Pavan Pidugu has a blunt diagnosis for what most often holds federal technology back: a resistance to change, even when agencies have the money and talent to move.
“The biggest problem is the ability to change,” Pidugu told MeriTalk. “Everybody wants change until it is their turn to be changed.”
Pidugu spoke with MeriTalk days before leaving his role as chief digital and information officer at the Department of Transportation (DOT). He was looking back on six years in federal service – first as chief technology officer at the Federal Motor Carrier Safety Administration and later as DOT’s top technology official – while offering a challenge to the leaders who will carry the work forward.
The mission, he said, remains the same: keeping people and goods safe as they move by air, road, rail, and water. What must keep changing is how technology teams support it.
“In the last 18 months, we were probably able to accomplish certain things that have been talked about for 18 years,” Pidugu said. “The important thing is for agencies to continue to keep the momentum and continue to keep delivering at the same pace, if not much faster.”
Building from the workforce up
Asked to assess DOT’s fiscal year 2027 (FY27) technology priorities, Pidugu described a hierarchy as opposed to a collection of separate investments.
Digital government and citizen experience are the goal, he said. Application modernization, cloud-native development, cybersecurity, and robust infrastructure are the means. Workforce transformation is the foundation beneath all of them.
“The workforce of yesterday does not meet the needs of what you have today,” he said. “The goal is the citizen experience. In order to reach that goal, the foundation is workforce transformation.”
Pidugu said public-facing digital services should meet or exceed the standards people encounter in banking, shopping, and other parts of daily life. When government experiences are difficult or fragmented, trust suffers.
He pointed to improvements in DOT’s aviation consumer complaint process as the kind of change citizens can feel directly. He also cited the department’s recent deployment of new surface movement radar at Newark Liberty International Airport – replacing a 30-year-old system – as evidence that long-discussed modernization can move faster.
Those examples should not feel surprising, he said. They should become routine.
Embedding AI in the work
To succeed with artificial intelligence (AI), agencies should define the return they expect, identify the key performance indicators they will track during a proof of concept (POC) or proof of value, and decide in advance what happens if the technology meets those measures, Pidugu said.
Today, agencies are launching disconnected POCs faster than they are putting successful tools into day-to-day operations, he noted.
“Definitely don’t stop with POCs,” he said. “AI has been tested. It adds value if implemented to solve a specific problem.”
At DOT, AI work includes agentic support for some IT service desk workflows. The department is also exploring AI to analyze large volumes of public comments submitted through rulemaking, support call centers, and accelerate software development.
Pidugu set an aggressive target for software development: shrinking an 18-month development cycle to a matter of weeks. He believes AI can help agencies get there if they define the problem and measure the results.
He also wants agencies to look beyond chatbots and use AI inside the workflow itself.
“AI can do a lot more things,” Pidugu said. “It can actually reduce friction in the work processes. It can eliminate the need for people to have 10 clicks. How do you reduce those clicks from 10 to two or one?”
Fixing the foundation beneath AI
Unfortunately, AI projects quickly expose weaknesses in an agency’s data and legacy systems, Pidugu observed.
An agency may want enterprise generative search or real-time decision intelligence, Pidugu said, only to encounter unstructured, siloed, and uncurated legacy data. Inconsistent taxonomies, limited master data management, and decades of incompatible application programming interfaces can stop a promising AI initiative from delivering useful results.
“The focus needs to be on reducing that foundational gap,” he said.
Pidugu called for stronger data governance and common master data management practices before agencies expect AI to synthesize information across the enterprise.
He offered similar advice on quantum computing. Agencies should prepare for it, he said, without skipping the modernization work their current systems still need.
Turning cyber plans into risk reduction
Pidugu framed DOT’s cybersecurity work around three priorities: reducing vulnerabilities in aging systems, using AI to accelerate security processes and continuous monitoring, and preparing for post-quantum risks.
To reduce vulnerabilities, DOT created an internal risk reduction task force to examine legacy infrastructure and systems and push remediation work forward.
“Somebody has to go fix the vulnerability rather than creating a plan,” Pidugu said. “What you have needs to be secure enough to begin.”
To accelerate security operations, Pidugu wants agencies to use AI. He said AI could shorten processes such as securing an authority to operate for a new application from six to nine months to 30 to 45 days. It could also help security teams analyze, assess, and triage new issues through continuous monitoring, he noted.
To prepare for post-quantum cryptography, agencies also need to reduce dependence on aging tools and prepare their encryption before quantum threats become operational, Pidugu said.
Finishing what you start
Pidugu’s approach to moving faster starts with finishing smaller pieces of work.
“If you are able to identify that by doing 10 small things, you would move 10 big mountains, then do those 10 small things first,” he said. “Start and finish. If you start everything and don’t finish anything at all, nothing gets done.”
In his farewell message to DOT employees, Pidugu acknowledged that the department had done some things extremely well and could have handled others better. That is part of working at a pace that is unfamiliar, he said.
His advice was to keep moving, keep removing complexity from systems and processes, and resist the comfort of the status quo.
“You will make some mistakes, but it also tells you that you are actually doing some work,” Pidugu said. “Refusing to accept that status quo and having the courage to say, ‘Let’s try,’ is the only way for that progress to actually happen.”