Lucy Frazer KC has seen the pressures on public institutions from the inside. A former Cabinet minister and twice Solicitor General, she served in six government departments, with ministerial roles across courts, prisons, probation, legal services and criminal law. Today she takes on an expanded role at Legora, leading our work with governments and public agencies, starting in the UK and EMEA. We asked Lucy about what she saw during her time in government, why she chose Legora, and what public bodies should expect from any AI supplier before they can trust it.
Which experiences in government most shaped how you think about the pressures on the public sector for those working in law, or law adjacent areas?
Government and public agencies face many challenges. Cases collapsing because of incorrect disclosure. Decisions in immigration cases being judicially reviewed because a step in the process had been missed. And significant backlogs in our courts.
I saw first-hand how the pressures on government were growing: the explosion of evidence in criminal cases, the rise in cases coming through our courts, and an overstretched prison system.
In government, we were looking to technology to bring new answers to old problems. In the NHS, for example, da Vinci robotic-assisted surgery was helping patients with prostate cancer, and predictive analysis was being used to bring down waiting lists. It was obvious to me that technology could fundamentally change every area of government, including the justice system.
What did you see in Legora that persuaded you to take on a bigger role with the company?
Three things matter to me: I want to work for a company that is bringing about positive societal benefits through AI. I want a can-do, positive culture, with dynamic leadership driving that change. And I want a company that truly works in partnership with its clients, listening to their needs and building its products on feedback and collaboration.
I see all of this at Legora. Its tools can fundamentally change the delivery of legal and legal-adjacent public services. Tools for legal research and factual analysis save significant time and cost, and more broadly they can bring order to disorder in the management of legal processes.
Its operating model relies on legal engineers, lawyers who work alongside the technical teams, so there is a real culture of collaboration with clients.
Under Max’s incredible leadership, Legora has come further, faster, than any company in legal technology. In a very short time, it has built a brand that law firms recognise and trust.
We want that same collaborative model to extend into the public sector, working in partnership with government to solve the problems it faces in delivering better, more streamlined services to citizens.
What are the biggest obstacles to modernising the justice system with technology, and what would give decision makers the confidence to move faster?
Many of our departments have legacy IT systems and those that work within government are understandably concerned about change and how to implement that change. We need to remind ourselves that not embracing technology also brings significant risk. What gives decision makers confidence is seeing it work for themselves. I would recommend that any public body tries a supplier before it buys. Legora included.
For the public sector, we say that trust is not a differentiator; it’s a precondition.
You’ve said the public sector needs to be confident in the AI it adopts. What does a public body need to see from a supplier before it can trust it?
For the public sector, we say that trust is not a differentiator; it’s a precondition.
The judiciary’s own guidance on AI sets a good benchmark: AI can support the work but must never make the decision, and every output has to be checked. Public authorities need to be confident that the tools they use protect this process, and to look for evidence that trusted organisations already use the tool, chosen through proper procurement.
At Legora, you can choose where your data is stored, and know that it is processed only for the purpose you define, and is never used to train models. There are full audit trails, role-based access controls, customer-managed encryption keys, ethical walls and isolation between matters.
AI can support the work but must never make the decision, and every output has to be checked.
What would you like to point to in a year’s time as evidence that the work is making a difference?
In a year’s time, I’d like to see significantly improved efficiency and productivity in every area where we are supplying the government, and that translates into real delivery for citizens.
Just last month, I was involved in a pilot with a UK regulator. They told me that a supervision data check that used to take two weeks took 10 minutes on Legora. Independent research points the same way. 65% of lawyers using Legora said work that used to take hours now takes minutes, and 71% said it picks up issues they would otherwise have missed. Among in-house legal teams, 97% said they respond to stakeholders faster. I’d like to see that sort of change replicated across departments and agencies.
