Home Platforms News Jim Cavellier: How To Build a ...
CIO Bulletin,
10 September, 2026
Author:
Guest
Most enterprise transformations stall not because the technology fails, but because business teams and IT groups work in separate worlds. At Cass, Chief Information Officer Jim Cavellier took a different route by tearing down the standard handoff process between departments. Rather than treating technical upgrades as isolated projects, the company structured its operations around shared ownership from the start. That cultural shift has turned technology into a shared discipline across the business, allowing teams to adopt new tools and ship products at a much faster pace.
In most large companies, business units write long lists of requirements and hand them over to engineers, only to wait months to see what comes back. Cass replaced that handoff model with a system where both sides share equal responsibility for project outcomes from the very first meeting. “With Cass, it means shared ownership across the entire organization,” Cavellier explains. “In other words, it’s not IT delivering solutions to the business, or the business waiting on IT. It is both organizations acting as one team and One Cass, with both sides committed and accountable for the outcome from day one.”
To make this co-ownership work, day to day, the company pairs technical leaders with business leaders on every project. This operating model ensures that neither side works in a silo nor loses track of what the other needs to succeed. “This model is built into everything we do. We call it our ‘2InaBox’ approach to business and technology innovation,” Cavellier says. “Every initiative gets a Technology Product Owner (TPO), a technology expert who is savvy with the business, and a Business Project Owner (BPO), someone who is a business expert, but also savvy in technology. The TPO and BPO are paired up and ‘joined at the hip’ for delivering against a common purpose and goal.”
Moving away from traditional project management was not simple, largely because teams were used to working behind departmental walls. For years, business leads wrote specifications and waited for delivery, while technical teams worked with full control over the code. “Like any other cultural change, the toughest part is changing muscle memory, meaning changing the way things have been done for a long time,” Cavellier notes. “Moving to joint definition, joint development using AI, and joint accountability for achieving business outcomes was different from the prior model, where business leaders handed over requirements (aka BRD – business requirements document) to IT and then waited. IT, on the other hand, was used to having sole ownership of delivery.”
Instead of passing documents back and forth, both sides now sit down together to build working software in real time using AI tools like Claude Design. This direct collaboration lets staff spot problems early and make adjustments before writing final production code. “Our 2InaBox model asks both sides to give something up,” Cavellier says. “We no longer write requirements that are thrown over the wall to IT, but rather business and IT subject matter experts sit down together in front of an AI platform (i.e., Claude Design) and build real-time working prototypes. IT then takes those working prototypes and ‘product-ionizes’ them.”
Transformation efforts often run out of steam when employees feel disconnected from the goals set by leadership. When changes are forced from the top down without input from staff, skepticism tends to grow quickly. “People stop trusting and start questioning transformation when it feels like something that is being done ‘to them,’ instead of ‘with them,’” Cavellier explains.
To avoid that pitfall, teams work together early to define the main goals and clarify what a successful rollout should look like. Cass also stepped away from long project timelines, choosing instead to ship smaller pieces of software at a steady pace. “Right out of the gate, we collaborate on business questions such as ‘What is the big-picture opportunity we have in front of us?’ and ‘What Success Looks Like’ (WSLL) in the end,” Cavellier says. “These are co-designed upfront so there is a shared vision throughout the development process. We also gave up doing large multi-year projects years ago in favor of a Minimum Viable Product (MVP) approach, meaning delivering smaller, incremental value at a fast pace that enables early recognition of business value.”
Delivering smaller pieces of work helped the company build trust through visible results rather than internal memos. Staff members saw the benefits firsthand during early pilots, which encouraged other teams to adopt the new process. “Everything we build follows an MVP approach now. This builds momentum, which results in early and quick wins,” Cavellier points out. “This change didn’t come about by sending out emails or publishing a new policy. Rather, we proved it in small pieces first. It was our Cass team members who did the heavy lifting. They showed up for the first few pilots, willing to work differently before there was any proof that it would pay off.”
Once the first few teams cut delivery times down from over a year to just a few months, enthusiasm spread naturally across the business. Seeing tangible proof on real projects turned skeptical staff into active supporters of the new workflow. “Once a couple of teams saw what used to be a year-plus project being done now in 3-6 months, they became evangelists for the rest of the organization,” Cavellier says. “This is the kind of team that leads and enables organizational transformation.”
Many business leaders focus on the technical side of AI. Finance and operations executives see it differently, adoption depends on people and trust, not the model. Any competitor can buy the same AI. What they can't buy is whether employees and clients trust the systems being build, and whether the guardrails around them hold up. Cavellier agrees: software along isn't a lasting advantage.
When staff feel confident in the safety and accuracy of their tools, they can move much faster without creating new risks. “I think finance leaders are actually right about that. Any company can buy the same AI models we have,” Cavellier points out. “That said, the model isn't the strategic advantage; it never was. What actually separates Cass from other companies is that our staff and customers trust the AI solutions we’ve built, and our controls are strong enough to allow them to be used quickly. That's culture, not tech.”
To protect that trust, Cass keeps automated models separate from its core transaction processing, using them instead to speed up workflows and handle exceptions. This setup ensures that high-volume payment operations remain completely predictable while still giving teams room to innovate. “On the trust side, we run AI deterministic-by-default,” Cavellier explains. “This speeds up how we build systems and handle exceptions, but it's not sitting at the core of high-volume invoice and payment processing. This means our team members and customers aren't second-guessing technology on transactions that matter most.”
At the same time, the company set up clear review processes before deploying new tools to live environments. This early governance keeps teams aligned and prevents rushed deployments from creating operational problems down the line. “On the control side, we stood up an AI Governance Committee before we even started scaling anything out,” Cavellier says. “Every AI tool and service goes through that committee before it touches production, without exception. That's another thing competitors can't just buy: an organization that's already comfortable moving fast while staying accountable. When others in this space are still debating whether to trust AI with anything real, our teams have already built the habits and guardrails to use it wherever it makes sense.
Many companies struggle to strike the right balance between rapid deployment and careful risk management. Focusing only on speed leads to unmanaged software rollouts, while focusing only on risk leads to endless policy debates and missed opportunities. Cavellier argues that successful transformation requires leaders to pursue both priorities at the same time without treating them as tradeoffs. The goal is to build an organization that can ship new capabilities quickly, while still standing behind its results.
Instead of getting caught in either extreme, businesses must learn to combine fast execution with dependable oversight. “Every company is doing the same thing right now: piloting new AI solutions, writing policies, and hoping they work,” Cavellier explains. “Leading the pack of your competitors requires something different, however: the ability to move fast while still maintaining consistent, reliable results. They are not mutually exclusive, nor should they be considered tradeoffs. Many organizations are focusing only on one half of that: speed. The ones chasing speed end up with ‘AI everywhere’ and no ability to control or explain any of it. That approach is a liability waiting to surface, not an advantage. Other companies that are chasing control only are getting stuck writing policy, while everyone else ships new products and services. Neither one truly delivers anything meaningful and sustainable.”
By pairing business and technical leaders together, testing through live prototypes, and setting up clear guardrails, companies can make change stick across their workforce. For Cavellier, long-term success comes down to building a team culture that values both speed and accountability. “At Cass, we focus on both at once, moving with urgency while controlling and standing behind measured outcomes,” Cavellier concludes. “This is more than just a technology play; it's a leadership commitment.”
Follow Jim Cavellier on LinkedIn for more insights on enterprise digital transformation, AI governance, and building high-performance technology cultures.
 consultants ltd.jpg)
Comments