In Conversation With Sarah Harris: Innovation, AI And The Future Of Regulation

In conversation with Sarah Harris: Innovation, AI and the Future of Regulation

Sarah Harris has had a unique career journey, moving from the Bar to regulatory law and now leading innovation and knowledge at Kingsley Napley. In this interview, Melinka Berridge, Head of Regulatory Law, speaks with Sarah about her role, the development of the firm’s AI assistant KNavigate, and the opportunities and challenges AI presents for regulators and legal professionals alike.

You’ve had a fascinating career journey – from barrister to regulatory lawyer to Director of Innovation and Knowledge. What was it that drew you in that direction, and what sparked that transition?

I am naturally quite a playful person, and so I was drawn to a role that encouraged others to be creative and get out of the day-to-day for a while. We spend a good proportion of our time at work, so it is important to me to find joy in that as much as possible. People are at their most creative and open-minded when they are relaxed and given space to think.

With my work in the regulatory sphere, my key remit for regulator clients was how to best harness technology for their processes and stakeholders. My broader innovation role is exactly that but at scale and across different areas. The added complication is the rapid development of new technology, which makes for an interesting ride!

One of the most exciting developments you’ve brought to Kingsley Napley is KNavigate – the firm’s very own secure AI assistant. With so many AI tools already out there on the market, what led you to build something bespoke for your colleagues here?

Having been at KN for over 13 years, I knew first-hand that the way our people think and operate is special. Our clients come to us because we approach their cases in a certain way, with independence of thought and courage. Those workflows, and that approach, deserve an AI strategy that respects that authenticity and difference. Buying generic, off-the-shelf products didn’t feel right for us. We didn’t want to hand the way we work over to others. We also wanted to learn how to do legal AI. I believe AI is a contact sport and if you are not learning about how the supply chain actually works (and how much it actually costs) by outsourcing it, you are missing a big part of the game.

With your background as a regulatory lawyer, you bring a really unique perspective to this role. How do you see KNavigate making a tangible difference to the services we provide to our regulatory clients?

Our KN regulatory team now have a safe space (and the skills and training) to start testing out what tasks AI can help them with. They have an appreciation of the benefits and drawbacks of AI and are well-equipped to understand how this technology can be harnessed for good. I think key areas such as redactions, finding a needle in a haystack of documents, transcribing witness interviews and reviewing large swathes of evidence are all key use cases that the regulatory team will utilise, leaving them with more time to spend speaking to witnesses, making expert judgements on legal issues and training their colleagues.

Our regulatory clients work incredibly hard with the resources they have, and keeping cases moving forward efficiently is always the priority. So, it’s a real concern when AI-generated communications, rather than genuinely valid complaints, start to pull their teams away from that vital work. What’s your advice to regulators on communicating clear boundaries to the public, so that AI becomes a help rather than a hindrance to their operations?

The difficulty is that AI-generated communications and genuinely valid complaints are not mutually exclusive. There may be some very valid complaints hidden in AI-muddled (or potentially even exaggerated) communications. I think regulators need to harness the same technology to address the issue. They should explore how they themselves could safely use AI to triage complaints, identify the genuine issues and draft communications that ask pertinent follow-up questions to distil those issues. This is something we can help clients with collaboratively, so please do reach out.

If you could point regulators towards one straightforward, practical ‘easy win’ when it comes to adopting AI, something they could act on today, what would it be?

I would say not to just look at the end of their noses, i.e. enterprise AI. There are plenty of providers out there offering enterprise-grade, secure AI. I would advise them to get expert advice on the key critical capabilities their AI needs, and then go out and have some conversations. Make sure someone owns it. If it is not someone’s job to do it, it won’t get done.

Looking ahead, if you could wish one development into existence for the regulatory sector’s use of AI over the next five years, what would it be?

It would be to harness AI’s pattern-finding ability to better understand what is happening in practice and feed that back into the sector through training, much more quickly. Rather than waiting for FTP allegations and analysing those (although even that would be helpful), it could identify issues right at the outset. Logging competence and capability issues in an anonymous, GDPR-compliant way could help identify training needs within a particular Trust or area, allowing those insights to feed directly into training and recruitment. It could significantly shorten and amplify the feedback loop across the sector.

And finally – we couldn’t let you go without asking – do you have an AI joke for us?

What was the AI drummer’s favourite beat? An algo-rhythm.