I’ve noticed that organisations often overestimate how much of their AI journey is about technology.
The conversations tend to focus on models, integrations, security reviews, and vendor selection. Those things are important, but after spending the last few years watching legal teams’ experiment with AI, I’ve become convinced they’re rarely the deciding factor.
What matters more is the environment AI is entering.
Throughout this series, I’ve argued that legal teams shouldn’t think about AI primarily as a technology decision. AI is a tool, but the value it creates depends heavily on the quality of the legal operation surrounding it.
An AI model relies on organisational context, institutional memory, and sound governance, rather than producing it. It can support judgment, too, but only where the foundations beneath it are already strong.
That’s why, before debating which model to use or which platform to adopt, I think there are a few more fundamental questions worth asking.
Not to determine whether AI works at all, but to help decide whether it creates clarity or amplifies the problems you already have.
The first question is surprisingly simple.
Can your team consistently find what it already knows?
Most organisations would confidently say they have the information they need. Contracts exist. Matter records exist. Email conversations exist. Negotiation history exists. Somewhere in the organisation, the answer is usually there.
The more difficult question is whether that information can be found quickly, consistently, and with confidence.
If a lawyer asked why a particular customer was given an exception six months ago, how easy would it be to reconstruct the answer? If a new team member needed to understand the history behind a risk position, could they do it without relying on institutional knowledge held by a handful of people?
These questions matter because AI tends to expose information fragmentation rather than resolve it.
One of the recurring themes throughout this series has been that context matters. The organisations seeing the greatest value from AI are rarely those with the most information. They’re the ones that have made their information usable.
The second question is the harder one.
Do you capture decisions, or just documents?
Legal teams are generally very good at retaining outputs. Contracts are stored. Policies are updated. Approval records exist. The outcome of a decision is often preserved long after the decision itself has been forgotten.
The challenge is that the most valuable legal knowledge often sits behind the outcome.
- Why was the exception approved?
- Why was the commercial risk considered acceptable?
- Why was one customer treated differently from another?
The longer I work with legal teams, the more convinced I become that institutional memory is one of the most underappreciated assets a legal function possesses. Yet it’s also one of the easiest things to lose.
People leave. Teams change. Priorities evolve.
The document survives. The reasoning often doesn’t.
That becomes increasingly important as AI enters the picture. An AI system can learn from decisions that have been captured, but reasoning that was never documented is simply gone.
Being well positioned for AI has less to do with the size of your repository than with whether you’ve preserved enough context to explain how important decisions were made.
The third question is the one I suspect will become increasingly significant over the next few years.
Could you explain your AI-assisted decisions to somebody else?
For much of the current AI discussion, attention remains focused on outputs. Is the answer accurate? Is the summary useful? Is the draft good enough?
Those are reasonable concerns. But as AI becomes more deeply embedded into legal workflows, organisations will need to think beyond outputs and towards accountability.
If the General Counsel asks why a recommendation was made, can the reasoning be understood? Or if an executive team challenges a decision, can the supporting context be explained? And if a regulator, auditor, or customer asks how a particular conclusion was reached, is there sufficient visibility into the process?
Those are governance questions more than technology ones, and governance is ultimately about trust.
The legal teams that benefit most from AI are usually those that have invested in understanding how decisions get made, how knowledge is managed, and how risk is governed inside their organisation.
That’s because AI tends to amplify whatever already exists, making strong legal operations more valuable and weak legal operations ones more exposed.
The technology itself is usually doing exactly what it was asked to do. The outcome depends on the environment surrounding it.
Looking back across this series, that’s probably the single observation I keep returning to. The conversation about AI often begins with models, features, and tools.
In practice, the organisations creating the most value usually end up talking about knowledge, workflows, governance, and decision-making. And while the technology matters, the operating model matters more.
That’s why readiness begins with understanding how your legal function operates today. The tool choice comes later.
If you can answer those three questions with confidence, you’re probably much closer to being AI-ready than you think.
If you can’t, that’s where I’d start.
This is the final article in a six-part series on what it takes to get a legal team ready for AI.
As the tools continue to evolve and the models become more capable, new approaches will emerge. But I suspect the legal teams that benefit most will be the ones that treat AI as an amplifier of a strong legal function, rather than a substitute for one.
This is the sixth of six articles on what it takes to get a legal team ready for AI. Click here to read the previous article, ChatGPT is free the way a shared drive is free.
