ProtoPie
Car HMI — AI-Powered Vehicle Diagnostics
Built an in-vehicle HMI concept that goes beyond static warning lights — live vehicle data (tire pressure, coolant, battery) feeds directly into GPT-4o, which diagnoses issues and explains them to the driver in plain, conversational language. Built with ProtoPie Connect and demoed live at IAA Mobility 2025 in Munich.
- Role
- Creative Technologist
- Team
- Solo
- Timeline
- 2025
- Year
- 2025
- Live AI diagnosis engine
- GPT-4o
- Live AI diagnosis engine
- Shown live in Munich
- IAA Mobility 2025
- Shown live in Munich
- Hardware-integrated demo
- ProtoPie Connect
- Hardware-integrated demo
1 / 2
Videos
Process
Concept
From static warning lights to conversational diagnostics
- Built a car dashboard prototype displaying real-time metrics for tire pressure, coolant levels, and battery status, driven from a separate controller device.
- Connected the dashboard to GPT-4o via ProtoPie Connect — when a metric moves into a problem range, the live data is sent to the AI for diagnosis.
- Designed the AI's response to be conversational rather than a generic alert: it explains what's wrong and suggests a specific next step, optionally read aloud via text-to-speech.
- Configured two parallel ChatGPT API connections in ProtoPie Connect so the dashboard could request and receive diagnoses without blocking the interaction.
Outcomes
- Demoed live at IAA Mobility 2025 in Munich — one of the world's largest mobility trade shows.
- Showed how ProtoPie can validate AI-driven automotive features using realistic, dynamic data rather than scripted demos.
- Established a reusable pattern — live sensor data into an LLM, paired with a conversational UI layer — for prototyping proactive vehicle assistant concepts.
Skills and tools
- Interaction Design
- Automotive
- AI Agents
- Creative Technologist