More and more legal departments are working to implement artificial intelligence into their workflow. The days of waiting to see what the technology can do are coming to an end, and legal departments are now starting to expect their firms to use AI.

Brian Kuhn is the co-founder and global leader of Watson Legal Practice at IBM. He will be giving a keynote speech at Legalweek in New York on Jan. 30. Kuhn spoke with Corporate Counsel about AI, what it takes to implement it in the legal department, and ways it can be used in the future.

This conversation has been edited for length and clarity.

Corporate Counsel: In the past there has been some hesitation on the part of legal departments in implementing artificial intelligence. Do you still see a hesitation, or are legal departments more willing to use it in their day-to-day work?

Brian Kuhn: They [in-house legal departments] are more on board with AI than law firms. I think a lot of that has to do with the fact that they [legal departments] are aggressively in-sourcing. I believe law departments are keeping 48 percent of work in-house, and that number is increasing. One of the things we've heard in having conversations with law departments in the U.S., Europe and Asia is that the law department has been discovered by enterprise senior leadership as a place to take out cost. So if a company needs to take out $1 billion worth of costs, the law department is where the C-suite is turning and where they're seeing success. I think that one of the key things we hear when speaking to companies is that any of the law firm tools are tools that companies are willing to buy and can serve themselves. We're absolutely seeing they are using artificial intelligence to accelerate the in-sourcing aspects and empower their full-time employees.

CC: In your experience, do you find that law departments want their firms to implement artificial intelligence?

BK: Yes. I think law departments are 66 percent more likely to hire a law firm if they believe that law firm is innovative and that innovation generates meaningful results. They don't want to develop software, and they are seeking law firms to do it because the law firms have all of the historical data representing previous representations of those companies that they can find.

CC: Are there steps that legal departments need to take before they begin implementing AI into their work processes?

BK: It really depends on whether they're in the market for a customizable solution, which is what we do, or a solution that it more general in nature. Both solutions have their merits. It just depends upon the complexity of the project. Supposing they want to work with IBM, we would need to go through a workshop [with the client] and identify not just what they want to do with AI, but their current workflow. We'd bring in the people that are doing that work today. We'd also look at what their business goals are. Is it to reduce outside spend? Then we look at the workflow and ask in that workflow, where specifically are they seeing paying points. Some aspects of the workflow might be just fine. Others might not be. Those are the areas we tend to target for AI. Then we identify a metric for success up-front, and with that as an anchor we go forth and build a solution if there is enough data and if that data is of a sufficient quality.

For companies and for law firms, AI is going to become increasingly customizable. Right now a lot of AI solutions are still analytics tools, which is one size fits all. What companies across industries are spending the most on AI is enhancing the customer experience. That could be the clients or their internal clients. A lot of thought and planning has to go into what do you want to do; why do you want to do it; what business impact will it make; do you have the data; and how will it improve your business over time. To do that companies need to commit to bring the business leaders in so we can understand the business needs.

Training of AI has become a lot less cumbersome, but still we need to expose the company to subject matter experts and [information technology] folks so that you can follow process. In the past it was programmers only and not creative people and businesspeople. The advantage of businesspeople and creative people in the mix is that there is less lost in translation compared to a data scientist trying to take legal concepts and create code around them.

CC: What are some of the popular uses of AI that you've seen in the last year?

BK: If you're an insurance company, you might want to know when a claim becomes a lawsuit it costs five times more to settle. We can continue to hire outside counsel and we should, but we want to understand and predict which claims are likely to become lawsuits in the first place and triage them for settlement. What you're seeing now is not just a trend towards applying AI to existing legal tasks but its greater potential, which is to predict outcomes and get upstream of business goals and business pitfalls.

CC: Is there anything that you think AI could be used for that legal departments have not begun to do so yet?

BK: Yes, and I would suggest that it is the most valuable use, broadly speaking. If I'm a corporation, I have preserved and have to hold on to the work product that outside counsel has historically created in the past. AI is not applicable to one-off company litigation. Most companies face repeating problems with variations.

If they could mine the work product that they paid for and own in the context of their problem with a solution that is smart enough to find it and identify why it's relevant and help them predict what a likely outcome will be when using this strategy or documenting this clause or repurposing for their specific needs, not in general, then they would be able to take more control over legal risk.