Leveraging AI in Legal Operations: Four Key Insights

The intersection of artificial intelligence (AI) and legal operations is a hot topic as legal departments look to drive efficiency, reduce costs, and keep pace with the business. However, there is always the understandable fear that AI will replace many legal operations roles in the future as tasks become more streamlined and automated.

In Axiom’s recent webinar, The Future of AI & Legal Operations, our speakers quelled the anxiety gripping much of the legal ops industry with a more optimistic perspective of the integration of AI into the day-to-day tasks of legal operations professionals. Their conclusion? AI will not replace roles, but it will replace tasks.

Moderated by Axiom’s own Maryam Salehijam, PhD., the following panelists shared their first-hand experiences:

• Brooke Van Dam, Johnson Controls
• Memme Onwudiwe, Evisort Inc.
• Ashley Miller, Capgemini

What did the webinar reveal? Four key takeaways:

1. The vast variety of AI and what makes them different
2. The difference between automation and augmentation
3. How to identify the right opportunities to use AI
4. The importance of vetting the right vendors

Looking for more educational resources on AI in legal ops? Sign up for our newsletter and stay updated on new webinars, blog posts, whitepapers, and more!

1.  The Different Types of AI

Before diving into AI implementation, it's critical to understand that AI is a broad term encompassing many different technologies. Some common types that legal teams are already using include:

• E-discovery tools leveraging natural language processing (NLP)
• Computer vision for converting documents via optical character recognition (OCR)
• Extracting data from contracts using NLP and machine learning
• Generative AI chatbots that engage in human-like dialog to answer questions

The latest generative AI models like ChatGPT have made AI more accessible by allowing users to interact using natural language, but corporate legal teams have been using AI for years in various forms.

Memme Onwudiwe explained: “I think generative AI, which was introduced to the broader public through ChatGPT, really changed how folks saw artificial intelligence. But when you think about tools like Netflix, Spotify, or even when you're on a plane and most of the time the pilot isn't even flying the plane, it's automated unless there's an emergency. We've been relying on AI for a long time.”

Onwudiwe continues to note that for AI in the context of legal operations, many have been using it for some time in the e-discovery space: “Natural language processing and deep learning tools are also used to extract data from agreements. When we think about generative AI and how it’s changed things, it allows for an artificial intelligence that, especially in the LLM perspective, can guess what the next best words produced will be, right? So what that means is that you're able to engage with these technologies at a level where it won't just be a pick list, but you're able to actually have natural language conversations. That helps a lot with querying when you have a lot of data or a lot of structured data about your agreements or your past litigations. If you don't - if you must approach in a mechanical way, which can be difficult.”

AI now, according to Onwudiwe, allows users to query their questions and engage with AI more naturally. It also allows the responses that users get from AI to be less mechanical and to be in a more normal human language. Onwudiwe believes those two capabilities alone really changed how people engaged with the tool in the last year, making it more accessible and discrediting the belief that AI is difficult to use.

These technologies are now big game changers, enabling legal teams to ask natural questions like “What risks are exposed to us because of the global pandemic across this corpus of contracts” and receive natural responses. While it’s easy to see how helpful these technologies will be for complex tasks, the key for legal ops teams is identifying specific use cases and matching them to the right AI solution.

2. Automation vs. Augmentation

Another important distinction for legal ops leaders to note is whether AI will be used for automation or augmentation. Automation involves having AI completely take over a manual task, while augmentation uses AI to assist and enhance human work. Examples:

• Automating NDA creation with a self-service portal
• Using AI to route legal service requests to the right attorney
• Leveraging NLP to extract contract metadata to augment human review
• Employing AI to check spelling and grammar in documents

Brooke Van Dam shared that her team looks at the following questions before deciding between the two needs:

• Is this a process that needs to be more efficient?
• Is this a manual process that could become a digital process instead?
• How many people need to have their eyes on the data being addressed?
• Would you need to create a new workflow around that task with new inputs?
• Does this process require collecting data and learning from it to make other decisions?

For Van Dam, automation allows her team to enhance their processes’ efficiency by removing the manual components or the paper components, eliminating the need for human hands on the task. Ashley Miller added that “automation comes into play where we can actually define rules and different layers to those rules and create workflows to get requests to the right person very quickly with the right information out the gate.”

But augmentation, Miller defined, is the “enhancement of the human task.” After a request has been routed automatically, high-value complex tasks can be augmented with senior lawyers in the department to focus their efforts and improve efficiency.

The choice between automation and augmentation depends on the complexity of the task, risk tolerance, and availability of clean data to train AI models. Legal ops leaders should consider that simpler, repetitive, routine tasks are better candidates for full automation. More nuanced work benefits from AI augmentation while keeping a human in the loop.

3. Identify the Right Opportunities

With so many potential applications, how do you prioritize where to utilize AI? The panelists recommend starting with areas that are high-volume, repetitive, and manually intensive. Some common examples:

• E-billing and spend management
• E-discovery document review
• Subpoena response workflows
• NDA and simple contract generation
• Regulatory and compliance policy distribution
• Legal request intake and triage

The key is analyzing your team's activities to identify bottlenecks and pain points. Review workload-to-headcount ratios, manual handoffs, and cycle times.

Weighing the workload against headcount is a major factor. According to Van Dam, “If the workload is more than the headcount and the work is what we consider non-critical to have people do it all, like an NDA, that’s where we want to think ‘how can we automate? Is there technology that can come in to reduce the time it takes a person to do this?’” She encouraged considering areas where AI could spot potential risks or opportunities across large datasets, like obligations in contracts.

4. Vet Vendors Thoroughly

While AI holds immense promise, not all tools live up to the hype. It's crucial to thoroughly vet AI vendors to ensure their capabilities match your requirements. Remember:

• Slick demos don't always reflect real-world performance
• Request a proof of concept using your own data and evaluate outputs
• Ensure AI is actually doing the work vs. humans behind the scenes
• Test with unusual edge cases, not just standard examples
• Check if cloud AI tools comply with your data security policies

Panelists advise taking a "trust but verify" approach to AI buying. Start with pilots before locking into long-term contracts. And don't hesitate to build your own tools if off-the-shelf solutions fall short.

Miller shared, that for her team, many software programs didn't “meet the needs of all of the audiences we're trying to consider, and they didn’t entirely get our purpose.” She emphasized that she wasn’t trying to say legal leaders couldn’t customize what they buy, but for her team’s needs, “sometimes it's really worth evaluating those needs very closely to get that customized system that you know is going to tick all of those boxes.”

Bonus Tip - Stay Curious!

A final piece of advice - stay curious and keep exploring AI's potential. Technology is rapidly evolving and new use cases are emerging all the time. Carve out time to experiment and consider how AI could enable entirely new capabilities, not just accelerate existing ones. With the right strategic approach, embracing AI can be a game-changer in optimizing legal operations and the legal industry.

In conclusion, the intersection of AI and legal ops presents immense opportunities - but also many complexities to navigate. By following these best practices and insights, legal departments can harness AI's potential to drive efficiency, reduce risk, and ultimately become a high-value partner to the business. The key is starting with the right use cases, choosing vendors wisely, and proactively managing the transition. With an open and iterative approach, the future of AI-powered legal ops looks bright indeed.

💡 Explore the ways you can start leveraging Artificial Intelligence.