How SpareLabs Put AI Voice on the Front Line of Public Transit And Watched It Take Off

Spare Labs is the enterprise software platform cities and transit agencies use to run more efficient transit services, with customers including BART and municipalities across North America and Japan.

Spare serves two audiences: the agencies that buy the platform, and the riders those agencies serve. That second group is where the problem started.

The Challenge: The Riders Who Need Service Most Were Waiting the Longest

A significant share of Spare’s customers operate paratransit services for people with disabilities. For many of those riders, a phone call is the only way to book a trip, check an ETA, or get an update.

The experience was often painful. Call centers were often outsourced to third parties and couldn’t keep up. Hold times stretched long. Staffing was thin. And every minute on hold was a minute a rider couldn’t plan their day.

Spare knew voice AI could fix this. They just couldn’t get it to work.

“Back in 2023, we were very excited by the latest in AI voice models. But we struggled to get them to work as well as we wanted them to.”

The team shelved the idea until early 2024, when they took another look.

Why Retell AI

Josh Andrews evaluated several vendors in the early days, including Bland. What set Retell apart was a combination of speed and substance.

“Retell had better latencies, it was a better product that we could interact with, and it was a bit faster to set up. Today, the built-in support Retell has and the ability to quickly innovate and stay up to date with the latest models has been really helpful. There’s a lot of functionality you guys provide that, if we had to build it ourselves, would take us a lot of time.”

In March 2024, Spare went live with Retell in early-access deployments.

The Solution: A Purpose-Built Voice Agent for Every Transit Agency

Because Spare’s customers all face similar operational challenges, the team built a single, configurable agent that can be turned on for any transit agency in a click. The agent already knows transit flows, ride-booking logic, and the tool calls required to reach into Spare’s platform.

“The configuration process is quite simple for us now. We don’t have to go in there and spend weeks to build it. You can basically turn it on in a click.”

The agent integrates with a single system of record: Spare itself. Because Spare is the operating system for the transit agency, every data point the agent needs—ride status, driver location, account info, service changes—is already available through one integration. Today, Spare’s agents handle a growing catalog of tasks:

Inbound

• Booking and modifying paratransit and microtransit rides
• Checking ride ETAs and driver status
• Answering service FAQs (schedules, luggage policies, lost-and-found)
• Creating support tickets

Outbound

• Calling riders on behalf of drivers who can’t locate them at pickup, confirming location, and relaying that information back to the driver

Most agencies deploy the agent as a first line of defense, with a fallback to human agents for edge cases. Increasingly, Spare’s customers are routing the agent 24/7.

From 30% to Near-100% Call Resolution

Within the first week of going live, Spare’s agents were already handling around 30% of inbound calls, immediately cutting hold times at customers that were chronically understaffed.

Two years in, the ceiling has moved dramatically.

“Our average customer is close to 70% resolution. For some customers, the simpler operations or those without as many complex cases, it’s close to 100%. We’ve seen a dramatic increase over the last couple of years.”

The lift came from steady iteration on prompts, tool calls, and voice quality paired with rapid model improvements from Retell.

A Record Snowstorm in Milton, Ontario

In late January 2026, a record snowfall hit the town of Milton, just outside Toronto. The transit agency rewrote its entire service plan overnight: buses on new schedules, routes altered, thousands of riders suddenly needing updates.

A human call center would have drowned. Instead, Milton flipped on Spare’s voice agents and absorbed the surge.

“The AI agent was able to handle all of those customers and the influx of calls about the changes of service plans. It was extremely positive for the town and enabled them to do something they would have not been able to do before.”

The Fastest-Growing Product in Spare’s History

Voice AI isn’t just working at the rider level—it’s reshaping Spare’s business.

“The growth of our AI product has been probably our fastest-growing product of all time. In the past year, the AI product has over 10x growth. And it was already looking pretty good a year ago.”

By the Numbers

• 10x+ year-over-year growth in Spare’s AI product line
• ~30% of calls automated in the first week of deployment
• ~70% average resolution rate across customers
• Up to 100% call resolution at top-performing customers
• ~1 month to build the initial Retell integration
• 500+ evaluation flows Spare now runs against its agents

Looking Ahead

Spare is now rolling the AI agent into new geographies, including Japan, and continuing to expand its catalog of supported workflows. Andrews sees the partnership as a long-term bet on innovation velocity.

“We see the opportunity to roll out more and more voice agents to cities and to our customers, which is quite exciting. It’s going to be very valuable to have a good partner that can continue to help us adopt the latest technology and latest greatest models. If we’re at 60% resolution, we’d love to get everybody to 100%. Having a partner that’s willing to innovate quickly with us is going to be really critical.”

For the riders on the other end of the line—often those with the least margin for error in their day—that innovation has already paid off. The hold time is gone. The ride gets booked. The bus is on its way. And in towns buried under record snowfall, the phones keep answering.