Every day, there seems to be an announcement about a new AI model being released, a new product that takes AI further, or the next big thing coming. It’s hard to keep up!

At the same time, there’s no shortage of stories about what AI can’t do and won’t be able to do. It’s easy to find examples of people testing some of the hardest things they do and reporting that AI can’t do that yet, so it’s not worth using.

I have more time and interest to explore AI than most lawyers, but I’m far from an expert. What I have found, though, is that there is plenty that AI can do which is helpful, and that’s what I’m most interested in.

Here are a few examples.

Perplexity

I regularly use Perplexity to help me research topics and get up to speed on an industry, a company, or a problem more quickly. I find it a great sounding board against which to test ideas.

CoPilot

We have Microsoft CoPilot. I like that it is part of M365 where we store most of our important information, but I do have my frustrations with it. I don’t trust its reliability when I need the results to be accurate and correct, as I have seen some bad examples of it getting things wrong. However, I do find it very helpful to quickly find and interrogate emails and documents within our M365 environment. Yesterday, I needed to find a contract in our contracts database and understand how much notice is required to terminate it for convenience.

Our contracts library agent found precisely what I needed without any issues.

I’ve also used CoPilot to generate simple clauses for contracts when I’ve wanted to add something, and its suggested clauses were fine for what I wanted.

ChatGPT

I struggle to write good copy for sales and marketing material. I’ve found that Perplexity can be a bit too academic and CoPilot’s guard rails too limiting, but ChatGPT has given me some excellent results when I add a good prompt and guidance as to some copywriters I admire and would like to follow.

ndMAX

For those that have NetDocuments, I’ve been testing ndMAX, as I like the idea of bringing the AI to your documents and the existing security permissions. I also like the idea of enhancing more traditional document automation (with all the structure and predictability it allows) with elements of generative AI for areas that lawyers often struggle with (such as drafting service descriptions or working out suitable performance standards or milestones). For now, I see it as necessary that humans remain in control of reviewing and refining the suggested outputs, but this can really speed up and improve the quality of this part of the process.

CoCounsel

I’ve been trialling Thomson Reuters’ CoCounsel. I had another contract that needed to be terminated, and there was a very complex set of clauses to determine if an early termination fee would apply, and I was working out what that would be. I found the tool very helpful and accurate in locating the correct provisions and applying the provisions to our situation with a bit of guidance and context to work out the fee quickly. If this was a regular part of my work, it would save me a lot of time and give me confidence that I wasn’t missing anything.

Screens

I’ve also recently trialled Screens.ai, which is a contract review tool. I asked it to review several SaaS contracts that we were drafting or reviewing against suitable playbooks. I just used community playbooks, so they weren’t a perfect fit for our situation to begin with, but they were close enough. While the tool identified several “issues” that I didn’t care about, I found it helpful to consider them and think about why they didn’t matter in this case. It also flagged - and marked up with comments - several issues that I did care about, and it was helpful to see those as part of my detailed review. I felt more confident in my review having used this tool. 

In summary

There are so many tools out there that it can be hard to know where to start. It’s also easy to worry about what AI can’t currently do and decide to wait because it will get better “soon”.

The risk in this approach is that you’re missing out on the benefits of what AI can already do for you, which is a lot. Even with the things it can’t currently do, there is value in understanding that because it helps you to know what to look for amongst all the noise.

I’m looking forward to being part of a panel at the upcoming ILANZ Conference from 7 to 9 May 2025 on the Use of AI for in-house teams, where we’ll be delving into this topic more. If you haven’t got your tickets, you can register here: https://ilanz.org/conference-2/register-3