Toyota transforms IT service desk with gen AI

“One of my bold bets is I want to eliminate our traditional service desk by 2025,” says Jason Ballard, IT executive and general manager for infrastructure and operations services at Toyota Motor North America.

Ballard is also the technology executive responsible for both the company’s battery electric vehicle (BEV) platform as it shifts to electrification, and its digital platform engineering and architecture organization, and he counts on conversational AI and generative AI as major components to transform HR and IT service requests.

He’s also a big believer in the agile DevOps concept of “shifting left” when it comes to technology — performing testing and evaluation early in the development process, generally before code is written — and “shifting right” when it concerns talent, where his vision for eliminating Toyota’s service desk is an example. It’s not about getting rid of positions, however, but “insourcing” technical talent by developing team members to fill more valuable roles.

“At the outset, it sounds like we’re cutting jobs, but really, the focus is shifting right with the talent,” he says. “A level-one technician’s job is kind of repetitive and I think most people who are in those roles want to get to level two or three. What we’re focused on is how we can up-level, upskill, those team members and move them into higher-value jobs, while at the same time push more automation across the enterprise.”

Toyota is an early mover when it comes to generative AI, with roots in its robotic process automation (RPA) efforts. In 2014, Toyota Motor North America operated separate headquarters for sales and manufacturing, but in 2015, under the mantra, “One Toyota,” the company brought them together in Plano, Texas.

“As you can imagine, there were a lot of varying systems that were used to run the enterprise for those separate team members,” Ballard says. “We were very focused on getting to a common technology stack so our team members could be more efficient.”

As part of that effort, Ballard brought RPA on board in 2017, saving more than 150,000 recurring hours of staff work in the first year, he says. The RPA initiative made it clear there was more IT could do to help automate and optimize processes. Then in 2020, Ballard took leadership of the digital platform engineering team with an eye on the IT service desk.

“I wanted to change the way our team members engaged with our IT service desk; that was my initial focus,” he says. “I wanted the ability to invoke a conversation and get to a fast resolution so our team members and I could get back to the important work at hand.”

Centralizing help

In 2021, Ballard’s team partnered with AI platform specialist Moveworks to create a central place for Toyota’s 45,000-plus employees in North America to turn for help at work. Dubbed AgentAsk, the service offers employees a ChatGPT-like experience that takes into account enterprise requirements, including permissions, integrations, security, privacy, and more.

In the past year, Ballard says AgentAsk has resolved nearly 70,000 issues and accelerated the resolution of about 100,000 more. He notes that the service now allows the company’s service desks to focus on important accelerated tickets such as hardware requests and software approvals, rather than password resets and account unlocks; each month, AgentAsk resolves about 458 tickets of the former and 164 tickets of the latter.

AgentAsk’s mean time to repair (MTTR) is 11.4 minutes, compared with an industry average of three days. The service has given employees and agents back more than 70,000 hours of productivity in the past year alone, according to the company. On average, AgentAsk is offsetting the work of about 25 level-one service technicians each week, Ballard says.

By way of example, Ballard notes that in the past, he’d spend a lot of time digging around in various enterprise systems trying to figure out how many outstanding approvals he had to deal with.

“Now I can just go to AgentAsk through Microsoft Teams and ask what approvals I have outstanding,” he says. “I can do that in a matter of seconds versus spending a lot of time digging into and logging onto each particular platform.”

Because AgentAsk was the first use of generative AI technology at Toyota, Ballard’s team built a strong partnership with the company’s cybersecurity organization to help it deal with any security considerations. In the end, implementing the technology was fairly seamless and within days, he says, the team was able to integrate AgentAsk with Teams and make it available to other team members.

“The challenge was not so much from the technology perspective but from the people perspective: the organizational change management, mindset, education, and awareness that you now have this solution at your fingertips,” he says.

Spread the word

So far, somewhere between 75% and 80% of the company’s employees have used AgentAsk to solve a problem, says Ballard, whose goal now is for every employee to use the service on a daily basis. The team promotes AgentAsk across Toyota’s digital signage, reminding employees of what the service can do, and every month they send employees gentle nudges telling them what AgentAsk can do for them.

While password resets and account unlocks are a good start, Ballard has grander ideas for the technology, like drafting a comparison of dental policy options for discussion. Employees have already begun intuitively asking questions that span finance, legal, travel and expenses, and more.

Based on several years of experience with generative AI, Ballard offers IT leaders seeking to get their arms around the technology a few pieces of advice.

First off, IT leaders should know the business problem and opportunity before adding another technology product to the stack. By starting small with scope and team size to prove out an initial hypothesis, CIOs can take a “crawl, walk, run approach,” validating the value delivered from the new solution and bringing the organization along.

“Never underestimate organizational change management,” Ballard stresses, adding that IT leaders should work backward from employees and structure deployment of the technology in collections that consider the end-to-end nature of the work being performed, rather than simply pursue whatever use case is presented.

And most importantly, declare a bold bet that’s forward looking. “Always keep the team’s efforts aligned to the objectives,” he says.