Allstate is best known for being the “good hands” company, founded in 1931 to insure cars and property. But like other companies of a certain age, it needs to venture into new areas.
That’s what the insurance giant—number 84 on this year’s Fortune 500 list with $36.5 billion in revenue—has done over the past few years, starting up its CompoZed Labs unit to speed up development of new software services. After all, Allstate (ALL) now has to compete not only with traditional insurance rivals, but a raft of startups aiming to change the industry just as Uber and Airbnb have shaken up the transportation and lodging markets, respectively.
Launched in 2014, CompoZed Labs now handles 40% of Allstate’s software development, according to Doug Safford, Allstate’s vice president of technology innovation. The unit uses Pivotal Cloud Foundry, a software workbench for building applications that can run on a company’s internal servers, or on third-party public cloud infrastructure from the likes of Amazon, Microsoft, or Google. Pivotal, a tech company backed by investors including [hotlink]Dell Technologies[/hotlink], General Electric @general(GE), and Ford (F) offers a commercial version of Cloud Foundry, which is free, open-source software.
Related: Pivotal Cloud Foundry Isn’t Just for New Apps Anymore
Allstate developers also use Jenkins, a tool for incorporating software changes continuously; Git to track those changes; and Atlassian Jira (TEAM) to manage projects.
That toolset, Safford says, gives development teams what they need to test new ideas quickly, and if they work, get the software out to users. This is in stark contrast to the traditional, monolithic way of building business software in which a new product would take months—or even years—to build and deploy.
“When you’re an 86-year-old company things have done a certain way, there are rules in place because of what someone did 10 years ago,” says Opal Perry, vice president and divisional chief information officer of claims for Allstate. “Now, instead of a 200-person team, you have small six- and eight-person teams working on things. It unleashes creativity.”
In the past, many enterprise software projects required millions of dollars, and even sign-off by the CEO. “By the time you got permission, ideas died,” explains Perry. “Now a senior manager has $50,000 or $100,000 to do a minimum viable product.”
Related: This Benefit of Cloud Migration Might Surprise You
Minimum viable product, or MVP, is another term for a software prototype that can be put out for customers to try. If they like it, all systems are go. If not, it gets killed.
Empowering developers has already paid off, Safford says. A few years back, two developers approached him with an idea to streamline Allstate’s roadside assistance program, the second largest in the country after AAA. With 25,000 calls on peak days, shaving a few minutes off of each would save Allstate money and improve customer service.
“So they went to the executive vice president in August, and by September, they were building an app,” Safford says. “By January, it was in the app store.”
Not only does the app cut the time it takes to get assistance to drivers, Perry says, but the “Uber-like” experience means the dispatcher knows where the caller is, and where the truck is.
Get Data Sheet, Fortune’s technology newsletter.
Use of Cloud Foundry, in particular, is both tactical and strategic. Tactical in that it fosters fast development, testing, and deployment. Strategic in that software built with Cloud Foundry can run on internal servers or elsewhere as needed.
“What we like about Cloud Foundry is it facilitates moving things around now or in the future,” Safford explains, echoing a familiar theme among corporate IT professionals who do not want to be tied to any one cloud provider. “We want to move to the cheapest supplier at the right time, and the only way to do that is to be agnostic.”
Related: Allstate Spins Out Startup Focused on Flagging Driver Risks
All companies have to be prepared for ever faster and relentless change and should use software that lets them adapt fast, Safford says.
“You have to be willing to disrupt yourself,” Safford says. “Three years ago, people were saying autonomous cars were 15 years out. Then they said 10 years out. Now GM has announced 120 Bolts that are completely autonomous.”
The lesson? No company can wait around for the future to happen because it’ll happen sooner than anyone thinks.
Note: (June 19, 2017 1:11 p.m.) This story was updated to add that Pivotal is a tech company backed by several big-name investors.