Europe’s urban shared mobility needs a brain — SWITCH built one.

Cities are under mounting pressure to manage an increasingly complex mobility landscape. Shared e-scooters, bikes and cars, on-demand delivery fleets, EV-charging needs, shifting commuter patterns, and new regulations have made urban transport harder to plan and optimise than ever. 

Operators are often left juggling siloed data, unpredictable demand, and costly manual decision-making, while city authorities struggle to design infrastructure that keeps pace with real behaviour on the ground. 

SWITCH (short for "Street Witch) is an Italian startup that provides agentic AI that can simulate, forecast, and act in real time, helping mobility and logistics stakeholders move from reactive operations to intelligent, data-driven systems that run efficiently and serve cities better. I spoke to SWITCH CMO Simone Ridolfi at the recent Bologna Gathering to learn more.

From the Rome car-sharing App to AI mobility intelligence

SWITCH was founded in February 2020. It originally started as a consumer app in Rome, to match demand and supply, enabling users to  "switch" cars with people.

Then COVID hit. And while everything stopped, the team used that time to listen to operators and pivot their app to cater to people eschewing public transport for car sharing and micromobility.  

Ridolfi says the turning point came when the company hired specialised talent — including a Chief AI Officer — and doubled down on solving operators' real-world problems with AI and data.

"We never stopped talking to operators. That's how we built what they actually needed," he says.

Inside SWITCH's AI toolkit

In response to industry needs, SWITCH has developed a suite of AI tech, including:

• Urbiverse, a simulation and synthetic-data engine for "what-if" modelling, fleet sizing, and infrastructure planning such as parking hubs and EV-charging networks.
• Urban CoPilot, an operations-optimisation platform that supports demand forecasting and fleet rebalancing;
• SWITCH AI Agent, which connects planning and operations through real-time forecasting, simulation, and autonomous or semi-autonomous decision-making.

The power of open data

Cities generate a significant amount of open public data, including street layouts, traffic flows and congestion, parking information, and event details.

"If you have a strong model to predict demand and use all this open data, the value becomes very interesting," contends Ridolfi.

The startup works with shared mobility operators, carpooling platforms and Demand-Responsive Transport (DRT) services —primarily in micromobility, but also increasingly in car-sharing.

Currently almost all shared mobility providers use historical data models, but they're not very precise — or they rely on gut feeling. 

"For example, they think, 'Okay, let's put all the scooters in the city centre because a lot of people go there.' That might work for one day, but if you distribute scooters across the city based on predicted demand, you'll get more rides. We help them do that and calculate the impact. On average, we see a 25 per cent increase in operational efficiency," shared Ridolfi. 

"We help operators forecast future demand with much greater precision," Ridolfi explains.

"If you know what the next weeks or months will look like, you can plan and operate far more effectively."

Further, a car-rental operator might use their demand-forecasting module to decide how many vehicles to buy/position in different zones, when to offer discounts, and when to relocate vehicles. SWITCH also uses flight event data — arrivals and departures — to predict demand so providers know how many cars to position at airports.

Real-time data responsiveness 

SWITCH's AI agent that connects a company's data with external platforms and with its demand prediction and optimisation models. You can receive data in a minute, and two minutes later, you'll get suggestions on what to do.

"So you can ask it — like you do with ChatGPT — things such as: "What do I need to do next month to reach €1,000 more revenue?" or "Which areas will be affected by Sunday's strike?" The agent analyses everything and provides recommendations."

A shared mobility provider can use SWITCH's demand prediction to understand demand concentration in different city zones to know where to place scooters. Its rebalancing and forecast tools can guide them to proactively move vehicles to high-demand spots, avoiding oversupply/undersupply.

For a new mobility launch, that means knowing how many vehicles to deploy, how many rides to expect, and when you'll hit break-even.  Crucially, SWITCH can also help companies determine whether to enter a new market. For example, one micromobility company wanted to operate in a city in Norway.

According to Ridolfi, "there are tenders, but they didn't know if they would be profitable — how many competitors, how many rides, etc. 

"We built a dashboard to assess whether entering that market made sense. Ultimately, they decided not to enter. So we saved them money and risk because they knew in advance what could happen."

It would have been beneficial in Berlin, which at one point had seven different e-scooter and e-bike operators competing for the same streets. Although operators are reluctant to share vehicle utilisation data, research indicates a single shared e-scooter is often used fewer than three times per day, for trips averaging under 1.5 km. That means long periods of idle time and significant public-space clutter relative to actual mobility output.

Urbiverse powered a launch with shared mobility operator Wayla by modelling every key dimension of their rollout — from fleet size and vehicle placement to ride volume and profitability.

The simulation, built on real-world data and dynamic modelling, delivered 92 per cent accuracy when compared to actual launch outcomes. 

This level of precision enabled the operator to transition from guesswork to a data-driven strategy, significantly reducing risk and refining deployment decisions.

Data-Driven policy (and clarity) for cities and operators

Further, SWITCH's Urbiverse platform enables local governments to optimise shared mobility fleets, strategically place micromobility hubs and EV charging stations, and simulate policy impacts before implementation. 

By generating synthetic data when real-time data is unavailable, Urbiverse ensures officials can make informed decisions despite data gaps. Urban Copilot enhances Mobility-as-a-Service by predicting fleet availability so vehicles are in the right place at the right time.

Another point is policy. Ridolfi contends that "cities need time to understand the impact of micromobility — but often they don't have the tools to measure it."

Take London, where public e-scooters have been in trial phase since 2028 and are set to run until 2028. Yet mobility providers invest now, and then one day the city might say "No more scooters," like in Paris, leaving dozens of vehicles to be sold.

"With SWITCH tools, operators can also become part of the city's planning process. They can be proactive, not reactive," shared Ridolfi.

"For example, the city might say, 'No free-floating scooters," and instead require hubs. Using our tools, you can plan hub placement based on real demand.

In Turin, we worked with the city to study where bike hubs should be placed so that bikes don't interfere with traditional mobility and still satisfy citizens' needs."

SWITCH raised €600,000 from private investors in January, including EIT Mobility and Berkeley SkyDeck — the startup was part of Berkeley SkyDeck's first acceleration program in Milan — as well as around €400,000 in public grants since its inception. The startup is currently part of the NVIDIA Inception Program.

And, as its reach expands, the days of idle scooters, underused fleets and reactive policy may finally be numbered.

Lead image: Freepik.