According to TechCrunch, FiveAI, which is a startup in the United Kingdom seeking to deliver a fully autonomous shared transport service for Europe’s cities, will conduct its first on-street trial in 2019.

The trial is aimed at commuters in the London outer boroughs of Bromley and Croydon, and will initially start with a 10-month “data gathering” exercise that includes five FiveAI vehicles—equipped with drivers—collecting information about road conditions, the movement of pedestrians and various vehicles, and other variables to help train its artificial intelligence platform.

FiveAI, which has thus far raised $35 million to develop driverless car technology, is avoiding central London as a testing location, according to Ben Peters, FiveAI’s VP of product, and one of the company’s co-founders. The center of London is extremely congested, but there are already several transportation alternatives attempting to address that issue.

Operating its technology in Bromley and Croydon provides FiveAI a unique opportunity to fill a gap that isn’t being addressed by other companies.

“There are a lot of problems to solve there, but they are very well served by current providers,” Peters says. “But in Zones 4 to 6 [the outer boroughs of London], about one quarter of people are still driving their own vehicles to and from work.” This, in turn, creates an opportunity for a shared mobility service.

​Another benefit for FiveAI to test its technology in the boroughs is that the insurance company that is a partner of FiveAI’s, Direct Line, recently moved its offices to Bromley from Croydon, so there are employees living in Croydon who have to travel to Bromley on a regular basis, giving FiveAI the opportunity to build a service specifically for that market.

According to the TechCrunch article, it will take approximately 10 months before the first FiveAI vehicles can offer rides, and will probably take more months before no driver has to be present to engage the vehicle if something goes wrong.

For Peters, this “slow early work” can prove to be beneficial, though, as it will help the company add more roads, areas and cities to the service more quickly down the line.

“It doesn’t have to be as slow in the future, maybe just a few months to get up and going,” Peters says. “We’re really targeting an urban service, which is the difficult part. Autonomous driving in urban areas is the hardest challenge, one that is still unsolved.”