Professional Service Firms in the AI Age
Although they may not admit it, most professional services employees — particularly those in the younger cohorts — will be using ChatGPT, Gemini, or other LLMs daily. Whether for fine-tuning an excel algorithm, rewriting important emails, or providing framework inspiration, LLMs have fundamentally changed the world of professional work. Those employing these tools are finding more efficient methods to deliver work, while those who don’t might notice their more tech-savvy peers advance quicker in their chosen field, as they produce articles quicker, solve excel quibbles easier, and write far cleaner.
Organisations would do well to take note. Rather than simply letting their technologically-aware hires leverage the vast power of generative AI, they may want to turn this lens onto themselves. Often a core issue facing companies is inherent knowledge. Those that have been there for years — even decades — are comfortable and familiar with the ways of working. They know where documents are, presentations lie, or other IP and intellectual assets exist. These employees, however, leave. Or get tied up in bottlenecks. And with this where the skeletons of effective presentations or the bodies of great work are buried.
The most innovative blue chip organisations have recognised this, and are using Generative AI models to leverage their deep pools of internal knowledge. From reviewing thousands of documents in order to produce consistent presentation structures, to scanning legal submissions for crucial kernels of information, applying generative AI models to a company’s existing knowledge bank can save significant amounts of time. Considering most professional services firms bill by the hour — and place value on time-saving applications — this results in significantly increased productivity and therefore profitability.
Leveraging Knowledge Bases With AI
McKinsey — a globally renowned consultancy — was one of the first to introduce a proprietary tool. ‘Lilli’, its AI chatbot, helps consultants by providing insights, generating reports, and performing various administrative tasks. Lilli leverages advanced natural language processing (NLP) and machine learning techniques to support McKinsey’s operations, all by using more than 100,000 documents and interview transcripts. This can turn work that would take weeks into a matter of minutes, with a considerable level of transparency. This is crucial considering the well-documented issues Generative AI models are facing with hallucinations and false information. When Lilli answers a question, it shows its work with all the sources, links, and even page numbers, so consultants can double check the results before a potentially awkward interaction with a well-paying client.
The ‘Big Four’ professional services networks have followed suit. Deloitte’s AI Studio, PWC’s ‘Halo’, EY’s Helix, and KPMG’s ‘Clara’ all employ Generative AI in a similar fashion. The technology works by augmenting existing Generative AI models with company assets. Built in a shared, private environment that only employees can access, these tools reduce privacy risks of sharing commercially sensitive data over publicly accessible self-learning tools, e.g. ChatGPT. This is either through building a proprietary Generative AI model — an expensive, labour-intensive, and technologically sophisticated approach — or by using an existing LLM as the “engine”, powering the process. The company’s assets are then layered on top using a machine learning process called Retrieval Augmented Generation (RAG), a technique for enhancing the accuracy and reliability of Generative AI models with facts fetched from different sources.
A&O Shearman, a merger between Allen & Overy — a full-service magic circle law firm — and Shearman & Sterling, is employing this latter approach. Harvey is an innovative artificial intelligence platform built on a version of Open AI’s latest models enhanced for legal work. Harvey empower’s more than 3,500 of A&O’s lawyers across 43 offices operating in multiple languages with the ability to generate and access legal content. It uses natural language processing, machine learning and data analytics to automate and enhance various aspects of legal work, such as contract analysis, due diligence, litigation and regulatory compliance.
Professional services firms should follow this lead. Although examples given are global companies, employing tens of thousands, with considerable capital expenditure at their disposal, there are ways of reducing these costs whilst still benefiting from Generative AI’s efficiency-saving powers. Organisations that are slow to adopt will see their competitive differentiation diminish. Resources should be directed to ensure employees are supported and empowered by their organisations to use cutting-edge Generative AI tools. They should be trained and upskilled to employ sponsored and consistent GenAI tools, rather than in a siloed and incohesive manner, to ensure everyone benefits from the ground-breaking advancements being made in Generative AI, not just the technologically-savvy few.
This article first appeared in The Playbook, download the entire edition for free today.