Cox Enterprises: from a trusted data foundation to AI-powered Finance

Now, we are fortunate enough to partner with Cox Enterprises, one of our customers, through our private preview program. And they helped shape the direction of all of these three agents, right? Let's hear about their story and how SensibleAI AI agents are making a real difference within their office of finance.

Please welcome Cox Enterprises, AVP of Finance Solutions Delivery, Allison Adrian. Hi, everyone. I am so excited to be here today to share with you our journey. It is very fitting because it actually started at Splash last year where we began conversations about being part of the private preview.

But before we move into finance or talk about finance or AI, I want to start with innovation. Because at Cox, innovation doesn't live in a lab or a tool. It starts with the business. This week, you will see that firsthand at the Innovation Hall where we are showcasing Cox Farms.

Cox Farms is the largest greenhouse operator in North America, using high-tech indoor agriculture to grow fresh food closer to where people live. It's a great example of how Cox invests in innovation to solve real-world problems like sustainability, food security, and supply chain resilience.

So in case you haven't heard of Cox, it's a family-owned Atlanta-based enterprise spanning communications, automotive, media, and sustainability. And while our businesses look different on the surface, what connects them is long-term thinking, disciplined growth, and practical innovation.

Finance plays a key role in enabling that, not just reporting the results, but supporting better decisions across the enterprise. Like many large organizations, our finance environment grew organically.

Multiple divisions, multiple systems, many, many manual processes, which meant time spent reconciling data instead of explaining it. We knew that if Cox wanted to scale innovation, finance had to scale with it.

That meant fixing the foundation first. One stream gave us a unified platform and a single source of truth. We started with consolidations and then expanded into planning and reporting. We created consistency and trust in our data.

And today, hundreds of users across accounting, FP&A, tax, and treasury work from that same foundation, which made AI adoption realistic and responsible. So you might ask, how did we approach AI? We didn't ask where we could deploy AI the fastest.

We looked where we could unlock the most value and build momentum. And that's why we started with accounting, not FP&A. Predictable workflows, high volume, and clear policies made it the right place to start.

Improved value and confidence before scaling it further. So we had several use cases. I'm going to give you an example in each of the agents. So finance analysts proved fast, reliable answers using our trusted one stream data, which reduced the back and forth and allowed our teams to really focus on analysis of data versus the retrieval of data.

Search enables self-service across policies and how-to guides for quick answers, which reduced dependency on our system administrators. And deep analysis scans revenue contracts at scale. Pulling out the key terms specific to our different users and allowing the teams to really make decisions rather than the manual review.

And with any new capability you learn along the way, and we had three main lessons learned. Configurability is powerful, but requires discipline, especially for AI. Iterative delivery beats big bang. And change management is just as important as the technology itself.

One stream is highly configurable and flexible, which is great, except when training AI. Context and naming conventions really matter. So we learned quickly that refining business context, example questions, and security setup in partnership with our users was essential.

And so use cases improved materially throughout UAT, previews, and the phased approach rather than a one-time design. Lastly, clear communications, role-based training, and early life support were critical to adoption.

Treat the AI agents like a new capability, not just a new feature, and plan accordingly. By treating AI as a new capability with enablement, guardrails, and a phased rollout, it helped us gain momentum and adoption.

So what's next, you might ask? As a private preview partner, Cox will continue to influence the roadmap direction, ensuring sensible AI agents evolve in ways that are practical, secure, and purpose-built for enterprise finance.

We asked OneStream in the beginning to move finance analysts into Excel, because that's where our users are. And by meeting them, they grew comfortable and more confident. But where we're headed next is orchestration.

Today, OneStream helps us see that a variance exists. But what excites us is answering why it happened. With MCP coming soon, we see one stream connecting securely to transactional details in systems like Dynamics and Workday, allowing finance teams to move from identifying an issue to truly understanding them, and all without leaving the platform.

So while we're still early in our journey, we anticipate thousands and thousands of hours saved across each of our functional areas. Innovation at Cox is practical and purposeful. AI isn't about replacing people.

It's about multiplying expertise. And with a strong foundation, thoughtful change management, and the right starting points, finance becomes a catalyst for innovation and not a constraint. Thank you for listening to our journey, Thank you.