Creating a Single Source of Truth for Enterprise Legal Work - with Christo Siebrits of AbbVie

**Daniel Faggella** (0:14)
Welcome everyone to the Emerge AI in Business Podcast. Today's guest is Christo Siebrits, Senior Associate and General Counsel at AbbVie. Christo discusses the implementation of an internal large language model environment that allows employees to select from various validated models for processing sensitive data. He emphasizes a management strategy of forced ranking to identify and fund only the top priority use cases across the organization. This approach prevents redundant technical investments by multiple departments and focuses resources on the most effective applications. By integrating legal and cyber security teams early in the planning phase, leadership can deploy these tools within established risk tolerances and compliance frameworks. Please note the opinions shared by Christo Siebrits in this episode are his own and do not represent the official position of AbbVie or its leadership. Before we begin, a quick note for our executive listeners, Emerj invites enterprise leaders who are driving meaningful AI initiatives to share what they're learning with a peer audience. If you're moving real projects forward and want to be part of that conversation, you can learn more at go.emerj.com/expert. That's go.emerj.com/expert.
Now the conversation with Christo.

**Matthew** (1:41)
Chris, thanks so much for being with us again. It's great having you back on the program.

**Christo Siebrits** (1:44)
Massive, great being here again. Thank you for having me.

**Matthew** (1:48)
Absolutely. Today, we're talking about something slightly different. We've had you on the show. We just recorded a moment ago talking about maybe the legal concerns, especially with those programs outside the organization, Shadow AI as it gets called. We're also hearing from legal leaders in this series and elsewhere that modernizing is an inevitable conversation in 2025 It's an absolute foregone conclusion. And in the way of that, information is scattered across systems, which gets in the way of answering basic questions about any initiative. And at the end of the pipeline, no one really feels confident yet about grading AI work in legal workflows.
At least as of 2025, we're still seeing a lot of unclear ROI metrics and governance structures. It all goes without saying that this is a landscape, you know, ripe for risk before even getting to constant demands for speed and accountability. I know we spoke a lot in our last conversation, like I said, about shadow AI and interacting with data from the outside. But where are core business problems? What are the core business problems for enterprise legal teams today from within?

**Christo Siebrits** (2:57)
Wow, that's a whole mouthful right there. I mean, the thoughts that crossed my mind as you were posing that problem statement is the adoption curve. So we're just making an assumption here. Well, firstly, let me say the potential of AI is huge. That's why we were deployed to find these solutions and build the machines that need to solve the problems, the AI machines.
But that presupposes that everybody that is interacting with AI actually wants to do that. And secondly, is capable of doing that. I would say, and we discussed this in our previous program, we have many thousands of employees, and I would argue a significant percentage of them want nothing to do with AI. So you have to deploy AI across an entire environment, and you have to convince people to participate in this experiment, and many don't want to reskill themselves to even understand what AI is. I would wager that there are a significant percentage of employees who haven't even gone to ChatGBT to see what that is and how it works. So the starting point is challenging. That being said, we have lots of opportunities with our internal data to do many things, and we are looking at that, you know. We're looking at how can we mine data that we have, how can we use that data that we have, how can we equip teams with the knowledge to use this information. And internally, we've started a program where, at least at the highest group level, we're putting teams together that are responsible for training people on, here are some AI tools, internal tools that you can use. Here is how you use them. And then, more importantly, sharing how people have benefited from using them. This I have found to be a very effective tool, just to spread the word and to build confidence.

**Matthew** (5:12)
Tell us a little bit, really, with that internal tool training. Just you spoke so eloquently last time, just about the build vs. buy question. And I know that kind of became on the spectrum of solutions for how to approach the different risk, as you were saying in the last episode, the different risk depending on your industry, your organization's tolerance for risk, for giving their data to a third party. For these internal tools, walk us through, especially as you're thinking about maybe the vendors, what does this look like as a success from a training standpoint that we know we can use this as an internal tool?