Tom Conti-Leslie is a full-stack software developer at Luminance, a legal A.I. company, in Cambridge, U.K., and former project manager for Nightline France, a charity working
for better mental health support and advocacy for students in higher education in France. He graduated from the University of St Andrews with a MMath in mathematics.
What was your experience like in St Andrews (e.g. societies, extracurriculars)? How did your background in Mathematics help you decide to go into A.I. software development?
I loved my time in St Andrews, both in terms of the academics and the extracurriculars. Within the School of Mathematics and Statistics, I specialised in fields including analysis and dynamical systems, and in my final year I was even able to participate in research into computational algebra. This was my first introduction to software at an industry level and it played an important role in my decision to take on software development as a career after graduating. Outside of lectures I was involved with several societies and student groups, including musical theatre, Nightline, the Doctor Who society, the maths society, and Ukelear Fusion, the ukulele band. I don’t consider that exact combination to have funnelled me into a career in tech! Rather, I value each of them for the unique and varied experiences that they gave me; they made my university experience well-rounded and allowed me to meet some amazing people. That said, I do remember taking particular pleasure in incorporating technology into the roles I held within those groups; helping to maintain websites, writing data extraction scripts, and even making a ukulele chord diagram generator. The first job I took on after leaving St Andrews was actually working for Nightline France in Paris, an extension of the mental health work I’d done at university. I was in charge of improving French students’ access to information regarding the mental health services available to them. While the job was fulfilling, I realised that I had a desire to do technical work more regularly, and that’s when I started with Luminance as a software developer.
Can you tell me a little more about what Luminance is? What are the legal A.I. capabilities the company seeks to achieve (different products/target areas that Luminance provides)?
Luminance is an industry-leading legal A.I. company that automates the generation, negotiation and analysis of contracts for over 500 customers around the globe. Luminance Diligence, the company’s first product, is tailored to the contract review process; capable of ingesting thousands of documents, it can pick up on patterns which link the contracts that it sees, and help teams collaborate to review a body of documents together. Luminance Discovery is a tool which can assist companies with investigation processes, conceptually scanning and filtering across items such as contracts, but also forms and email archives, to find important information during tasks like Early Case Assessment and data regulation compliance. Finally, Luminance Corporate accelerates the contract negotiation process, allowing both legal admins and other business professionals, such as sales users, to track the evolution of a draft contract from its creation all the way through to signature. Luminance Corporate includes a powerful Microsoft Word plug-in which can apply A.I.-powered Traffic Light Analysis to an incoming contract – this means it will highlight risky areas in red, clauses potentially in need of review in amber, and acceptable areas in green . At the core of each of these products is a common A.I. engine which is capable of reading, understanding and analysing any legal document, extracting key information, and classifying clauses. Every document uploaded to a Luminance system can be visualised in our application overlaid with its A.I.-extracted tags, which include everything from clause types to contract parties, identification numbers, currencies, dates, and locations. If a legal concept isn’t supported out of the box by our pre-trained models, users can add a custom one simply by
tagging a few examples, and then the A.I. can begin to recognise and tag instances of that concept independently in new contracts. Around this core A.I., the company has built a range of tools to make it easy to visualise and explore a corpus of documents, alert users when termination dates are near, generate custom reports, bulk-redact sensitive information from contracts, and plenty more, all through the same user-friendly application.
What does a typical day as an A.I. software engineer look like? Have there been any memorable moments or breakthroughs during your time at Luminance?
I have worked in a number of different software areas since joining the company, but most of my time has been with the Application Engineering team, whose role is broadly to develop the interface between Luminance’s users and our A.I. engines: everything from the buttons in the Luminance web application, to the database which stores the contracts that users upload, to the servers which extract text and other important information from those contracts and pass them to our A.I. for analysis.
My work on an average day is typically split between developing new features for our Corporate product, and fixing small issues with our existing features when they are reported by our QA team, or by our customers. We have a high standard for the quality of our code and are constantly improving and testing our product; we also have a fantastic Customer Support team who are able to quickly address any issues encountered in the field. I am also involved in the company’s code review process, through which every change to our software is looked over by a second pair of eyes, a standard industry practice to ensure quality remains high and potential issues are caught early. A few times at Luminance, I’ve had the chance to work on larger features which introduce completely new functionality to our software, and I’ve found those particularly memorable and rewarding. Experimenting with new ideas brought forward by our Product team is a great way to ensure that Luminance’s software brings the maximum possible value to our customers and saves them as much time as possible in their contract review and negotiation process. Some of these larger features are now part of the product’s standard offering—one example is fully customisable “Knowledge Banks”, which is our software’s representation of a legal playbook, allowing businesses to set the specific terms under which a draft contract should be accepted, with preferred wordings and acceptable/unacceptable ranges for entities such as the contract’s term period and liability cap.
What is your opinion on how A.I. will change legal work? Do you think it will be an overhaul or more of an assistant to lawyers?
To a great extent it already has. I was speaking to a friend the other day who is training as a lawyer, and they told me that the work of junior lawyers joining firms today has completely changed because of the introduction of A.I. tools; where much of the work used to be manual review of boxes and boxes of documents, now their job is A.I.-enhanced from day one. There’s the potential for these sorts of tools to overhaul the work entirely for sure, but it will never replace lawyers entirely. Rather I think it will allow legal professionals to spend more of their time performing high-skill tasks, while some of the repetitive work can now be automated. We have heard reports from our customers that our products have won over even the tech sceptics within their companies; just like Microsoft Word became a natural part of people’s workflows twenty years ago, I think that AI will become more normal to use within the legal profession in the next decade.
What do you think is something to watch in the legal tech sphere? (e.g. any news/ events/ people/companies to follow, upcoming regulatory decisions, future changes you foresee)
This year has obviously seen widespread interest in, and adoption of, generative A.I. (such as OpenAI’s GPT models) in all kinds of fields, and the legal profession has been no exception. For Luminance’s customers, our new A.I.-powered chatbot “Lumi” has been available for several months already, combining generative AI with Luminance’s proprietary legal AI engine to provide users with new insight into the contracts that they are reviewing. Other contract lifecycle management apps are beginning to feature similar integrations. While the world has been quick to adopt generative technology in many places, I think that we are just now beginning to discover the specific areas in which its presence can really make a difference, and to distinguish these areas from those where users might end up less interested in being able to augment their work with A.I..
Our chatbot, Ask Lumi, was initially available only through a generic chat window while reviewing a document. Since the initial launch, though, we’ve discovered that it is particularly skilled at redrafting clauses from a contract under negotiation in a way which reconciles your counterparty’s proposed wording, and your own company’s legal playbook. We’ve now introduced a more streamlined redrafting interface where our users can harness this ability in a single click. I believe that we can expect generative A.I. to really find its place in our workflows in this way in the months to come.