CCBJ: Matt, please tell us about your background.
Matt DenOuden: I’m a lawyer by training who transitioned to legal technology 21 years ago. For the last six years, I’ve led Onit’s sales and account management teams and been a member of co-founder and CEO Eric M. Elfman’s executive team.
Artificial intelligence has long been a subject of sci-fi movies, but it is quickly establishing itself in real life. Can you tell us what AI is?
In a nutshell, AI is technology doing something that a human mind would have otherwise done. My favorite definition of AI is the most practical one. Fundamentally, it’s a technology that does a task or suggests decisions that a person would have otherwise done. And I emphasize that it’s often a part of a workflow instead of accomplishing the whole thing. It’s helpful to think about it in those simple terms.
In the legal enterprise software world, AI has gained a foothold and is quickly growing. Why is it becoming so interesting for lawyers and legal operations?
It’s because there is both top-down interest and pressure and bottom-up enthusiasm. We hear from companies that legal has AI as part of their overall technology strategy, whether its use relates to contract management, invoice review, or an overall digital transformation of the business at large.
The bottom-up enthusiasm is there as well, driven in part by the belt-tightening that happened during the height of a pandemic and for the good fiscal performance reason of trying to find the most efficient ways to do things. There are very demonstrable ways to save time on work through AI. It has a tangible effect. For example, a study found that those who use AI to review contracts can be 52% more productive. If you compile that across everyone in corporate legal reviewing contracts, many people are suddenly finding a task much easier to complete. Results like that stoke enthusiasm from people who are doing the heavy lifting.
How does AI allow attorneys to focus on higher-value work and better serve their businesses?
When IBM came out with Watson, they had to show it doing things like play chess against Russian masters. This wasn’t because the question-answering computer system couldn’t do other things, but because chess-playing was something that people could see and understand. When we set out to determine how legal can use AI most effectively, we focused on areas where there was a lot of activity or some inefficiency, or on ways the lawyer was being pulled down into the weeds of a task that might otherwise be handled by AI.
That led us to focus on AI use cases in the contract and invoice spaces. The principle is that you’re not just making the workflow happen faster. You’re helping the lawyer and knowledge workers work at the top of their licenses rather than being mired in some of the less important processes. The prework, if you will.
Let’s go back to the contract review study for AI. Why is AI so helpful in this context? It thinks and moves through work like a lawyer. It functions like a sous chef. After AI’s first pass review, a lawyer opens up a contract and finds flagged issues, suggested edits, and an overall risk profile. That level of contribution not only removes work from that lawyer’s plate but also gives them a running start on review before they even open the contract.
You mentioned that AI “thinks and moves through work like a lawyer.” What do you mean by this?
From our point of view, AI should help lawyers where they need the most help – which is often where they face repetitive, high-volume and manual tasks. We’re drawing from our history in legal when we identify these areas. I have a background in legal that goes back more than two decades, as do many people on our teams do. The founders of our AI Center of Excellence are all practicing lawyers, so we’re using our experience to train AI to think like a lawyer. There is always a focus on making sure that we’re looking at AI through the lens of legal, practicing in-house lawyers, and even outside counsel when we’re thinking about contracts.
We also try to make sure that we think about things as a process and that the AI is walking the user through the process that’s familiar to them. Rather than reorganizing their work, the technology takes them along a familiar path. For example, a lot of what we do with contracts in the AI space involves AI redlining. An AI-based Microsoft Word add-in allows faster legal contract review and editing. We’re taking the technology to lawyers where they work without removing them from the tools they rely on every day. It’s the same sort of place they would’ve otherwise been if they were manually redlining that contract and running their playbooks.
Can you give us some examples of AI-enabled solutions used now and how successful they are?
In the last couple of years, we came out with AI models against legal spend that now have over a thousand customers and well over 500 of whom are using some form of our e-billing. We have decided that there ought to be a technology-aided prereview for interested customers who want to make their work lighter.
For a standard e-billing system, you have billing rules. But billing rules can’t capture every potentially non-compliant charge in a legal invoice. They rely on words and descriptions. If law firms use a slightly different word choice, that line item could be approved for payment. Our AI is looking for findings and guidance to catch “between the rules” errors.
For example, consider rules in your outside counsel guidelines around the kind of work that can be done by what kind of charging person – whether it be an associate, partner or paralegal. AI can interrogate the work, even if it isn’t coded in a way that draws attention. It can say whether it’s chargeable, but also if it should have been chargeable by a paralegal and not by an associate.
Rather than a human sussing out possible incorrect or misallocated charges, the AI does that prework for them. For example, we chatted about using AI in the contract space earlier. The benefits extend far beyond just the reviewing attorney. Rather than having the front-to-back review done by multiple different lawyers and people in the process, the AI can scrub the agreements. These are usually third-party agreements, but they can also be first-party agreements that have been heavily redlined. AI then runs them against the customer’s playbook, risk ratings, and intentions around contracting. AI makes redlines to the playbook or at least draws attention to the areas where a lawyer should focus rather than starting from the beginning.
We have customers who offer self-service, where the salesperson or another person in the organization sends something simple to legal like an NDA. AI can review and redline the document without a lawyer even being involved if the corporate legal department wants it that way.
Ultimately, speeding up the pace and effectiveness of contract review benefits everyone involved – regardless of which department they are in. It’s the entire enterprise that wins. Each deal closed means more revenue for your business or establishes a valuable partnership, initiative or technology.
Where do you see AI-enabled software heading in the coming years? Do you have any predictions for more amazing advancements?
At Onit, we are really excited about where we are going. We’re determined to finish the things we’ve begun, which is to continue to drive AI into e-billing and contracts. There are also some exciting opportunities in risk management and compliance matter management. We have customers who use our overall technology for what we call “legal service requests.” Someone within their business “knocks” on the door of legal asking for help of some kind. If, as is often the case, the answer to their question already exists in legal, there is no need to trouble any particular person to rearticulate it. AI can be used along with search functionality to identify whether the question is an FAQ and provide the answer.
Then there’s the modeling of disposition data. In other words, what were the outcomes of matters in our customer’s matter management systems and what do they tell the customer about where there is risk and where there’s success? Where certain strategies were particularly successful, that knowledge is beneficial from a compliance perspective.
Many of our customers are not only using Onit for matter and spend management but also to enhance compliance workflows. The fundamental point of compliance programs is to avoid risk and address issues prospectively. Sometimes that’s easy because of transparency. But sometimes, AI can find patterns and connections in the data that would not have been apparent to humans. We see a lot of receptivity in the legal market, and there’s more road ahead of us than in the rearview mirror.
Attorneys as a group have the reputation of being slow to adopt new technology. For the AI-hesitant, what are your final words of advice?
Use your instincts about where AI will be helpful and where it can do prework. Look for those pain points and places where AI, when properly set up, can automate prework and take a load off your team. Talk to customers who’ve used the technology for the same use case. Then try it out yourself. If you focus on AI’s very tangible outcomes, the decision-making process will be relatively easy. You won’t have to interpret your success. You can see your success.