‍

A car-sharing marketplace serving millions of guests worldwide replaced dense, step-by-step roadside scripts with Zingtree guided workflows, shortening its longest call type, lifting quality scores, and raising satisfaction for both guests and hosts.

The challenge: High-variability calls and heavy cognitive load on agents

Roadside assistance is one of the hardest interactions Turo's agents handle. Every call carries variability: the vehicle, the location, the guest or host on the line, and the situation on the ground. 

Turo's underlying process lived as a linear, step-by-step block of text that agents had to read, interpret, and map to a live call.

• Roadside calls branch constantly on “ifs and thens” that a linear script can’t capture cleanly
• Agents faced a wall of text and had to work out which step applied to the caller in front of them
• Accounting for every possible scenario in a static document was hard to maintain
• Roadside was one of Turo's longest calls on average

“Where it was really difficult was when we hit ifs and thens. With a roadside call, there’s a lot of variability, a lot of things for an agent to consider. It left agents looking at a block of text that might apply to their situation, and they were trying to figure out what step to pick up next.” – Betty Miller, Sr Program Manager, Turo

Implementation: Trained agents, owned by the people who write the policy

Turo took its time with the rollout. Every agent joining the roadside team learns Zingtree during onboarding, and the training covers why they're using it, not just how to click through it. 

The workflows are built and maintained by the people who write the policies. When Sr Program Manager Betty Miller writes the roadside processes herself, she can update the tool right away when a policy changes, from a major process change down to small tweaks like reformatting a step or switching "continue" to "next" because an agent asked for it. 

When Betty sat with agents earlier in the year to gather input on the tool, she was able to incorporate all of their suggestions in about two hours.

“It was important for somebody like me to be able to own this. I’m the one who writes the processes, and I make changes really quickly, and we wanted to be able to execute those changes whenever we needed to.” – Betty Miller

Solution: Guided workflows that prompt only what’s relevant

Turo built its Roadside Assistant workflow in Zingtree so agents are guided step by step through each contact. Agents click through the flow based on what the caller tells them, and Zingtree surfaces next-step actions and suggested talking points and scripts. 

Crucially, the workflow only prompts an agent to do something when it’s relevant to the contact at hand, which keeps every conversation efficient and focused on the caller’s actual situation.

• Agents are guided through each call and prompted only when a step actually applies
• Next-step actions and optional talking points keep conversations on track
• Relevance-aware prompting removes guesswork and shortens the path to resolution
• Separate workflows per region keep US and Canada consistent while allowing for local nuance

“It only prompts an agent to do something if it’s relevant to the contact that they are working on. So it makes their conversation more efficient.” – Betty Miller

Results: Shorter calls, higher quality, happier guests and hosts

Roadside was one of Turo's longest calls on average, and reducing that length was the problem the team set out to solve. Over the past year, Turo focused on driving adoption of the tool, and the results have spoken for themselves. 

Reporting features help the team see where agents end their interactions or aren’t using the tool, so they can hunt down the why and keep refining the experience.

• Shorter average handle time on what was one of Turo's longest call types
• Higher quality scores, reflecting more consistent handling across agents
• Rising satisfaction scores for both guests and hosts
• Consistent processes across US and Canada teams, with room for country nuance
• Reporting surfaces exactly where to refine the workflow next

“I’m most proud of the way that our satisfaction scores have climbed for guests and our hosts. Being able to know that guests and hosts are walking away from that conversation feeling good about that interaction is really cool.” – Betty Miller

Looking ahead

The results from roadside caught the attention of Turo's leadership, and the company is now extending Zingtree into other parts of the business. Interest has come from several departments, and the claims team is already building its own workflows, partnering directly with Zingtree engineers and CSMs on the build.

• Extending Zingtree to new work streams beyond roadside
• Executive sponsorship driving the expansion, prompted by roadside results
• Claims team already building, working closely with Zingtree engineers and CSMs

“This came from one of our executives. They’ve been impressed with the results we’ve gotten from Roadside, and they wanted to explore what it would look like to use it across other work streams.” – Betty Miller

About Turo

Turo is a car-sharing marketplace where guests book vehicles from a community of local hosts. Founded in 2010 and headquartered in San Francisco, Turo operates across the United States, Canada, the United Kingdom, Australia, and France, connecting hundreds of thousands of hosts with millions of guests worldwide. Its operations teams write and maintain the policies and processes that support guests and hosts whenever they need to contact Turo.

About Zingtree

Zingtree is the agentic context and decision layer that orchestrates complex, high-stakes support workflows safely, accurately, and at scale.