Mobility Global Challenge Lab 2026

Who owns the data from your commute?

Two weeks to answer a question nobody asks. We built RoadVerse, and the useful part was killing the idea we started with.

The problem, in one moment

A delivery rider hits an unseen pothole and falls. Three parties need the same information immediately. His insurer wants proof. His platform wants to know what happened. The city does not know the pothole exists.

Three stakeholders. Zero shared intelligence.

177,175
Road deaths in India in 2024, close to 20 every hour
28.5%
Of Indian delivery riders crashed last year, driven by delivery time pressure
Under 2%
Of India's vehicles carry any connected intelligence
Road deaths, MoRTH 2024. Rider crash rate, IIT Bombay 2024. India holds roughly 1% of the world's vehicles and the highest road death toll on earth.

What we killed

We started with an AI smart helmet. We scrapped it. A rider who owns a helmet costing ₹2,000 will not spend ₹12,000 on one because it detects potholes.

So we changed the question. Not how to build better hardware, but what is the least hardware this needs.

What we built instead

Almost none. The phone already has the camera, GPS and processor to run the detection model on device. For cars that is the entire product. For two-wheelers a low-cost module handles what a phone cannot survive on a handlebar: vibration, heat and rain.

Running inference locally and uploading only flagged events holds infrastructure cost at roughly £0.30 to £0.90 per user per month.

The model

Core safety stays free, permanently. We monetise what the data becomes, not the safety itself: premium insights for riders, fleet analytics for delivery platforms, road intelligence for insurers and cities.

Free adoption builds the dataset. The dataset is what makes everything above it worth paying for.

What the judges pushed on

They liked the software-first approach and pressed us on detection accuracy across weather conditions, sharper differentiation, and a tighter link to the sustainable development goals. All three were fair. All three are on the roadmap.

The takeaway

The strongest version of an idea shows up once you are willing to kill the first one.

Every vehicle on the road is already a sensor. We are still finding potholes by hitting them.

Team

Built by Srishti, a four-person team across three cities and four disciplines: two engineers working in electric vehicles and robotics, one specialist in finance, risk and governance, and myself on strategy, business model and go-to-market. Global Challenge Lab 2026, powered by Imperial Enterprise and supported by Futurize.

All writing