CIOs reimagine business processes to reap AI benefits

Every business process reflects the constraints that existed when it was first devised, says IT exec Maria Cardow.

Those constraints, Cardow explains, typically derived from technology limitations at the time of the process’s initial implementation. As a result, many business processes to this day involve workflows that still require manual actions or cumbersome jumps between multiple computer applications to accomplish essential tasks.

But artificial intelligence and other modern technologies can knock down those old constraints, “which is why it’s incredibly important to reassess processes” before attempting to automate what may likely be an ineffective, outdated process, says Cardow, CIO of managed security services provider LevelBlue.

“You have to check the assumptions in processes before you kick off a transformation if you want to allow for greater innovation,” she emphasizes.

Executives are feeling the pressure to transform or optimize their organizational workflows with AI. J.P. Morgan’s 2026 Business Leader Outlook found process automation to be the most common AI application that midsize business use or plan to use, cited by 62% of executives surveyed. And EY’s CEO Outlook 2026 found that 43% of CEOs identified optimizing operations and improving productivity as a transformation priority, making it the top cited priority, with enhancing product and process innovation coming in at No. 3.

Cardow is in the thick of leading such efforts, and she’s finding that IT is well positioned to do the work.

For example, when examining business workflows, her IT team often finds that processes span “at least a dozen different systems and that the number of assumptions baked into workflows were dependencies on technologies two generations back,” she says.

Such instances show why CIOs should lead their organizations through process optimization before automating with AI, Cardow says, citing the longstanding admonition against automating bad processes.

A natural fit for optimization

Anytime a technology is up for license renewal, LevelBlue endeavors to review any business process that depends on it, Cardow says. The goal is to determine whether the process is ripe for optimization and transformation before renewing or replacing the existing technology — and before applying further automation or AI.

That timing means Cardow, as CIO, takes a lead role in optimizing enterprise processes. Moreover, she brings more in-depth knowledge than her business colleagues of what technology capabilities can optimize and transform a process. And as CIO, she has a remit to improve productivity and efficiency while driving down costs.

The CIO is a natural fit for such work, Cardow adds, pointing out that CIOs have had responsibility for optimization for most of the role’s existence.

IT also brings objectivity to process optimization, she adds. Workers are often comfortable with the status quo and the tools they’re using, whereas CIOs and IT teams don’t have such attachments.

“Being able to critically challenge how processes are executed is absolutely essential to my role now,” she says. “[The business] will still say, ‘This is how it gets done,’ but CIOs are well positioned to ask about baked-in assumptions and whether certain steps are still appropriate as they partner [with the business] to streamline processes and develop solutions.”

AI heightens the significance of process optimization

As Cardow notes, CIOs have been involved in process optimization since the role’s inception, implementing technologies to automate tasks. But the CIO’s role in process optimization has taken on heightened significance in the AI era.

“It’s vastly different now,” says Doug Gilbert, CIO and chief digital officer at Sutherland, a digital transformation services company. “There’s been an evolutionary shift.”

Gilbert says CIOs focused on improving individual tasks during prior waves of automation, such as when robotic process automation (RPA) was first rolled out. Now CIOs must optimize entire processes and workflows to transform larger swaths of their organization’s operations.

“If you’re looking at AI to be a more complex RPA to solve minute tasks and activities, you’re going to fail,” he says. “You’ve got to look at how humans work across systems and across tasks; automation today has to go across an entire process flow.”

He cites as case in point his company’s Insurance AI Hub, a project launched and completed in 2025.

“As CIO, I was responsible for the overall technology direction and architecture. This meant overseeing the implementation of a full ecosystem of domain-trained AI agents for underwriting, claims adjudication, and policyholder servicing across multiple lines of business, while ensuring clean master data foundations, embedded governance, observability, and human-in-the-loop controls were built in from the start,” he says. “The result is production-scale agentic AI that delivers up to 30% faster claims cycle times, lower leakage, higher satisfaction, and stronger compliance.”

Gilbert cites as a second example his company’s agentic AI platform for healthcare provider credentialing, saying this multiagent system “reimagines a highly complex, regulated, cross-functional process at the macro level — automating and contextualizing information across many previously siloed processes, internal systems, and external sources.”

These agents “work in tight coordination to tackle a much larger, macro-level complex process — sharing contextualized data in real-time, maintaining quality and compliance throughout, and enabling a far more comprehensive level of automation and problem-solving across the entire end-to-end credentialing workflow,” he says. “This turns what used to be a slow, manual, multiweek process into a fast, accurate, governed operation.”

A point of inflection for CIOs

IT execs say process optimization today requires more from CIOs than in the past. CIOs must now understand how work happens within the organization in more detail, as well as how those processes sit within broader workflows, how processes and workflows arrive at decisions, and what defines a good or accurate decision.

“Agentic AI — the ability for LLMs to actually take action — is raising the bar significantly,” Gilbert says. “When designed properly, these systems don’t just follow rules; they plan, decide, act across systems, and learn. That means CIOs can no longer optimize in silos or bolt intelligence onto broken processes. We must now lead the complete reimagination of workflows so they are natively designed for autonomous agents while staying inside clear guardrails.”

Those are more complex — and more significant — requirements than CIOs have faced to date, Gilbert adds.

“In the past we would receive a process from the business, map it, automate the obvious parts with RPA or traditional tools, and hand it back. Today, because AI brings reasoning, contextualization, and the ability to work across systems, the CIO must lead the charge in fundamentally reimagining how work gets done,” Gilbert says.

“We now own end-to-end process intelligence, ensuring the underlying data is clean and contextualized, governance is embedded, and the optimized process actually delivers trustworthy business outcomes,” he adds. “The CIO must sit at the strategy table and, in large part, drive it going forward.”

Organizations across industries are indeed turning to their CIOs to reimagine workflows, says Catherine Malkova, senior vice president of Kyndryl Consult and Practices US.

“There is a focus on business workflow, not just working with old processes and making them more efficient but instead remaking them from the ground up into an AI agentic workflow. It’s about building an AI-native enterprise,” she says, noting that “the ‘agentification’ of workflows is what will drive the next wave of AI adoption.”

Kyndryl’s 2025 Readiness Report highlights the amount of transformation executives are expecting, with 87% of the 3,700 business leaders across 21 countries saying that AI will completely transform job roles and responsibilities at their organizations this year.

Challenges abound

Expectations of transformation are high, but so too are the challenges that CIOs and their C-suite colleagues face.

Kyndryl’s research identifies “five specific readiness challenges ahead of them: getting solid tech foundations, managing global data, evolving workforces, pressure to scale AI pilots, and aligning leadership.”

Rigid workflows, fragmented systems, lack of trust in AI, lack of required skills, and resistance to change also present challenges to CIOs, their colleagues, and their teams as they seek to optimize and transform processes and workflows, Kyndryl’s Malkova adds.

Others list the lack of good documentation for many processes as another barrier, along with concerns about data security, privacy, and AI hallucinations.

Dom Profico, who as CTO of consultancy Bridgenext advises CIOs, says enterprise ambition itself can be problematic, as executives and teams may rush to adopt AI without putting in the work required to transform the processes AI is meant to help optimize.

“AI makes it so much easier to do that automation, so you have a higher risk of automating a bad process. And there’s a lot of pressure in the age of AI. Everyone feels like they’re behind. That adds to the likelihood and risk of going too fast,” he says.

Profico stresses that while automating existing processes may create some efficiencies, it may not lead to the optimization and transformation that delivers the significant returns that CEOs and boards now want from AI investments.

CIO as chief instigator

CIOs must also bring rigor to the discipline of process optimization “to make sure the teams are automating the right things and using the technology in the right way to do that,” Profico says.

He lists systems thinking as a must-have skill for CIOs, as it allows them to take a holistic problem-solving approach where they see how the different parts interact within a larger whole.

That, he notes, requires CIOs to deepen and broaden their knowledge of the business, how work gets done, and what adds value.

Sutherland’s Gilbert agrees, saying needed skills include deep fluency in process intelligence platforms and mining tools; strong data strategy expertise (especially master data management, lineage, and context engineering); AI governance and risk management (particularly around autonomous agents and accountability); advanced change leadership (i.e. designing new operating models where humans and agents collaborate effectively); and business translation skills — “the ability to link technical decisions directly to P&L and customer outcomes.”

“Technical depth is still essential, but the real differentiator today is the ability to think systemically about human-plus-agent workflows and to lead at the intersection of strategy, technology, and risk,” Gilbert adds.

Merim Becirovic, CIO, managing director, and partner at Boston Consulting Group, has a similar take on the process optimization work that CIOs and their teams are leading.

Like others, Becirovic says CIOs in the past focused on optimizing processes within siloes. “But what we’re doing now is not about sending people across the different systems. It’s about transforming the journey and the experiences within that journey. It’s about connecting the dots and connecting the systems that create those experiences and that journey,” he says.

That requires CIOs to focus on how processes and workflows connect to deliver outcomes.

“That’s a natural evolution for me as a CIO,” he adds. “Technology gives you a lens into that work and gives you access to products that allow you to break siloes down.”

Becirovic points to his IT department’s delivery of Deckster, a homegrown AI tool for slide deck creation, as an example of successful process optimization and transformation.

BCG workers create more than 30 million slides annually, making it a process that takes up significant work time. Rather than focus on incremental gains by using generative AI for specific tasks, such as crafting the language on individual slides, BCG teams reimagined the entire slide-creation lifecycle. Using OpenAI’s API alongside a curated library of BCG templates, Deckster produces fully formatted, client-ready slides in just three seconds instead of the 15 minutes it had taken.

BCG is now bringing agentic AI capabilities to Deckster to further optimize and transform the process.

“Those are the opportunities where CIOs can step in and say, ‘Is there a better way to work?’” Becirovic says. “I really think CIOs today and tomorrow need to be far more aggressive in looking for change and driving change. The CIO needs to be the instigator, influencer, enabler, collaborator, and the innovator to make that happen.”