The federal CIO job is changing. For Adam Starr, chief information officer at the Office of Personnel Management (OPM), technology leadership is increasingly moving out of the back office and into the center of agency strategy.
“The CIO seat has become a strategy function. Every business is becoming a technology business, and government’s not any different,” Starr told MeriTalk in a recent interview. “The job is designing what a digital government can look like and how we can better serve our constituents.”
A year into government, Starr said one thing has surprised him most: the speed.
His biggest concern coming in from the private sector was that bureaucracy would prevent his team from moving quickly. Instead, he said OPM has been “delivering value to constituents at an incredible pace.”
“I continue to be blown away by the pace that we can move in government,” Starr said. “I think that will just continue to accelerate.”
Start with the mission, not the technology
Asked to rank his agency’s top priorities for fiscal year (FY) 2027, Starr put digital government and citizen experience first, followed by application modernization and artificial intelligence (AI) and machine learning.
But he views the No. 1 ranking differently from the others.
“For me, it’s not just the priority; it’s the mission,” Starr said. “How do we have better services at scale through technology?”
His team is working toward five goals. Two center on how OPM operates: rewiring the agency as a tech-first organization and creating a world-class developer experience. The other three focus on delivering world-class experiences across federal human resources (HR), retirement, and health benefits.
OPM is already applying that approach to major modernization efforts. The agency has moved the federal retirement application process online, replacing a largely paper-driven process with a digital system.
Starr also pointed to OPM’s push to consolidate the government’s HR environment into a single, government-wide human capital management system.
“We’re going from 119 different HR IT platforms in government to one,” he said. “That allows us to rethink shared services, move faster, be more responsive to our employee base, and be more efficient for the taxpayer.”
For Starr, the goal is to build an organization that can keep improving services at scale. Doing that, he said, depends on having the technical talent to support it.
Creating a tech talent flywheel
Starr sees maintaining “talent momentum” as one of his biggest challenges in FY27.
OPM recently launched and expanded the US Tech Force, a government-wide effort designed to recruit technologists into public service and embed them in agencies working on modernization priorities.
For Starr, the talent strategy operates as a flywheel: bring strong engineers into government, give them visible opportunities to make a difference, use those successes to attract more talent, and then deliver more.
“The biggest fear that I hear from engineers is the same fear that I had – that they’d get stuck in bureaucracy, that they wouldn’t be able to actually use their talents to have impact,” he said.
Starr said his first year at OPM has shown him otherwise. He sees government offering technologists both the pace to get meaningful work done and the chance to have an impact at a scale that is difficult to match in the private sector.
“These are all mission-oriented people that want to have impact, and they can in government,” he said.
That makes the employee experience itself part of the recruiting strategy. Starr wants Tech Force participants to see that they can move quickly and do consequential work, then carry that message back to other engineers.
AI’s biggest constraint: time
That same focus on people sits at the center of Starr’s AI strategy.
In OPM’s most recent quarterly pulse survey, employees said the biggest barrier to using AI more was not access to the technology, but time.
“AI adoption is broad,” Starr said, “But we only really see high value where people have made the time to rethink their work.”
OPM has responded with AI office hours, an ambassador program, and bonuses for employees finding useful applications for the technology. But Starr said those efforts only go so far. Some teams need hands-on technical help to get over what he calls the “activation energy” required to change established workflows.
That realization led to OPM’s AI Catalyst program, which embeds engineers directly within business functions, including retirement services, finance, and legal teams.
Rather than arriving with a predetermined AI use case, the engineers start by asking questions: Walk me through your day. How do you do this task now? Where are the rough edges?
The goal is to reach what Starr described as an “aha moment” that shows employees what is possible. But the model depends as much on business ownership as technical expertise.
“You can’t just parachute an engineer into these teams and expect transformation,” Starr said. Business leaders must commit someone to work alongside the engineer, explain the process, and co-create a different way of working.
Starr sees that adoption strategy working across two levels – what he calls “little OPM” and “big OPM.”
At “little OPM,” he said the agency is testing how its own workforce can work differently with AI.
At “big OPM,” the agency is taking that work governmentwide. Starr pointed to OPM’s AI Gap program, which aims to help other agencies manage the human capital changes around AI.
“We are critical to the ability for government to rethink how their employees engage with AI,” Starr said, describing the shift as “the biggest workforce transformation of the past century, at least.”
Build security into the team
Starr is applying a similar embedded model to cybersecurity.
“Security is everyone’s job, and it’s job number zero,” he said.
Rather than relying on a centralized cybersecurity team to review systems near the end of development, OPM is embedding security expertise into engineering teams so that security is addressed throughout the development process.
“You can’t just throw something over and expect it to be secured by a different team,” Starr said. “Everyone owns the risk and the business outcome.”
The distinction matters because, in Starr’s view, “security is a risk-management exercise.”
That means measuring what actually changes risk – not simply tracking activity.
Starr said agencies can spend significant time on what he called “security theater,” including checklists and policies driven by outdated compliance requirements. He is pushing his team instead to identify a small number of executive-level metrics tied to decisions and outcomes.
If a metric changes, he said leaders should know what impact to expect and what action must follow.
Make room for measured risk
Technology modernization will also require government to change how it thinks about failure, Starr said.
“I think government has to be better at embracing measured risk,” he said.
Federal employees have strong incentives to avoid becoming associated with a project that fails, Starr explained. The unintended consequence is that leaving a broken process untouched can feel safer than trying something ambitious and risking an unsuccessful outcome.
“If you’re constantly trying to optimize for lack of failure, you’re not going to take the risks that you need in order to innovate and wow the customer,” he said.
To combat this, OPM developed a risk appetite statement to make clearer where the agency is willing to experiment and where it is not.
The larger goal is to remove friction around talented people.
Starr said he would also like to see changes to the General Schedule that better recognize knowledge and engineering work and allow government to reward exceptional technical talent without tying advancement so closely to time in government.
“There’s incredible talent in government that’s eager to drive progress,” Starr said. “I view my job as clearing the path for them.”