5 ways to make AI work for your legal team
AI continues to generate plenty of discussion in legal teams. The harder question is how to use it when workloads are already high, expectations keep rising and resources remain tight.
That challenge came up repeatedly at the recent Running Legal Like a Business event in Las Vegas. As Lacey Near, LOD's Global Lead, Key Accounts, put it: "Legal teams don't need another legal transformation programme competing for time they don't have."
Making AI useful isn’t simply a tech question. Legal teams need to look at how work is organised, who does it and where tech can genuinely help. It will add a little more work initially, but for long term gains.
Here are five practical ways to build, or utilise your existing, legal team to make AI work for you.
1. Start with the work
Before looking at the tech, understand the work coming into the team. Are you answering similar, if not the same, question multiple times? ? When should it come to Legal? Does it genuinely need legal judgement, or could it be handled differently? Perhaps you are reviewing a significant volume of contracts and rarely deviating from your standard terms, or it could be pre-and post-merger support where AI can review contracts for post-merger alignment and playbook updates.
Looking at demand can highlight opportunities for guidance, self-service or different ways of resourcing work. It can also show where repetitive drafting, review or information gathering is taking up time better spent elsewhere. Understanding that picture before investing in tech makes it much easier to focus on the problems that matter most.
Make a start: Choose one category of work and map it from request to completion. Look for delays, repeated effort and unnecessary hand-offs.
Tip: An easy place to start is any process you already have playbooked (which is why commercial contracting is a common win). You will need to spend some time making the playbooks compatible with your tool of choice. This can be drafted by more junior team members or consultant lawyers, before review and approval by your leadership team.
2. Create a clear route into Legal
For many teams, work still arrives through email, Teams, meetings and informal chats. It feels easy for the business but it makes demand difficult to see, harder to prioritise and almost impossible to measure
Without a clear route into Legal, it’s difficult to understand what work is coming in, where the pressure is building or whether resources are being used effectively. A clear intake route gives the team better information from the outset and helps ensure work is directed to the right place. It also creates the structure needed for self-service, workflow tools and AI to be genuinely useful.
Make a start: Agree the minimum information Legal needs and create one simple way for the business to provide it. If the process is difficult to use, people will work around it.
Tip: This can be as simple as reiterating to the business to use your legal email alias. Draw firmer lines around where work will be accepted and you’ll soon minimise the noise of multiple avenues for receiving work.
3. Solve a specific problem
“Let’s use AI” isn’t a useful brief. Which task takes too long? Where is effort being repeated? What would make a visible difference to the team or the business?
The first use case doesn’t need to solve everything. A focused improvement can remove a genuine frustration, build confidence and help the team learn what works before it makes a bigger commitment. Is it automating repeat contracts like NDAs and simple commercial agreements, or perhaps you need to update all contracts with new regulatory clauses, and AI could identify which contracts are yet to be updated.
Make a start: Pick a task with a clear beginning and end. Decide in advance how you’ll know whether the change has helped. If you can’t measure the improvement, it’ll be difficult to know whether it was worth doing. Having clear precedents or examples of good outcomes will also make it easier to assess the results and build confidence in the approach.
Tip: Look for work that follows a repeatable pattern – even if it’s not the whole process, but a recurring part or stage of the process. The more consistent the process you identify, the easier it is to test, refine and improve.
4. Create capacity for implementation
AI may save time, but getting there takes work. Processes need to be understood, information needs to be prepared and outputs need to be tested. Teams also need to agree where human judgment is still required and how exceptions will be handled.
This effort is easy to underestimate. When it’s squeezed around already full workloads, progress slows and the results will be harder to trust. A narrow scope, realistic timetable and clear testing criteria make progress more manageable.
Where the in-house team can’t create that space alone, bringing in additional support for a defined project or workstream can provide the breathing room to move forward without losing momentum on BAU work.
Make a start: Choose a contained use case, agree who needs to be involved and test it against real examples, especially the awkward ones.
Tip: Keep the testing group small but include relevant team members with a range of experience levels.
5. Design the team around the work
Not every step should be automated. We see legal teams get the best results when people, processes and tech work together in a way that reflects the work being done. Tech may support an initial review draft, while legal professionals apply the organisation’s standards, deal with exceptions and make the final judgement calls.
Lacey shares: “The goal is to match the right resource, process and tech to the work as it is now, with clear ownership and oversight – and all these elements should evolve with your business, in line with your needs. Once your process/es are mapped, you’ve built the foundation upon which resource, process and tech can be applied, and iterated, as your business needs inevitably evolve.”
External specialists, and Structured Teams, operate within this principle. Both are, at their core, flexible models, built on agreed process and outcomes – delivering what you need, even as those needs change.
Make a start: Write down what the tech will do, what a person will check and who makes the final decision. If that is difficult to explain, the process is probably not ready to scale.
Tip: Start with the outcome you want to achieve, then decide which tasks are best handled by AI, which require human expertise and where oversight is needed.
What next?
There’s no single formula for building a legal team that can make AI work. But we’re seeing the teams making the most progress tend to start in the same place. They work out where their time goes, what genuinely needs legal expertise, and where work could be handled differently.
Tech has a role to play, but it’s only part of the answer. The bigger challenge is deciding how legal work should be delivered in a way that is sustainable for both the team and the business.
If these are questions you’re already grappling with, we’d love to share what we’re seeing across the in-house legal market.