Lyft unveils AI 'co-pilot': The secret tool promising to boost driver earnings

Ride-hailing giant Lyft has rolled out a new artificial intelligence tool designed to help its drivers make more money. The 'Earnings Assistant' provides personalised tips on where and when to drive, potentially changing the game in the competitive gig economy.

For the millions of drivers working for services like Lyft and Uber, learning the city's secrets to maximise earnings can take months or even years of trial and error. With intense competition for every fare, figuring out the most profitable strategy is a constant challenge. Now, Lyft is introducing an AI-powered solution to level the playing field.

Over the past year, the company has been rolling out its 'Earnings Assistant', a tool that uses artificial intelligence to give drivers a competitive edge. It synthesises vast amounts of data - from real-time ride requests to historical patterns of demand - to offer personalised recommendations.

"Drivers want to earn more, and they see various signals in the app, but those signals are not tailored to drivers' personalized needs," a Lyft senior software engineer, Xiaoyi Duan, told Business Insider. The new tool aims to change that by acting as a smart co-pilot.

A tool for every driver

The Earnings Assistant has two distinct features designed for drivers at different stages of their careers. For those just starting out, 'plan guidance' offers advice on where and when to work to increase their chances of finding fares. This feature, available to all US drivers, is meant to combat the initial feeling of being overwhelmed.

Yuko Yamazaki, vice president and head of driver at Lyft, told Business Insider that many new drivers are swamped by options when they first use the app. The plan guidance simplifies this by offering suggestions like, "Here is what the next two hours could look like."

For more experienced drivers, a feature called 'real-time guidance' is being tested in a few cities. This function pinpoints exactly where customer demand is surging at any given moment. An example could be guiding a driver towards a stadium just as a concert is ending, or suggesting a final trip from the airport during the busiest arrival window. "The real-time signals are becoming more of an interest for drivers as they become more mature on our platform," Yamazaki said.

The bigger picture in a changing industry

Lyft’s AI push comes as the entire gig economy faces a period of transformation. Its main rival, Uber, is reportedly beta-testing a similar AI feature to help its drivers earn more.

This competitive environment is forcing platforms to offer more transparency and control. In early 2026, Lyft moved away from what some called a "black box" algorithm and introduced features like wait-time compensation and a cap on the percentage it takes from passenger payments.

Driver earnings remain a key focus. A 2026 analysis of data from the previous year indicated that the median gross hourly pay for a Lyft driver was $20.38 before expenses, with vehicle costs estimated at between $5 and $10 per hour. Lyft’s own data for the latter half of 2023 showed a median gross earning of $30.68 per hour of "engaged time," meaning time spent on a trip.

However, the growing use of AI is not without controversy. Concerns have been raised about the potential for "algorithmic wage discrimination," where platforms could use AI to personalise pay rates, leading to unpredictable income for workers. This comes as regulators worldwide are increasing their scrutiny of the gig economy.

Lyft said it developed the Earnings Assistant over the past two years, testing early versions with drivers at events in Dallas, Las Vegas, and Miami. The company plans to expand the tool's capabilities, with future versions potentially alerting drivers to nearby events or explaining how weather is affecting earning opportunities.

As the industry moves forward, with autonomous vehicles looming on the horizon, the focus remains on the human element. "We are not trying to build an AI product because we want to use AI," Duan said. "We're trying to find what are the actual driver needs. That's the fundamental thing."