From Coast Guard to CDO: Todd James on Scalable AI and Delegation

As the former Chief Data and Technology Officer at 84.51°, the data science arm of Kroger, Todd James has helped to modernize one of America’s largest grocery retailers using AI. But his biggest insights weren’t about models or infrastructure. They were about leadership.

On a recent episode of Data Radicals, James shared how scaling AI requires not just technical acumen—but also trust, humility, and an obsession with solving business problems. Now founder of the consultancy Aurora Insights, he brings these lessons to organizations navigating the next wave of data and AI transformation.

Lead less, empower more

At the heart of James’s approach is a simple leadership truth: great data leaders don’t try to do it all. They delegate—not to offload work, but to create opportunities for others to grow.

“You delegate everything that if it breaks, and you can put it together, even if it’s painful… Those are opportunities for your team,” James said. “The things that you can’t fix, the things that are super high risk—yeah, you got to stay more involved in those.”

It’s advice he’s carried with him since a formative dinner early in his career at Deloitte. And it’s a mindset that shaped his leadership at Kroger, where scaling AI meant empowering domain experts to own parts of the process.

AI in retail: From science to store shelves

At 84.51°, James helped shift the business from a consulting-led model to a scalable, technology-driven operation. One standout initiative? A real-time routing algorithm for store associates picking online grocery orders.

By optimizing their paths through the store, “we were able to put in advanced routing algorithms that reduce the distance traveled on a pick order by about 10%,” James explained. The impact? Shorter lead times for customers and a more efficient, less chaotic experience for employees.

But the innovation didn’t stop there. The same algorithmic principles were reused to optimize truck routes from distribution centers—leading to an 8.63% reduction in distance traveled. That reuse wasn’t accidental; it was by design.

“We have to start to think about building a layer of algorithmic infrastructure that allows for reuse,” James said. “That’s how we’re going to get speed. That’s how we’re going to be able to manage at scale.”

Start with the business problem, not the platform

James is passionate about flipping the script on data architecture. Rather than starting with a data lake or catalog and hoping for adoption, he insists that teams should “start with the business.”

He shared: “Let’s line up all the use cases, figure out where there’s critical mass around data, and we’ll move the data as we do the use case… I would much rather be a little bit inefficient, create a little bit of throwaway, but start driving business value today.”

This mindset was key to embedding AI across Kroger, beyond marketing and merchandising. James and his team physically visited stores, shadowed operations, and built trust—grounding their science in reality and boosting adoption.

Why the CDO role may disappear

James also believes the Chief Data Officer role, as we know it, won’t last.

“I don’t think this job lasts forever if properly managed,” he said. “At some point, there is going to be enough knowledge around the application of AI in the business that a lot of the really exciting aspects of deploying advanced analytics are going to be part of the business roles.”

As AI becomes embedded into marketing, supply chain, and finance, James predicts that more domain leaders will be trained in data science—and more data leaders will move into the business.

Data governance, in turn, may move back under IT. But for now, the opportunity for CDOs is to embed, empower, and educate—so their roles become obsolete not from irrelevance, but because the business has grown more intelligent.

Scaling AI Isn’t just technical—It’s cultural

Not every AI effort at Kroger succeeded. Some failed despite technically sound models.

“The people at the consumption end of your sciences… if they aren’t bought in, they don’t get implemented,” James said. “The biggest mistake you make is not killing them [unsuccessful models].”

Instead, James advocates for investing at least half of your energy into managing organizational dynamics—talking to end users, building empathy, and earning trust. AI adoption, he argues, is “more about people than the math.”

From CDTO to consultant: What’s next for Todd James

Now, through his consultancy Aurora Insights, James is helping other organizations scale AI the right way. Some want to get started. Others are stuck and struggling to see ROI.

“I think it’s important for their competitiveness. I think it’s important for the future of the U.S. and global workforce,” James said. “I’m energized by the fact that I think there’s a fundamental purpose that I can serve by helping people get through these challenges.”

It’s clear that James’s unique background—from the Coast Guard to consulting, from financial services to retail—isn’t just a résumé. It’s a playbook for modern data leadership: adaptable, collaborative, business-first, and always learning.

Final takeaway

Great AI isn’t just built. It’s embedded—into processes, decisions, and most of all, people.

As James put it:

“Ultimately, the desire for every company should be, I want to positively impact every single decision point with data and advanced analytics.”

That’s the goal. And as Todd James reminds us, the best way to get there might be letting go.

Curious to learn how a data catalog can support your AI initiatives? Book a demo with us today.