MiCa.

Step 01

Research and Problem Identification

We began by asking how trust is constructed and value is perceived during guest-host interactions. From the guest side, users like 'Eric' dislike generic recommendations and expect hotel-quality responsiveness from hosts. From the host side, non-professional hosts like 'Carlotta' find management exhausting and costly, with no way to predict how future guests will behave. The core insight: psychological research shows that place attachment — built through chatting and shared activities — contributes to the meaning guests attribute to a space. Designing for emotional connection helps guests care for the property.

Step 02

Conceptualizing the Solution: MiCa

We developed MiCa, an IoT system that welcomes and guides guests throughout their stay. It acts as a digital co-host, supporting the guest while the actual host remains behind the scenes. MiCa provides essential information — Wi-Fi passwords, local pub recommendations, house rules — in a way that feels personal rather than automated.

Step 03

Prototyping the MiCa System

Our solution consists of two primary components. The MiCa App lets hosts 'train' their AI co-host using reinforcement learning and handles the administrative side — check-out checklists, booking management through platforms like Airbnb. The MiCa Totem is a physical IoT device in the apartment that gives guests context-aware recommendations and instructions. See the video below that demonstrates how MiCa works in Airbnb home.

Step 04

Evaluation and Findings

We refined the prototype based on specific interaction goals.  The first prototype underwent a peer review using the think-aloud technique. We identified several navigation issues, and found that limiting interaction with the AI to chat alone may not help users engage effectively with the product. The final UI was updated to convey a calm, premium look. We shifted from training the host AI, MiCa, entirely through chat to enabling interaction through a map that recommends spots for guests and maps objects within the house.

Step 05

Future Directions

We established a roadmap for further technological integration. The next step is testing to measure how this emotional bond influences long-term guest satisfaction and engagement. We plan to include Tiny Machine Learning to enable presence-based automation, allowing the totem to enter an energy-saving 'sleep mode' when no one is around. Finally, we aim to create a MiCa-enabled network that connects with other automated home systems, such as lighting.