The future of legal isn't fewer lawyers. It's fewer decisions requiring lawyers.
The best legal departments aren't using technology to replace people. They're using it to remove friction, automate repetition and create space for the work that needs human attention.
That sounds simple until you try deciding where the boundary lies.
Every new system promises efficiency. Every AI tool promises scale. The difficult part isn't implementation.
It's knowing exactly when you should.
People tend to imagine the GC appearing near the end of a project carrying a red pen and firing off a broadside of inconvenient questions.
In reality, there's much more to the role than that.
During his time at JLL, Geraint moved between commercial strategy, operational firefighting, regulatory change, training, technology and occasionally the sort of problems nobody expected to crop up.
Legal may dig up issues that the business didn't know it had. More often, issues arise that no one in the business expected and it's then for legal to resolve.
That's because companies rarely experience problems in neat legal categories.
They experience business problems with legal consequences.
The best in-house teams understand that distinction instinctively. Their value isn't measured by how many contracts they review. It's measured by how many problems never become contracts in the first place.
That requires something technology still can't automate: judgement.
AI dominates almost every boardroom conversation.
The temptation is to see technology as a substitute for experience. Geraint sees something rather less dramatic.
Contract review becomes faster. Admin work becomes cleaner. Repetitive requests are automated. Intake systems force stakeholders to provide the information legal actually needs instead of rattling off an email containing approximately half of it and leaving them to figure out the rest.
Progress, in other words, often looks suspiciously mundane.
These kinds of improvements rarely make headlines. They do, however, free lawyers to spend more time on decisions that genuinely require experience.
Nobody hires a senior commercial lawyer because they're exceptionally gifted at migrating information from one system to another.
The catch is that efficiency creates its own risks.
Hallucinations remain real. Governance still matters. Today's market leader may become tomorrow's abandoned platform.
Geraint compares today's legal AI market to the old Betamax versus VHS battle. The technically superior product doesn't always win and predicting the eventual survivor before the dust settles can be an expensive speculation.
The smartest GCs aren't trying to predict the future.
They're trying to preserve their options.
AI has created a new army of armchair lawyers.
A customer who previously wouldn't have the determination or competence to file a complaint can now produce something that reads like a junior lawyer's first draft. AI has democratised legal access.
That changes the volume as much as the quality. Firms that once received ten carefully drafted letters now receive a hundred sloppier ones. Businesses also face better complaints and more sophisticated contractual challenges from the professionals.
They inevitably respond with AI of their own, reviewing, triaging and often generating replies. Before long, machines are writing to machines while humans supervise from the sidelines.
Ironically, the easier AI makes written communication, the more valuable human relationships become.
When all communication sounds as though it's been run through AI, the ability to write well stops being a differentiator. Authenticity becomes crucial for trust.
Technology hasn't removed the need for judgement.
It's simply moved where judgement is exercised.
One of the more revealing observations isn't actually about lawyers at all.
It's about property.
Most people think about buildings in terms of what they see. Houses. Shopping centres. Offices with a city skyline painted on the wall.
The interesting work happens underneath.
Planning, financing, sustainability, construction, leasing, facilities management, tenant relationships and operational risk all sit beneath the visible surface. Property isn't a single industry. It's dozens connected to one asset.
Technology has a habit of making difficult problems look deceptively easy.
Property is a useful antidote.
Buying stocks and shares takes seconds. Buying a building still takes months. Lawyers, surveyors, lenders and regulators remain a stubborn drag. Plenty of clever people have hailed tokenisation and fractional ownership as the silver bullet to this liquidity problem.
The underlying purchase process is still frustratingly slow.
Geraint highlights a forgotten topic.
Sustainability.
Organisations can't claim AI is central to their future whilst treating its environmental cost as somebody else's problem. Data centres consume extraordinary amounts of electricity and water. Those costs don't disappear simply because they're outsourced.
AI and sustainability are discussed in separate rooms.
They probably shouldn't be.
Every new AI deployment carries an environmental cost as well as a commercial benefit. Judging one without considering the other feels increasingly difficult to justify.
Legal has always been good at managing contradictions.
Although this one feels unusually significant.
Which brings us back to the beginning.
The future probably doesn't belong to the organisations using the most AI.
It belongs to those exercising the best judgement about where technology genuinely improves decision-making and where human experience remains irreplaceable.
AI may draft the contract.
AI may review the contract.
The question is whether the contract ever needed drafting in the first place.