Lyft has been a fringe player in the self-driving car race for the past few years, but its new plan for the technology takes a more pragmatic approach than a lot of its competitors.

On Friday, the company announced it was going to formally start a branch in Palo Alto, California, to develop autonomous car technology. Company executives announced the branch is expected to hire “hundreds” of new employees, who will work on an “open self-driving platform” that can be integrated into Lyft’s existing network. Most of this is par for the course, but Lyft’s strategy relies on a “hybrid ecosystem of both human-driven and autonomous vehicles,” which means human drivers will still have a role. Lyft’s decision represents the most realistic acknowledgment of the limitations of autonomous vehicle technology a company has made so far.

Other companies, like Lyft’s main competitor Uber, have thrown millions, if not billions of dollars into developing a fully self-driving car system to replace their entire driver workforce. While that’s undoubtedly Lyft’s eventual goal as well, the company’s strategy is explicitly aimed at the near-future, when limited self-driving cars will share the road with traditional vehicles. Lyft’s new network relies on “smart dispatching” that will send the right vehicle for a customer’s trip. This is an important distinction, as experts predict that self-driving cars will be relegated to specific areas, called “geofences” for years to come. In other words, if you order up a Lyft to get from, say, 14th street to 34th street in Manhattan, New York, a heavily-mapped area on an easy grid system, the service might send you an autonomous car. But if you wanted to take a Lyft on a longer trip, like from the center of a major city to the suburbs, Lyft would send you a human driver that could navigate in areas that potentially aren’t yet accessible to autonomous cars.

“It is not just going to happen tomorrow,” Taggart Matthiesen, senior director of product at Lyft told the Financial Times. “It is not going to be some mechanized fleet blanketing San Francisco. It will take time. What you will see is in small pockets, in isolation, these vehicles will start providing service.”

Lyft has explicitly stated that its goal is to be “Switzerland” in the self-driving car wars, avoiding costly lawsuits and scandals and trying to instead forge partnerships with the various mega-companies working on the technology. Lyft has taken $500 million in money from General Motors and signed a high-profile partnership with Waymo, the biggest threat to its rivals at Uber.

The company’s overall share of the ridesharing market is still less than Uber, forcing it to work with comparatively fewer resources to make an impact. It makes sense for the company to focus on a middle-ground network blending autonomy and human drivers, so it can nail the next five years, instead of shooting for far-off plans like urban air transit and fully autonomous fleets. The company’s deal with GM wants to put an autonomous Chevy Bolt on the road by the end of this year, so we may see if this strategy is going to pay off sooner rather than later.