We work with a number of public-sector leaders who want technology innovation to be a cornerstone of their public service and spend much of their tenure sprinting after flashy new tech tools. Some have a clear and genuine goal to improve services or efficiency; others might see tech as a way to motivate headlines and enthusiasm. Regardless of motivation, this approach is a trap.

Sending public funds after unproven technology -- and pinning services on tools that still have big unknowns -- translates into costs and reliability issues we don't yet understand. The private sector is already willing to make big bets and take big risks to sort out innovation challenges; government doesn't need to be at the bleeding edge of technology to provide valuable services to constituents. Leaders should instead invest in the skills and organizational flexibility that will let them be fast followers of proven private-sector practices and get smart about how to integrate technologies into efficient, effective, equitable governance.

Being a fast follower rather than an innovator is particularly critical for public digital infrastructure, the database and application layers that agencies use to build and provide services for citizens and the private sector. We aren't arguing that government shouldn't innovate in other areas. If the existential risks of falling behind are extreme (in national defense, for instance, or regulation of cutting-edge technologies) or the private sector doesn't have good incentives to invest in critical development (such as open-ended medical and scientific research), government may be the best party to drive progress.

That should be the exception, not the rule. Yet when it comes to technology in the public sector, the current conversation inevitably focuses on the latest innovation. You can read about applications for artificial intelligence, machine learning and blockchain across the public sector, before some of these technologies have scaled and matured in the private sector. The problem is that governments have a bad track record of investing in shiny objects before the underlying technology has matured.

One Laptop Per Child, for example, was supposed to transform global education, but it was a mistimed investment in a fast-moving space, without the network and application infrastructures that would have made the technology useful. After the 2000 presidential-election recount debacle, states invested heavily in electronic voting machines before security and privacy protocols were ready; cybersecurity experts are now suggesting a wholesale return to paper ballots. And federal law-enforcement agencies were early investors in drone technology -- so early, in fact, that some of the drones they bought couldn't actually fly.

In all of these cases, public-sector decision-makers saw brand-new technology as the answer to a legitimate service need. These ideas might be extremely interesting and promising. But when governments invest in new, untested technology with a limited user base, such as Utah's Google Glass transit app, they risk wasting public resources on services that will never see widespread adoption.

The better approach is to be fast followers of things that work. A great example is the success of digital teams like the U.K.'s Government Digital Service (GDS) in bringing Agile development principles into public-sector technology, culminating decades of work on user-centered design, iterative development cycles and cross-functional scrums in the private sector. GDS wasn't spending time figuring out how Agile worked; those lessons had been learned the hard way by a generation of software engineers and designers. Instead, GDS focused on how to integrate Agile into government. Choosing a proven private-sector practice meant that GDS could spend its time innovating on governance rather than on the methodology itself.

Advocates of government innovation like to point at how countries like Estonia and India are pushing the envelope in public-sector digital services. However, the real story of development in these nations wasn't government inventing technology, but rather top-flight engineering teams that followed private-sector practice very closely and made smart bets about where technology would be at the time of service delivery.

The lesson of these efforts is that being a fast follower will require governments to spend more of their innovation budgets on adoption and governance. That means adapting technology into a public-sector context; it requires insiders like the members of the United States Digital Service who translate technologies into public-sector practices. Ultimately, the goal is to build organization-wide capacity to confront orthodoxy.

When it comes to technology, there are good reasons to want government to move slowly and resist change -- it's how we protect core services and ensure continuity of values. Big or small projects with boring technologies might not garner as many headlines as putting city contracts on blockchain, but it's the right way to ensure that we're good stewards of public dollars. If government can adapt, revamp and re-skill faster by learning from private-sector innovation, it will generate better, cheaper and more sustainable services for all of us.