As the global cycling community gathers in Gdańsk this week for Velo-city 2025, the Polish city’s urban mobility credentials are firmly in the spotlight.
With nearly 850 kilometres of bike paths and a strong cycling culture, Gdańsk has long been recognised as one of Poland’s most bike-friendly cities.
In 2023, a new era of public cycling began with the launch of the RowerMevo bike-sharing system, a vast, multi-municipal network encompassing over 4,200 bikes and 717 virtual stations across 16 metropolitan areas. Operated by Inurba Mobility, MEVO has rapidly become one of the largest and most ambitious bike-sharing schemes in Poland, already clocking over a million rides and four million kilometres travelled in less than a year.
At the centre of this rollout is a deep operational challenge: how to manage such a sprawling, decentralised, mostly electric fleet with limited infrastructure. To meet that demand, Inurba partnered with French tech company Qucit. Its AI-powered tools now help drive the MEVO system’s day-to-day efficiency, guiding bike redistribution, flagging missing assets, and prioritising battery swaps across a complex urban landscape.
Łukasz Jezierski, City Manager at Inurba Mobility, and Raphael Cherrier, CEO of Qucit, tell Zag Daily how AI and local expertise are driving this ambitious micromobility project forward.
How AI improves daily bike operations
Managing the MEVO system is no small feat. “We could write a book about it,” jokes Łukasz Jezierski, City Manager at Inurba Mobility. The scheme now includes around 4,300 bikes, 75% electric, spread across a wide area, all without physical docks or integrated charging infrastructure. That means over 1,000 e-bike batteries need to be manually replaced each day, a process that demands a small army of field staff.
While charging stations are in the pipeline with 51 scheduled under a regional transport authority initiative, MEVO’s present-day reality is grounded in boots-on-the-ground logistics. Each day starts with field teams navigating an urban sprawl to locate, service, and rebalance the fleet.
The reliance on virtual stations—designated parking zones without physical docks—introduces a significant challenge that goes beyond simple unpredictability: data reliability. A bike can be perfectly parked in a designated zone but remain a ‘digital ghost’—unassigned to the station due to a back-office flaw. As Qucit CEO Raphael Cherrier explains, “We have to solve the data problem before we can solve the logistics problem.” Free-floating bikes can also end up in hard-to-reach or inaccessible locations, complicating both battery management and service levels.
The operational targets reflect the sophistication of the service required. Inurba must meet a detailed framework of service indicators with targets that nest within multiple geographic levels—from individual stations to entire city zones. “It’s like Russian matryoshka dolls: KPIs within KPIs,” Cherrier chuckles, acknowledging the level of precision involved.
Qucit’s system helps manage that complexity by using predictive algorithms to proactively anticipate rider demand, allowing teams to stay one step ahead. It generates optimised routes for battery swaps and fleet rebalancing, functioning as a live operations map that suggests tasks, prioritises hotspots, and monitors fleet health. “You can’t manage this kind of operation with Excel files,” Jezierski says. “You need a system that can keep up.”
Building trust in new technology
Deploying an AI-based system in a live environment requires more than just technical integration. It also demands cultural buy-in. At first, Jezierski admits, field staff were sceptical of the app’s instructions. “They trusted their instincts. But over time, they realised the system sees more than they can from the ground.”
Feedback loops between field teams and developers—a core part of Qucit’s collaborative approach—played a major role in refining the system. “Every day we collect input from the ground,” Jezierski says. “Sometimes it’s a small tweak that makes a big difference in practice.” The result has been a gradual shift: early trust in the application was low, but now sits around 90%.
From Qucit’s perspective, Poland posed unique operational and cultural dynamics. “A lot happens at night, and at first it wasn’t easy to track or fully understand everything that was going on,” Cherrier recalls. Local insight proved critical. “This is why we had a highly committed Polish team member who helped coordinate with field operators and also translated the app.”
The real-world scenarios were not always what the algorithms had anticipated. Bikes were found in rivers, lakes, forests, even inside damaged buildings. While unpredictable bike placement continues to be a daily challenge, GPS-equipped bikes help, as do internal protocols. If a bike has not moved or had a battery swap in several days, the system flags it. Field staff then follow up based on known high-risk zones. With virtual stations, that level of active management is essential.
The partnership’s maturity is now reflected in the technology itself. The Qucit platform has evolved to support specific in-app commands for operators, like unlocking a bike or releasing the battery, and is constantly refined based on direct user feedback. “We’ve come a long way since the start,” says Jezierski. “We understand how we work and what we need now.”
Pioneering Polish micromobility
What makes MEVO particularly noteworthy is not just its scale, but its timing. While many Western European cities have long embraced electric shared mobility, Poland is only beginning to adopt these systems at scale. MEVO is the first major public network in the country to feature such a large share of electric bikes.
That’s helping shift attitudes. “People used to associate cycling with sweat and sport,” Jezierski says. “Now they see it as a clean, fast way to get to work.”
Although environmental metrics are typically tracked by local government partners, Inurba has taken steps to green its own operations. Electric vans and 16 e-cargo bikes are now used in larger cities to handle swaps and rebalancing.
Looking ahead, both partners recognise that complexity will only increase as the fleet expands. But they’re optimistic that their joint approach of blending operational experience with AI-driven operational intelligence has laid a solid foundation.
“In Gdańsk, we’re seeing up to 8.6 trips per bike per day,” Jezierski notes, a usage rate that compares favourably with leading systems across Europe. Across the full network, averages are slightly lower due to less dense areas, but the numbers still require daily rebalancing and around-the-clock battery management. “Without a smart platform, this level of service would be impossible,” he says.
As the Velo-city conference unfolds, MEVO and Qucit will be showcasing the benefits of their partnership. The city plans a parade of 400 mechanical bikes and will open its operations hub to visitors. “We’re proud to show how the system works in real life,” says Jezierski.
For Cherrier, it’s also a chance to highlight what AI can do when paired with the right people. “We don’t build tools to replace decision-makers,” he says. “We build tools to support them.”
