Builders,
Not Bystanders

Scan the start QR

The team captain — your Radio Boss — receives a printed QR code. They scan it to create and open the team in the app.

Set up together

With the whole team gathered around, the Radio Boss creates the team — name, motto, mission, roles, selfie. Do this together. It's the first challenge.

Radio Boss shares access

The Radio Boss's phone generates a unique QR code for each team member. Each person scans from the Radio Boss's screen to join the team in the app.

Game on

Everyone is in the app on their own device. Challenges unlock as you move. The Radio Boss is your captain — keep your team together and start planning your route.

💊 Choose your pill — Morpheus is waiting

The team decides together during setup — this sets your route

Live demo of the mobile app built for today — the same app teams use to submit logo hunts, GPS checkpoints, spark ideas and scores in real time.

Demonstration of the Data Spaces use case — connecting, querying and surfacing structured data across organisational boundaries. For those interested, this feeds directly into the build-along session.

An inspirational exploration — not a product pitch — of where AI agents are heading. Built around OpenClaw, the open-source agentic framework by Peter Steinberger.

What it does

A live demonstration of a structured data space — GPS-tagged, permissioned operational data flowing between teams and a central platform. Models how Droneport Rotterdam could manage real-time data across multiple parties.

How it works

Game data (GPS, submissions, team check-ins) streams into a permissioned data layer. Teams and agents query it in real time via the API. Demonstrates sovereign data exchange — each party controls what they share.

What it does

A live intelligence layer combining three data sources — client portfolio (who we work with), project allocations (who is on what, at what %), and game data (logo hunts, spark ideas, GPS) — into a single queryable view via GitHub Copilot (with the model of your choice).

How it works

Photos from the A–Z logo hunt are run through image recognition to extract company names, auto-building a prospect pipeline from Rotterdam streets. Spark Register ideas are matched against client gaps and allocation white space to surface real proposal opportunities.

What it does

GitHub Copilot scores every game submission using a custom prompt per challenge — evaluating correctness, quality and creativity — then aggregates results into a live leaderboard. Human judges can review and override any score. Copilot can run against any model you choose.

How it works

Eric's MCP server wraps the game API and exposes submissions to GitHub Copilot. A structured prompt per challenge type instructs Copilot on the scoring rubric — and you can point it at any model. Results are written back to the leaderboard in real time. The full judging module is being completed during the build-along.

What it does

Game data visualised live in Power BI on the Microsoft stack — combined with an AI agent layer that can query, narrate and surface insights from the dashboards. Kevin built this and will walk through how it was done.

How it works

Live game submissions flow into Power BI. An AI agent sits alongside the dashboard and can answer natural-language questions about what's happening — scores, trends, anomalies. Kevin will share the build approach during the session.

What it does

Everything generated today — team submissions, GPS, logo photos, spark ideas, scores — is available as structured data via a live REST API and an MCP server. Use it however you like: build a new agent, a dashboard, a tool, or connect it to your own system.

How it works

Steven's API exposes all game data as queryable endpoints. Eric's MCP server wraps that API so GitHub Copilot — running any model you like — can access it with natural language. Bring your laptop, pick a use case, and start building.