IT isn’t holding AI back, your business processes are

Most CIOs and other IT leaders are confident in their teams’ ability to meet the coming AI challenges, but many believe business operating models and processes need an overhaul.

More than 80% of senior IT executives surveyed for the 2026 Global Leadership Technology Study from Deloitte are confident in their organizations’ ability to deploy and govern AI capabilities at scale. However, 75% think their operating models and processes must change in the next 12 to 18 months to drive greater value.

The survey results don’t signal an overconfidence from IT leaders in their teams’ ability to roll out AI, and they don’t seem to suggest a major AI choke point within the IT organization, says Anjali Shaikh, MD of Deloitte Consulting and leader of the Deloitte global CIO and US tech executive programs.

They do, however, point to a need for major changes to operating models and processes across the organization, not just within IT, to achieve better AI results, she says. And while CIOs can’t control how business workflows are designed, they have a role to play in pushing for processes that better make use of AI tools, she adds.

“What we’re hearing now is that AI isn’t just exposing those technology limitations, but enterprise design limitations,” Shaikh says. “What’s new is that technology is no longer the bottleneck, it’s the operating model and the ways of working.”

Business workflows that need to change are the manual processes that aren’t friendly to AI integration, like using spreadsheets for some accounting, or manual data entry.

CIO as advocate

CIOs can use their expertise to champion the redesign of enterprise workflows so they better align with AI tools like agents, Shaikh adds, creating opportunities for CIOs in a changing landscape.

 “The AI era is starting to demand a new type of leader,” Shaikh says. “They’re going to have to guide the organization through this change, build AI-related teams, and figure out the skill set required for not only their team, but the overall business.”

Another conclusion from the Deloitte report is that CIOs and other IT leaders are experiencing a make-or-break moment as they face major new expectations in their roles, including the ability to lead change and build AI-ready teams.

Leadership is important, Shaikh says, in that the organizations with the most successful AI deployments tend to be those with C-suite support for operating model and process change.

“It’s an investment and a commitment from the top down —  the board, CEO, and C-suite have made a commitment to saying we’re going to invest not only time, but budget, resources, and talent into helping us fundamentally shift the way we work,” she says. “When you get that alignment and commitment from the top, that sends a signal to everyone in the organization.”

Several IT leaders agree that several operating model and process changes are needed at many organizations.

Focus on the right thing

Many organizations are embracing AI by buying ChatGPT licenses and hosting lunch-and-learn sessions on prompt engineering, but they’re solving the wrong problem, says Peter-Paul Schreuder, chief cloud officer and VP of support at enterprise asset management software vendor Ultimo.

“Teaching people to use ChatGPT takes about an hour,” he says. “Teaching an organization to fundamentally rethink how work gets done takes four to five months of hard, unglamorous process work, but delivers exponentially more value.”

A major problem is that most organizations don’t understand their own workflows well enough to determine where AI could add value, he adds.

“They don’t know where the information gets stuck,” Schreuder says. “Which tasks consume disproportionate time? Where do we repeatedly reinvent the wheel because knowledge lives in someone’s head? Until you can answer those questions with specificity, no amount of AI training will matter.”

Getting the most out of AI takes a huge reset in the way businesses think how they complete their work, he adds, and if AI initiatives lack deep involvement from business operations leaders, they’re doomed to fail.

“If you can’t diagram how your work flows from intake to completion, you’re not ready for AI augmentation,” Schreuder adds. “If your list includes ‘summarize documents’ or ‘draft emails,’ you’re thinking about AI as a fancy word processor. Transformative applications are process specific.”

So some companies are clear about how they link new AI tools with workflow redesign. When financial services firm IMA Financial Group deploys a new AI capability, the company doesn’t treat it like a technology rollout, but as a transformation for how work gets done, says Megan Cullen-Meyer, its VP of data and AI. The company examines what tasks need to change, what roles need to be realigned, and what training and reskilling is needed, for instance.

With AI, IMA is automating routine, transactional work that used to burn up a lot of human time and energy.

“We talk a lot about how applying AI to old, outdated processes isn’t how we’re going to get the value we want from AI,” she says. “We’re taking a hard look at our core tasks and workflows to determine where we automate the work, where we augment work, and where we keep humans at the center.”

Setting the right foundation

The Deloitte report suggests that the IT team can’t build AI tools in a vacuum, says Luis Mesas, VP of product engineering at IT services provider Sngular. The value of AI comes when the organization develops a good governance framework to start with, can keep learning from what’s happening in its workflows, and can adjust the system without turning every improvement into a new initiative, he adds.

“This requires IT to work much closer with the business problem from the beginning,” he says. “Engineering teams can’t build AI systems in isolation and expect adoption to follow later, because they need to understand the process, the quality of the data behind it, and the risk of getting the answer wrong before they can design something that’ll hold up in day-to-day use.”