A client centred approach to AI – the path to enlightenment or the road to perdition for law firms?
Our Innovation Group considers the lessons learnt from our own AI exploration.
Often dubbed the Fourth Industrial Revolution, the increased use of artificial intelligence (AI), especially in the professional services sector, continues at a fast pace. Our starting point was creating an Innovation Group, led by a member of the firm's Board, to drive the internal agenda for technology and, externally, to own and seize the opportunities with a mandate and budget to run a number of proofs of concept.
Over the last 12 months, we have made some transformatory decisions in the life of our business. We have learnt much through research and as we have explored AI, we have seen numerous demonstrations and applications of AI that have been genuinely impressive. On the flip side, we have seen promises made by organisations' sales and marketing teams that outstrip the ability of the organisation to deliver.
We have learnt many lessons along the way, including:
- An Innovation Group needs to have Board level support and is the enterprise level framework which enables a part of the organisation to lift its head up from the day job and focus on R&D. In addition, it lays the ground work for business process and organisational change and provides governance and validation for Innovation with a capital "I" (as opposed to innovation with a lower case "i" – the day to day improvements which is part of doing business and which for us is lead by our Director of Business Change).
- In terms of AI, everything still starts with collecting or extracting the raw data reliably, using legal reasoning to make a decision before recycling those insights to inform or change future actions.
- There are clearly regulatory and legislative implications that you have to understand before you embark on any AI project.
- Much of what is termed AI is simply "clever search". Very clever to be fair, but "search" nonetheless.
- We cannot innovate alone. Our role in the AI world is as legal subject matter experts deployed in collaboration with clients, academia and software experts to deliver practical solutions that fix our clients' problems.
- Some of the technology that relies on machine learning needs to be taught what to look for in order "to learn".
- AI is not a magic wand nor a universal panacea for all ills. In the short term, AI based technology is not likely to provide the whole answer but does appear to us to be capable of augmenting human legal decision making.
- Any machine learning system is only as good as the human validating it, and a governance structure should be implemented with different levels of users and permissions to teach the system.
- Any solutions that are developed ideally need to be plug and play for clients on universal platforms which can in some circumstances offer a 'self serve' option.
- Not unreasonably, there is an expectation that technology will both reduce price and improve quality, or at least maintain it. However, this view does not take into account the up-front investment costs which can be significant.
- Technology may start to change the nature of the job itself. For example, with technically assisted review of legal documents, this shifts the focus of the efforts of a lawyer from supervision of a junior fee earner or document review to the training and validation of an AI process. Alternatively, a whole new class of knowledge engineers may emerge whose role is to understand the underlying capability of the technology so that the lawyers can concentrate on the law.
- In our view, AI solutions need to be auditable to allow us to work backwards to understand, if required, the decisions that had been applied. This is perhaps more important in the legal domain as a law change could result in the re-engineering of an AI system so being able to show how a decision was reached and then to be able to see how the change could affect an ongoing legal case or claim is important and another important step in the risk management of an AI solution.
- Sometimes what you end up with is not what you started out to fix but the exercise of Innovation can nevertheless result in products and solutions that turn out to be useful in their own right.
With our work well underway, in conjunction with commercial and academic partners, we are excited about the future and breadth of our client offering. Should any of our proofs of concept appear to be economically viable, we are looking forward to delivering AI solutions that will help solve some of our clients' legal problems. Effective legal AI solutions will not be developed without collaboration from all the potential stakeholders. As Socrates said "the secret of change is to focus all of your energy not on the fighting of the old but on building the new", we would like more of our clients to come on this journey with us.