Inspiration

We were inspired by how often dining decisions still happen inside messy group chats: “Where should we eat?”, “Who is joining?”, “Any allergies?”, “Which branch is closest?”, “Can someone book?”. Even with many restaurant apps and websites, users still need to search, compare, discuss, call, fill forms, and wait for confirmation.

We believe the next interface for hospitality is not another booking form, but an AI agent inside the conversation. For private occasions, group dinners, business meals, and celebrations, users should be able to ask naturally in ChatGPT, Slack, or similar platforms — and the agent should help decide, coordinate, and complete the reservation.

What it does

Norra AI Reservations is an agentic dining assistant that helps individuals and groups move from conversation to confirmed restaurant booking.

The agent understands dining intent, group context, cuisine preference, time, location, budget, allergies, and occasion. It can recommend suitable restaurants or branches, check reservation logic through Norra’s hospitality infrastructure, and return the result in the same conversation.

For Norra merchants, the reservation can be made directly through Norra’s reservation management layer. Because Norra already has reservation management and POS as core products, the long-term goal is to support the full dining journey: booking, dine-in context, guest preferences, staff preparation, payment, feedback, and post-purchase engagement.

How we built it

We built the project around an agent workflow that starts from a chat conversation and routes the user’s request into restaurant operations.

The current hackathon prototype focuses on Slack as the first group-chat interface. The architecture is designed to later extend into ChatGPT and other personal or group chat platforms. We used workflow orchestration to connect chat input, intent extraction, routing, restaurant search, reservation logic, and response generation.

Norra acts as the core operating layer. The agent connects to Norra’s reservation and POS context so it can move beyond simple Q&A and perform useful hospitality actions.

Challenges we ran into

The hardest challenge was not generating a response — it was making the agent operationally useful. A restaurant booking requires structured data: branch availability, booking rules, table capacity, guest preferences, allergies, timing, and confirmation logic.

Another challenge was designing for both group chat and personal chat. Group chat needs coordination and consensus, while personal chat needs fast decision-making and personalized recommendations. We had to think about an agent that can support both modes without changing the core reservation flow.

Accomplishments that we're proud of

We are proud that the project shows a clear path from AI conversation to real-world restaurant action. The agent is not only a chatbot; it is designed as a connector between chat platforms and restaurant operating systems.

We are also proud of the roadmap: starting with Slack for group dining, then extending to ChatGPT and similar platforms for personal dining decisions, and eventually expanding into the full hospitality experience through Norra’s reservation management and POS foundation.

What we learned

We learned that dining is a highly contextual decision. People do not choose restaurants based only on date and time. They consider mood, occasion, distance, budget, dietary needs, guest preferences, and group dynamics.

We also learned that agentic AI becomes much more powerful when it connects to real operating systems. The value is not just in answering “where should I eat?”, but in helping the user decide, coordinate, book, and continue the experience after the meal.

What's next for Norra AI Reservations

Next, we want to strengthen the Slack MVP and connect it to more live Norra merchant data. After that, we want to expand the agent into ChatGPT and similar personal AI platforms so users can make faster dining decisions for private occasions.

Our roadmap is:

Start with Slack for group dining coordination.
Extend to ChatGPT and other personal or group chat platforms.
Expand from reservations into the full dining experience.
Later, apply the same agentic commerce layer to similar industries such as nightlife, events, wellness, beauty, and personal care.