Artificial Intelligence
CIOs Have Crafted a Winning Recipe for AI Success
Leaders are taking what they learned in 2025 and reshaping the year ahead.
Sydney Scott
Editorial Strategist, AI
Workday
By the time 2025 rolled around, CIOs were like home chefs caught up in a craze of new kitchen gadgets.
AI had changed everything and everyone was rushing to be the first to have it. Leaders stocked their kitchens with every appliance that promised to automate work, convinced that the tools themselves would cook and serve dinner.
But much like a good chef, smart CIOs began to brush off the hype and realize the truth. What mattered most was not the latest gadget. It was how they used those tools and the quality of the recipe—their strategy.
Here is a look at exactly what successful CIOs got right in 2025 and the lessons leaders must carry into the new year.
Report
C-Suite Global AI Indicator Report: AI Is the Ultimate Level-Up
1. Trust First, Speed Second
The first big lesson of 2025 was about human nature.
As AI tools arrived, there was a disconnect between intent and impact. The goal was to boost efficiency, but the sheer speed of the rollout and rapid changes to the workplace created an unintended side effect—anxiety.
Where leaders saw a powerful productivity engine, many employees saw a potential replacement. When employees feel like AI is a black box that might judge them or displace them, they don't experiment. They freeze.
The breakthrough came when leaders stopped mandating usage and started building trust.
Take Accenture as an example. When former CIO Penelope Prett needed to roll out a new career platform to 700,000 employees, she refused to just scrape their data and fill in the blanks for them. She knew that would feel like surveillance.
Instead, she asked employees to co-create their own profiles. She gave them control over their own data. They felt ownership rather than fear, and it became the "single smoothest install" in the company's history.
The same thing happened at Workday. Internal research revealed that only half of employees were using AI tools, often because they were afraid of doing it wrong.
That’s when CIO Rani Johnson stepped in to shift the culture. She partnered with Workday CLO Chris Ernest to launch EverydayAI, an initiative that focused on small, safe wins. The program used global "Promptathons" to let thousands of employees practice in a fun, low-stakes environment.
By allowing space for employees to safely explore, adoption of AI tools at Workday surged from 40% to 85% in just six months.
When employees feel like an AI tool is a black box that might judge them or displace them, they don't experiment. They freeze.
2. Give People What They Actually Need
We often fall into a trap when new technology arrives. We start with the solution and go hunting for a problem. We ask, "How can we use this agent?" instead of asking, "Where is my team struggling?"
To build a pantry of agentic workflows that actually matters, we have to ignore the technology at first. We have to look at the friction.
Let's look back at the manufacturing giant Jabil. Their problem wasn't a lack of computing power, but a lack of continuity. They noticed that as their most experienced engineers retired, decades of knowledge went walking out the door too.
They didn't solve this by buying a generic AI tool. They built an agent specifically designed to interview senior engineers and bank their wisdom into a chatbot, allowing junior engineers to query that knowledge bank long after the retirement parties have ended.
Other times, it’s simply about finding ways to make the human experience better.
At Seattle Children's Hospital, instead of relying on system logs to pinpoint issues, CIO Dr. Zafar Chaudry walked the floor and talked to patients.
When a 12-year-old bluntly told him the hospital was boring, Chaudry didn't upgrade the bandwidth. He partnered with Microsoft to build a digital twin of the hospital inside Minecraft, allowing kids to virtually explore the facility and play with one another from their beds.
This approach had tangible results. One patient even credited the immersive gaming as a critical factor in surviving post-operative pain.
Leaders and employees should bring this same level of curiosity to their own orgs. They can watch the finance team during the quarterly close or observe a customer support rep navigating their monitors to see exactly where the roadblocks occur during a client call.
In this way, we can pinpoint where the recipe falls apart. We’re looking for the cracks in the digital pavement we've called "admin" for decades. When we identify real problems, we stop buying technology and start building solutions.
We’re looking for the moment a human has to physically turn from one screen to another to re-enter data they just typed; or export a CSV file from one dashboard just to upload it into another. We’ve called these cracks in the digital pavement "admin" for decades.
When we identify real problems we can find the right ingredient to get the desired outcome. We stop buying technology and start building solutions.
3. Always Include a Human in the Process
In 2025, leaders learned a simple truth. AI agents are not just software, and like human employees, they need a manager.
The IT team cannot just install agents and walk away. Organizations have to recruit them, train them, and track their work. Successful companies looked at central systems to handle this, acting like an HR department for their bots, tracking every agent so the company knows exactly who did what.
At Seattle Children’s Hospital, Dr. Chaudry did not just turn on his new Pathway Assistant agent—a generative AI system that synthesizes clinical guidelines to help doctors make faster, evidence-based decisions. First, he knew it had to be safe. So he had 50 doctors spend hundreds of hours testing it first. They made sure every answer it gave was medically perfect before a real patient was involved. The agent acts as a smart helper, but the doctors are still the experts.
We see this in business, too. Payroll agents do the math, but do not change the paychecks on their own. Agents draft the work and ask a human to review it. The human spots errors, edits the draft, and hits "approve." It allows digital teammates to handle the rough drafts while humans make the final call.
Demanding immediate returns on early experiments crushes innovation before it can breathe.
4. Rethink How to Think About ROI
There’s a question many CIOs run into when launching new programs: "What is the ROI?"
But Toby Stuart, chair of Workday’s AI Advisory Board, offers a radical piece of advice: "Remove ROI from the conversation."
During a conversation on the Future of Work podcast, Stuart argued that demanding immediate returns on early experiments crushes innovation before it can breathe. Instead of promising a specific dollar amount, CIOs need to change the currency of the conversation.
Sure, there’s value in ROI as it’s defined now when looking at small-scale projects, but companies can miss opportunities if it’s their sole focus. “This is the way the world is migrating. Your job is to build, and prepare the company and the product it makes for the future. That's an AI future.”
If it’s too soon to measure monetary ROI, measure momentum.
When the manufacturing giant Jabil rolled out their AI initiatives, they didn't just look at the bottom line. They realized that in a fast-moving market, speed is currency. They tracked how fast they could go from a whiteboard idea to a working pilot. They also tracked how quickly that idea spread to their 100+ sites around the world. They changed the question from "How much did this save?" to "How fast is this moving?"
But a quick caveat: You can’t optimize for speed if you haven’t solved for trust. Momentum hits a wall when people are afraid. And sometimes, the best way to fund the future is to point out the danger of the past.
In legacy-heavy industries, like the public sector, efficiency isn't always a strong enough argument to unlock budget. But safety and security might be. One public sector CIO faced a wall when trying to upgrade ancient systems. The team pivoted from arguing for productivity to arguing for security, pointing out the risk these ancient systems posed. That reframing worked.
The lesson is clear. You can’t build a next-generation company using last generation’s metrics. Rethink your ROI and invest in the areas where AI is scaling and delivering real impact.
5. Fix Old Risks Before Moving Forward
We all have that one piece of infrastructure we’re afraid to touch—the messy data lake we ignore because we’re terrified that fixing it will break the business. We convince ourselves that if it isn't broken, don't fix it. But often, the risk is a ghost story we tell ourselves to avoid the hard work.
During the height of the COVID-19 pandemic, Dr. Chaudry faced a nightmare scenario: He had to migrate the hospital's core clinical and revenue systems to Epic—a massive undertaking that usually requires armies of on-site consultants.
The traditional risk assessment would have screamed "Stop." In-person training was impossible due to lockdown protocols. The safe choice would have been to delay the modernization and stick with the old, familiar systems until the world went back to normal.
Dr. Chaudry didn’t wait. He realized the greater risk was stagnation.
His team pivoted to a fully virtual training program for 13,000 staff members. They ignored the conventional wisdom that said clinical training requires a hand on the shoulder. The result? Despite the perceived danger of a remote deployment, the staff achieved a 92% first-time pass rate on their proficiency exams.
They proved that high-risk modernization could succeed even under the most extreme constraints.
Whether it is moving to the cloud or cleaning up data taxonomy, the perceived risk of action is almost always higher than the actual risk. The real danger lies in doing nothing.
We’re no longer just managing technology. We’re leading a new kind of workforce.
The Beta Test Is Over
We spent the last year learning the rules of the road. We found that safety actually makes us faster. We learned that digital workers need human bosses. And we proved that when we make people feel safe, they will try amazing things.
These are not just nice stories. They are the building blocks for 2026.
Next year, the focus shifts. We are done asking, "How does this work?" Now we ask, "How do we scale?"
We’ve built trust. We’ve set the guardrails. We’ve trained our people. Now we can stop playing defense and start playing offense.
This means moving from small experiments to big actions. It means trusting the digital team to handle the heavy lifting so CIOs can attack the problems that have been stuck on their to-do list for years.
Today’s CIO is no longer just managing technology. They’re leading a new kind of workforce.
The foundation is set. The team is hired. Now, go let them do their best work.
Report
C-Suite Global AI Indicator Report: AI Is the Ultimate Level-Up
AI has been a gamechanger across nearly every industry and marketplace. But realizing the true value of AI requires buy-in between different stakeholders. In this AI Masterclass, learn how you can make the most of AI for finance, HR, and more.