How solicitors are using AI tools for contract review

Your digital documents might have arrived well-structured and well-ordered, with file names that make sense.

Or they could arrive as a bunch of numbered scans. In that case, you might first use a machine learning tool to extract the date and the parties from those scans, and if there are titles, what the document is.

You might use that same tool to understand and group the contracts – maybe there are 50 employment contracts and 50 commercial agreements in that bundle.

Even just dividing them into two electronic piles is valuable. For example, you could seek instructions from your client and decide to only review the commercial ones, or delegate each pile to employment and commercial specialists.

What are the potential benefits of using AI tools in contract review?

It saves you time and boosts efficiency – and it could also help enhance how you deliver your services to clients.

In the past, to meet your clients' cost expectations on a significant transaction, you might have worked on a sampling basis, for example only reviewing 10% of employment contracts, leases or commercial agreements.

Now, using technology as an ‘assistant’, you can potentially review them all while still aligning with price points.

AI could also help improve your internal contract processes. You could, for instance, automate reminders of key contract expiry dates, rather than having to manually add them to your calendar

Today, documents are more likely to arrive as digital files – rather than the paper bundles of the past – and technology can be used to help review them.

What’s one thing you wish more people knew about using AI?

When GenAI gets bad press in the legal sector, it’s often due to hallucinations. In some cases, hallucinations can pose a real risk.

Depending on the context, a hallucination isn’t always bad. Sometimes, you are looking to creatively analyse a unique or unusual situation. If you’re using AI to generate new solutions to your client’s problem, or to provide a fresh perspective, hallucination could be a useful feature.

But don’t assume that GenAI is the right tool for everything. GenAI is built to create new text and new information, so if you’re looking for a strict data extraction, there will be an alternative machine learning tool available that's specifically for that purpose.

What are the risks of using AI for contract review?

We’ve covered the risks that hallucinations bring, and making sure that you know which AI tool is best suited to each use case.

A big risk is using a public GenAI chatbot. Your data won’t be protected in the same way as if you’re using one on an enterprise level that is sanctioned by your organisation and has gone through an established risk assessment and approval process.

Iain Murdoch, Head of Legal AI at Mills & Reeve.

What would you say to someone interested in investigating this as a use case?

Beware of the hype. There’s a flood of sales pitches on AI tools for lawyers on LinkedIn, on emails – all proposing solutions.

But you don’t want to end up implementing a ‘solution’ that doesn’t address your real problems. You must interrogate whether the tech company’s claims align with the challenges that you, your firm, or your in-house team, face.

My go-to questions are, “What does my client need from me? What is my client trying to achieve?”

You should explore AI implementation while considering good business practice, not just technology-specific practice.

Is it wise to go to a newly established startup to service one of your business-critical practices, even if it has a wonderful offering? Perhaps not. If it’s a core need, you might opt for a more established player.

But a new business area may offer a lower risk opportunity which you can use to explore a new provider.

What are your predictions for these tools?

The legal technology market moves at pace. Leading players are reacting to each other's new products very quickly. Recently, a new legal plug-in for an existing AI tool was announced, which uses prompts and workflows designed for a specialised legal context. On the same day another technology provider’s shares fell by 18%.

These platforms, I believe, will continue to innovate. The drive for growth will be very intense as they are funded by private equity who will be looking for a return on their investment.

But not all specialist providers are targeting large firms. There are opportunities to engage with these technologies, whatever size firm or practice you’re working in.

We’re also hearing more about the development of ‘agentic AI’ systems which will take a human instruction and then act autonomously to break that task down into components, and send different AI agents off to complete them. These could plug into other tools – for example, having access to certain documents, to the web, and other software services.

Remember that a public GenAI chatbot won’t protect your data in the same way as one on an enterprise level that is sanctioned by your organization.

Where should I start?

Review your existing systems and tools and find the ‘jagged edges’. In other words, define the things that they’re already great at, and the things that they’re less good at.

Only worry about the things they can’t do right now if that’s a particular need for your business – in which case, go off and explore that.

If you are at a firm dealing with volume work, there may be great opportunities to integrate elements of AI and GenAI into those workflows.

I don't think that you can just ditch an existing workflow and say, “I can just ask an AI assistant instead”.

Often the value is embedded in the workflows that you've developed through your existing practice management system and through your people. Bear that in mind – and don't throw the baby out with the bathwater.