16-17 June 2027 – London, InterContinental O2 | Magazine
Partner
A&O Shearman
Partner
A&O Shearman
Theatre is buying licences, counting logins and talking earnestly about the future of law.
Transformation happens at the level of data, people and use cases. For us, that means changing how work is actually priced, staffed and delivered. For clients, it means finding new ways to deliver value to their internal stakeholders, whether through their own initiatives or by changing how they work with external vendors.
The test is depth, not breadth. One or two use cases embedded properly into daily work, and built so they can scale, are worth far more than fifty shallow pilots. And the hard part is rarely the technology. AI transformation is a change-management problem more than an IT project. Persuading clever, busy and sceptical professionals to abandon habits built over decades, and to invest time rethinking something that already works, is where the real effort goes.
AI adds genuine value where the work is high-volume and pattern-rich, and where a lawyer’s time currently goes on reading, finding and typing rather than deep thinking. Think of covenant checks across a portfolio, clause extraction across thousands of contracts, first-pass review of due diligence, or first drafts built from term sheets and precedents.
Each of those tasks is different, and the value compounds when you design around a specific legal workflow. Before we take on a use case, we apply three tests: does it require legal judgement that a generic tool cannot replicate; can our lawyers’ know-how, precedents and market insight make the output materially better than an off-the-shelf answer; are we comfortable with how the AI will perform and do we have lawyers in the loop at the right places? If the answer to any of those is no, it isn’t a use case for us. The expertise is the answer; AI is what makes the expertise scalable. When you combine that with a client’s own governance parameters and commercial playbooks, the system delivers actionable insight rather than just extracted data. This is not “what we did before, but a bit faster”. It is something that was not possible before.
DORA – the EU’s Digital Operational Resilience Act, which sets resilience and oversight requirements for financial services – has been a useful forcing function. One financial services client came to us with several thousand ICT and outsourcing contracts and a hard regulatory deadline to meet the Article 30 contractual requirements.
We used AI to triage that population by criticality, extract the existing provisions on subcontracting, audit rights, exit, incident reporting and the rest, and map each contract against the DORA requirements clause by clause. We then overlaid the client’s own internal policies and their prioritisation of each relationship.
The output was not a due diligence report. It was a renegotiation plan, sequenced by risk and by the criticality of each contract, together with tailored amendment agreements. Instead of a mass of contractual terms, the client had something it could use to make decisions on risk, budget, staffing and regulator engagement.
And this is not limited to one-off remediation projects. The same approach applies to business-as-usual contracting. We have a client growing a particular business line, where the challenge is balancing client acquisition and speed of onboarding against risk management and regulatory compliance. By turning their portfolio into actionable data, and applying market benchmarks alongside their own playbooks and policies, they move from a stack of documents to information they can use for commercial decisions, to streamline ongoing negotiations, and for compliance monitoring and default planning. The information takes on a different value, and legal becomes an enabler of business growth.
Accountability has to be human and named. A model cannot be accountable; a partner, a general counsel, or a reviewing lawyer can. That is exactly why how the technology fits into real workflows matters far more than which model you happen to be using. We ground outputs in trusted precedents and commercial playbooks curated by our own lawyers, which reduces the risk of false or misleading results. The system produces a structured starting point that a lawyer then validates and tailors in context, with governance, auditability and human oversight built into every step.
“Good enough” is not a fixed standard. It is calibrated to the stakes, and for us it means being transparent with clients about the role of AI and the balance of cost and risk. For a first-pass review of 10,000 contracts, 95% accuracy with human spot-checking may be more than good enough. For the high-stakes transactional and advisory work that makes up most of our business, the AI is a drafting aid and never the decision-maker, and a lawyer owns the final output entirely.
The mistake is to ask “can we trust the AI?”. The better question is: “what is the consequence if it is wrong, and have we built the controls to match?”.
That is a real risk, and one we spend a lot of time on, but it is solvable through governance, controls and training. If you look at the recent public examples, things went wrong because of a failure in those things, not because of something inherent in the technology. And it is not a genuinely new risk — AI is simply a new piece of an old puzzle.
Talent and training. AI will start to dismantle the apprenticeship model through which junior lawyers have historically developed judgement and expertise. Remove that, and if you do not find another way, you get a generation who can prompt a model beautifully but cannot tell when its answer is subtly, dangerously wrong.
We need to be redesigning how lawyers learn now, deliberately, rather than discovering the hole when we fall into it. So we are rebuilding our training in two directions at once: teaching lawyers to learn by reviewing and challenging AI output, and making the training itself more deliberate and less by osmosis. We are also building systems that capture the knowledge and expertise of our most senior lawyers. The real question is how we make sure there is a next generation good enough to keep that knowledge alive and up to date.
The future of law is at LegalTechTalk – events for legal transformation
The legal industry is undergoing a major transformation, driven by technological advances, changing demographics, and new business models. Lawyers who are able to adapt to these changes will be well-positioned for success in the future.
Upcoming Events:
EUROPE · 16 – 17 June 2027 →
USA · 13 – 15 December 2027 →
Get news and insights delivered straight to your inbox!
Join 5,000+ legal leaders shaping the future of law for two days of insight, innovation and unmissable networking — now with your biggest saving!
