Beyond The Hype: Use Cases And Strategies For Implementing AI

Artificial intelligence is, indeed, the future of legal operations.

We’ve already written about using business case studies, whether your personality is too legal, the importance of eSignatures, and the nine ways the legal profession needs to change. We’re now in our final article in our series based on learning from the Corporate Legal Operations Consortium (CLOC) annual meeting.

Ginger Dolgow, senior manager at NetApp, Tami Baddeley, operations lead at Microsoft, and Mike Naughton, senior manager at Cisco, all agree: though it is a highly hyped topic of conversation, artificial intelligence is, indeed, the future of legal operations. “The good news is that AI is more than hype and can offer strong returns on investment,” they explain. “The bad news is that getting a solid return on investment in AI often requires a further investment of time and money. These solutions do not deliver as promised right out of the box.” Dolgow, Baddeley, and Naughton have shared the secrets and best practices of using AI to support contract management.

Choose high volume and low risk to start: All three recommend starting with an agreement that is high volume and low risk, with a well-established process. Many companies start with NDAs or internal human resources documents. This way, you can leverage your existing templates, playbooks, and cheat sheets to automate the process. It is also a good idea to slowly expand to increasingly more complex contracts allowing you to gradually learn what works for your organization and the best way to leverage AI. Naughton observes, “It takes time to learn and to do it right, both generally and for your organization specifically.” For example, a second phase of AI implementation may involve applying AI to M&A documents, RFPs, or other short, relatively routine contracts.

High quality matters: Dolgow also emphasizes that AI should be applied to high-quality documents. “You must have high quality documents and may run into issues with historical agreements. You may need to manually extract relevant data to make sure that your AI output is accurate, predictive, representative, and comprehensive.” Dolgow also notes that you will need a large number of documents to achieve an accuracy level that allows you to reduce or eliminate the level of human touch. “Vendors tend to promise that an AI implementation requires 300-500 contracts,” she says. “, it took us a couple thousand because of the variations in third-party paper and the non-standard provisions we are tracking” Because of this, Dolgow recommends being patient and working through the process.

Be prepared to invest money: Naughton adds that preparing for a AI project involving the analysis of thousands of contracts is not cheap. “It is a major investment and it requires management and maintenance,” he says. Because AI is such an investment, Dolgow recommends asking for the total cost estimate upfront. All costs should be accounted for, including training, hosting fees, consulting, and others. “You want to know the costs of implementation, operation, and maintenance upfront,” she says. “Otherwise, you will be surprised. And most legal departments don’t like being surprised, especially when it comes to their budgets.” Dolgow also emphasizes that costs structures may vary. For example, additional searches could cost more, especially if you create a separate project in the software for a certain subset. Therefore, she recommends “working with your vendor to structure the right interactions and avoid additional costs.”

Invest in a data repository: It is also a good idea to have an in-house repository for your data. “Once the data is extracted, we move it to our metadata repository. This way we can be flexible. We are not dependent on the provider, and it allows us to try different AI technologies through our legal services provider,” Dolgow explains. “This AI market is changing and you may not want to be locked in,” she adds. Baddeley agrees, suggesting that “it is worth investing time and money into a data mart, or data repository. This data can be later used for other unrelated projects. Once you have enough data, you can play around with it and do some interesting things.”

Choose providers carefully: Before you begin your AI journey, Dolgow recommends defining success and choosing providers carefully. “Make sure you align with your AI providers,” she says. “Check references. It is a good idea to find out where they failed and where they succeeded. This way you can learn from others.” Because the AI market is so new, Naughton recommends that companies be mindful of their expectations. “It is an immature market – both customers and vendors are immature. That means you need to be open minded when you start this process,” he says. “The truth is that no one knows everything.” Naughton suggests asking your AI vendor what space they were in before. “This will help you think about context and their baseline,” he explains. “While AI is a powerful engine, it is not a specific concept. And asking people for their background may provide you with enough missing information to form a complete picture.”

Sponsored

From Dolgow, Baddeley, and Naughton’s experiences, it’s clear that AI isn’t just the hypothetical future of legal operations: it’s here now and it’s here to stay. Their fearless yet strategic approach is inspirational for anyone seeking to implement not only AI, but also any other new legal technology. Collaboratively sharing experiences and collectively developing best practices is the only way we can thrive through the legal profession’s technology-rich future.


headshotKathryn Rubino is an editor at Above the Law. AtL tipsters are the best, so please connect with her. Feel free to email her with any tips, questions, or comments and follow her on Twitter (@Kathryn1).

Sponsored

CRM Banner