With tech leaders under pressure to build successful AI programs while delivering ROI, there are lessons for enterprise CIOs within the lean, cost-conscious innovation work happening in the public sector.
Dru Rai, chief information officer for the State of New York, underwent a mindset shift when he left the corporate world for the public sector in 2023. He spent his career at Quaker Houghton Company, the Ball Corporation, DuPont Coatings, GE Advanced Material and EY.
Like the private sector, his innovation team is aiming to use AI to save time and resources, but instead of aiming to improve a company’s bottom line, the focus is on state residents, which underscores the need to ensure the tools and data they use are safe.
“Our policy is to make sure we serve people,” Rai said. “We do believe [AI] is a productivity tool — we can juice out more taxpayer dollars to get more work faster, better — but we cannot sacrifice the health, the safety and the privacy of folks.”
Enterprise AI budgets have steadily risen for the last several years despite economic headwinds, but C-suite executives have faced more scrutiny on the cost and return for AI projects in 2026.
Earlier this year, many enterprises encouraged employees to use as much AI as possible to encourage innovation and experimentation, and prove value to boards. But as AI providers shift their pricing models to usage- or outcome-focused structures, and enterprises feel the real cost of a “per-token” model, those incentives faded at many companies.
Enterprises could begin thinking about their IT spend in the same way government workers have always had to, Rai said. His team is forced to rely on less money and deliver more.
“When you have less money, you find more creative ways to utilize it,” he said.
One of the strategies Rai brought from private sector work into his work for the state was a culture and process to “try and fail fast in a highly controlled environment." He applies this thinking to AI pilots, saying that his team needs to recognize early when a pilot may have issues that keep it from scaling so that they can cut their losses, rather than keep investing in it.
“That has built us more resilience to make sure that our dollars are just going in the right place because I can't afford to make giant mistakes,” Rai said.
Timing innovation work
Although it may seem against many CIO’s instincts, Rai said it could be beneficial to stop thinking that teams need to achieve results in the fastest way possible. Budget-conscious decisions may mean choosing an AI model that is less expensive, and works slightly slower, but will ultimately help an employee arrive at the same conclusion or output.
Being first at something isn’t his team’s priority, Rai said. The focus, instead, is on finding a sustainable solution. “I’m happy to be second,” he said.
The CIO role is changing for the public and private sector, Rai said. CIOs are now expected to predict problems and find solutions for them, guiding teams through technical transformation. Learning how to best use AI has been similar to grasping other tools he’s encountered across his 30-year career, calling it “an event in the evolution” of technology.
Rai said he feels future IT leaders will need to navigate the continued reliance on hardware like GPUs, and in the future, quantum computing, to constantly update their tech stacks.
Tech leaders will likely need some resilience to weather changing expectations of their budgets and their roles, but a civil servant mindset and good tech partners can help make it work, Rai said.
“We're never going to have enough,” Rai said of his AI budget. “So you just have to figure it out, change your priority, and find a way.”
