Amara - Personal AI Concierge

Inspiration

Modern hospitality demands instant, reliable, and premium engagement. We wanted to build a true digital concierge for the Eko Grand Hotel in Victoria Island, Lagos, that moves completely past standard, rigid chatbots. By naming her Amara, our goal was to create an intelligent assistant capable of providing the high-touch experience expected at a world-class luxury hotel.

What it does

Amara acts as a personal AI concierge for hotel guests. At runtime, she seamlessly greets users, handles real-time room availability inquiries, provides transparent pricing details, checks up on active hotel amenities, and assists automatically with room cancellation requests.

How we built it

Amara is built entirely from the ground up using a strictly approved Google Cloud and partner ecosystem with no competing third-party cloud or AI services:

• Google Cloud Agent Builder: Orchestrates the runtime session management and intent handling logic.
• Google Gemini: Powers the natural language conversational brain, allowing fluid and context-aware responses.
• MongoDB MCP Server: Handles the backend data storage layer, dynamically serving live hotel data variables directly to the agent.

Challenges we ran into

One of the key hurdles was building out a strict, real-time data sync via the Model Context Protocol (MCP). Tightly structuring instructions so that Gemini accurately handles pricing and availability parameters dynamically without hitting runtime exceptions requires continuous prompting iterations.

Accomplishments that we're proud of

We are incredibly proud of building a highly responsive, custom-styled front-desk user interface that connects perfectly to real data backends. Seeing Amara seamlessly switch between room lookups and policy FAQs with zero lag feels great.

What we learned

We learned how to efficiently utilise the Model Context Protocol (MCP) to supply external knowledge databases directly to large language models. We also mastered the art of optimising prompt tokens for fast execution inside Google Cloud Agent Builder.

What's next for Amara - Personal AI Concierge

Next, we plan to scale Amara by integrating voice capabilities so guests can talk to her over local messaging channels, expanding her integration to handle physical smart room automation (like lighting and temperature controls), and adding multi-language support for international travelers visiting Lagos.

Built With

• google-adk
• google-cloud
• google-cloud-agent-builder
• google-gemini
• javascript
• model-context-protocol
• mongodb
• pymongo
• python