Building your legal team’s AI skill set
Key insights
- Take a people-first approach to building your team’s AI expertise
- Start out by teaching what key terms mean. Make sure teams understand the definitions of fields and sub fields, such as AI, generative AI and machine learning
- Set ‘golden rules’ for AI, which explain when AI tools can be used at work – and which tools can be used
- You could highlight your team’s ‘tech champions’ who can recommend the best ways of using these tools
Using the right technology tools to perform your client's instructions is increasingly a part of the skill set the modern lawyer needs. That means we need a training system which caters to that.
Imagine that a relatively stable piece of software, such as your work email platform, was gradually replaced by something else.
There would be training about where to click, how to send an e-mail, what the button for attachments looked like and how to save a file to a document management system. We would be given a fixed checklist of training items to allow us to understand this new system and why it is an improvement.
But when it comes to generative AI chatbot tools, it’s not just about understanding which ‘buttons to press’.
A new approach to technology training
We’ve only very recently been able interact with a series of algorithms that can generate human-like text about almost anything on demand. Generative AI is a new field to navigate.
That means we need to teach people how to understand the technology so that they know what they're working with.
We need our teams to develop the skills to interact with AI platforms – to delegate tasks, to describe tasks effectively, to discern whether an answer is useful and to spot potential problems.
And we also need to make sure everyone exercises diligence – using these tools in a way that meets our professional, ethical and client obligations.

Iain Murdoch, Head of Legal AI, Mills & Reeve
Setting the golden rules of AI use
At first, the complexity of the tools can feel quite daunting.
If you’re new to AI, start out by learning a little about the underlying technologies. Find out what we mean when we talk about the broad spectrum of AI, the narrower subfield of generative AI, and the slightly broader subfield of machine learning.
At Mills & Reeve our 2030 strategy includes technology– but also a firm focus on our people. We say that clients are at the heart of our business – and we mean that.
Our technology strategies are about equipping our people with the right knowledge, the right information, and the right technologies to serve our clients better.
We’ve set golden rules for generative AI, which explain how and when AI tools can be used at work.
For example, one of the golden rules is ‘don't use generative AI to produce a client output if the client's asked you not to use generative AI to generate it’.
We have a list of approved software platforms that incorporate generative AI. Only those approved tools can be used for work.
Using a memorable name like ‘the golden rules’ helps those guidelines to stick in people's minds – and it means they’re easy to find when team members need to double check an approach.
We need to allow our people to flourish in an environment where, for the first time, they can delegate tasks not only to a human colleague, but also to a ‘silicon’ colleague.
We've never lived in that world before.
Find your tech champions
At our firm, we have a technology adoption team who train people on effective AI use, different features, and how AI can be used in word-processing or spreadsheets, for example.
We work with the providers of some of our technology tools for training. But we also have our own internal ‘tech champions’ to recommend the best ways of using those tools.
For example, our real estate team has access to a couple of specialist tools – on our approved software providers list – that include generative AI features. There are tech champions and innovation lawyer specialists within that team who can really promote their use within that context, to meet clients’ needs.
Everyone's got access to a wealth of information on our intranet, and we also offer both in-person and online events.
We are developing the next round of events as we speak because we see it as a continuing process. I think that the teaching and sharing of best practice around generative AI use is going to be an ongoing process for the foreseeable future because it’s evolving so quickly.
It won’t be a one-off ‘big bang’ approach. It can’t just be a matter of saying, “here's a new piece of software and here's how to use it”.
That's why I believe in a skills-based and people-centric approach. It's about building up our people's confidence and awareness.

Iain believes that a skills-based and people-centric approach is crucial in an effective AI strategy.
Positive mindsets
My advice to lawyers is to be deliberately optimistic.
Acknowledge the real risks of bias, hallucination, or error. But recognise the opportunity to deliver work that might not have been possible before.
For example, you might be able to better meet the needs of a charity client, by giving them a level of detail that might have previously been unaffordable. Those opportunities are equally real.
You can be sceptical – and it is right to be sceptical about claims that present generative AI as the solution to everything. But avoid cynicism. Lean into the technology, understand and seek the benefit from it.
Key terms
Artificial intelligence: The theory and development of computer systems able to perform tasks that usually require human intelligence, such as visual perception, speech recognition and decision-making.
The data used to train an AI system is referred to as the input data, and the results produced by the system as the output.
Machine learning: A subset of AI that enables computer algorithms to evolve and refine their performance by processing and learning from data over time.
Generative artificial intelligence: Generative AI is a subcategory of AI that uses deep learning algorithms to generate novel outputs based on large quantities of existing or synthetic (artificially created) input data.
These outputs can be multimodal such as text, images, audio or video.
Read more in our Generative AI: the essentials guide glossary of terms.