to work
A way to build for production
How to build, test, and release changes to an AI agent: the Agent Development Life Cycle, ready to run in your team.
Your use case, working
A voice agent for a process from your company, built and demoed live in one day.
Tests you can reuse
Simulated conversations, regression tests, and mocked versions of your APIs that show what a change breaks before customers see it.
AI Agents Academy
Starting point
No experience building AI agents
You have built an AI agent before
The day
Build a first AI agent with a coding agent
Build your own use case as a voice agent and test it
Who comes
Teams new to AI agents
Academy graduates and experienced builders
You leave with
A working AI agent
A way to build for production, plus your use case as a tested voice agent
What you work on
Five things you work on during the day
Coding agents through MCP
You brief Claude Code, Codex CLI, or Gemini CLI, and it builds and configures your AI agent in Zowie through the MCP.
Functions and code in Git
The AI agent's functions and code sit in a Git repository, so changes are versioned and can be reviewed.
Simulations and regression tests
You run simulated conversations end to end. Your coding agent runs them, reads what failed, fixes the AI agent, and runs them again.
Mocked APIs
You mock the back-end APIs of your own use case, so you can build and test without the live systems.
Connecting to voice
You connect the AI agent to a test phone number and call it.
Who it's for
For people who build, run, or integrate AI agents
Academy graduates and experienced builders. Each team needs people who work in Git and with a coding agent. The other seats can come from design, telephony, architecture, and QA.
Come alone or bring colleagues. You can register with up to two colleagues, though every registration is reviewed individually. Colleagues from one company work on their own use case. If you come alone, you join a team and pick one use case together. We never put competitors in the same team.
Conversations
Voice and chat
Conversational AI designers and product owners of digital channels. Deciding what the AI agent says and does.
AI agent logic and models
Prompts, functions, model versions
AI, ML, and software engineers. Writing the AI agent's logic and functions in code.
Voice and contact center
Telephony, queues, routing
Contact-center and telephony owners. Connecting AI agents to the phone lines customers call.
Architecture and integrations
APIs, back-end systems, security
Solution architects and integration engineers. Wiring AI agents into back-end systems.
Testing and release
Simulations, regression tests, releases
QA, test automation, and DevOps engineers. Proving a change is safe to release.
How the day runs
One day, from your use case to a live demo
Welcome and ADLC keynote
Registration and coffee from 9:00 a.m. At 9:30 a.m. the Zowie team opens with a keynote on what the Agent Development Life Cycle is and how to run it. From 10:15 to 11:00 a.m., a guided demo from the first MCP prompt to a voice call.
First build block
11:00 a.m. to 12:30 p.m. Your team builds its use case with its own coding agent. The block is open: you decide what to work on.
Second build block
1:00 to 2:00 p.m., after lunch at 12:30 p.m. The goal is a working, tested voice agent by demo time.
Live demos
2:00 to 3:00 p.m. Teams demo their voice agents live. One team wins.
FAQ
Questions about PRO
Bring a use case. Learn how production teams build it
AI Agents Academy PRO, Google Office, The Warsaw HUB, Tuesday, December 15, 2026.