DriveMate Daily

Inspiration

Tasco's brief was clear: they already have a large mobility ecosystem, but they want to move user engagement from monthly active to daily active. That inspired us to focus less on "one more car app" and more on a daily AI companion for car owners. We asked a simple question: what would make a driver open the app every day, not just when they need to pay tolls or book service?

What it does

DriveMate Daily is an AI assistant for the full car ownership journey. It combines toll, parking, service history, trip context, maintenance timing, repair signals, and user habits to generate daily-value actions.

Core features in the MVP:

• daily driving digest with parking, toll, weather, and route tips
• smart maintenance reminders based on mileage, service history, and symptoms
• photo/chat-based repair triage for common car issues
• personalized upsell recommendations for wash, quick service, insurance, and partner offers
• engagement loops such as streaks, savings insights, and "things to do today"

The goal is to give users a reason to come back every day, while also creating new service and revenue opportunities inside the Tasco ecosystem.

How we built it

We built a lightweight web MVP with a React frontend and FastAPI backend. Qwen is used for intent understanding, symptom-to-action reasoning, and natural-language recommendation generation. A rules layer handles deterministic triggers such as overdue maintenance, expiring insurance, or location-based parking/toll prompts. We modeled the user around a unified vehicle profile, service history, driving habits, and task timeline.

The prototype flow works like this:

1. ingest basic vehicle, trip, parking, toll, and service events
2. generate a daily driver context
3. rank relevant nudges and service actions
4. let the user chat with the assistant for explanations or next steps
5. push the result into a simple dashboard and mobile-friendly assistant UI

Challenges we ran into

The hardest part was deciding what "daily value" actually means. Most mobility apps are opened only when there is a transaction, so we had to design for habit, not just utility. Another challenge was balancing helpfulness with noise; too many notifications would feel like spam. We also had to separate deterministic safety-critical actions from generative suggestions so the product stays practical and trustworthy.

Accomplishments that we're proud of

We're proud that the concept is tightly aligned with Tasco's real business problem instead of being a generic chatbot. We turned a broad "increase DAU" challenge into a concrete product loop with clear triggers, user value, and monetization paths. We also built the flow so the same assistant can support both users and internal service teams later.

What we learned

We learned that the strongest mobility AI products are not only about navigation or chat. They win when they connect fragmented car-owner moments into one continuous experience. We also learned that Qwen is most useful when paired with a rules engine and structured user context, rather than used as a free-form answer generator for everything.

What's next for DriveMate Daily

Next, we want to connect the MVP to real event streams such as parking entries, toll activity, service bookings, and vehicle health signals. We also want to add a stronger recommendation feedback loop, so the assistant improves based on what each driver actually clicks, ignores, or books.

Built With

• alibaba-cloud
• analytics
• fastapi
• maps-api
• notification-service
• postgresql
• python
• qwen
• react
• redis
• typescript