◈ RASA / HACKATHON
--:--:-- UTC
01 / 09
Boarding Now · 5-Hour Build Sprint

Build an AI harness that turns a strategic prompt into a fully-featured Rasa bot — then let it help Lisa pick where the whole company flies next.

Ten teams. Six builders each. One question: can your harness build a complex, multi-domain bot that a real person would actually trust?

11:30
Kickoff — clock starts
7:00
Final cutoff, bots live
50–100
Hidden test conversations
1
Metric: goal adherence
GATE A1

The mission

A harness is a reusable, AI-orchestrated system. Feed it a business use case; it ships a production-ready Rasa bot.

01 · INPUT

A strategic prompt

One well-crafted prompt describing the use case, the flows, and the goals. No hand-coding the bot.

02 · HARNESS

Your orchestration

Agents, an LLM prompting layer, integration patterns, and human checkpoints working together as an architecture.

03 · OUTPUT

A working Rasa bot

CALM flows, dialog logic, error handling, clarification loops, and escalation — live over ngrok or a deployed endpoint.

DELIVER

Harness codebase

Documented code + README explaining architecture and design decisions, with example skill definitions.

DELIVER

Architecture diagram

Use Case → Harness → Bot, annotated with agents, rules, decision points, and human checkpoints.

DELIVER

Human-in-the-loop doc

3–5 real checkpoints where a human validates or adjusts harness output — not just checkboxes.

DEPARTURES

Today's board

The day runs like a flight schedule. Tap any row for the gate notes.

Time
Phase
Status
Info
PASSENGER

Meet Lisa

Director of HR. She needs to decide where Rasa holds its next company offsite — and she wants an agent to do the heavy lifting.

Rasa · Passenger Record
🧭

Lisa

DIRECTOR OF HUMAN RESOURCES
Group size
~100 people
Spread across
Europe · US · India (3 continents)
Cares about
Visas · flight time · budget · weather · activities · safety
Budget
Tight — and she'll reveal it when she's ready
P<RASAOFFSITE<<LISA<<<HR<DIRECTOR<<
100PAX<3CONTINENTS<VISA<WEATHER<BUDGET<42

Entry stamps · what she'll ask

VISA EDGE CASE BUDGET WHAT-IF WEATHER COMPARE FRESH CITY

Click a question to load it into the bot demo below ↓

REAL DATA ONLY

Flights & hotels

Amadeus, Kiwi Tequila, SerpApi, Booking.com, Google Places — pull live prices and availability.

REAL DATA ONLY

Visas & safety

VisaGuide.World, IATA Travel Centre, government advisories — requirements by nationality.

REAL DATA ONLY

Weather & activities

Open-Meteo, WeatherAPI, TripAdvisor, Google Maps — climate by month and group-friendly things to do.

GATE H

Anatomy of a harness

Tap a stage to inspect it. The middle is where your team's point of view lives.

Input

Use case prompt

The offsite brief + goals fed in as a strategic prompt.

The harness

Orchestration

Your architecture that plans, generates, and validates.

Agent / skill definition system
LLM prompting layer
Integration patterns (APIs / MCP)
Human-in-the-loop checkpoints
Output

Rasa CALM bot

Flows, actions, escalation — live endpoint.

LIVE

Talk to the bot

A taste of the conversation your bot will need to win. Pick what Lisa asks.

Offsite Assistant● connected
HOW SCORING SEES THIS

An LLM judge reads each exchange and asks one thing: was the answer correct, useful, and well-reasoned? Handling trade-offs and edge cases is where points live.

LISA ASKS →

FINAL

How you win

One number decides the leaderboard. Everything else is a tiebreaker.

0%GOAL ADHERENCE

After 7:00pm, every bot is run against 50–100 labeled conversations. Each specifies a user message, an expected action, and a desired outcome. An LLM judge scores whether your bot achieved the stated goal.

RANKING

Goal adherence %

% of conversations where the bot's output matched the expected outcome. Ranked descending.

TIEBREAKER

Coverage %

% of flows your harness successfully generated.

🔒 The dataset is hidden until after cutoff. The use case is open from day one — the test conversations are not. Build a robust bot, not one tuned to the evals.
BOARDING

Submission checklist

Everything below is due by cutoff. Check items off as your team ships them.

0 of 7 boarded

BUILD SOMETHING LISA WOULD TRUST · GOOD LUCK, TEAMS

← / → to navigate