In 2023, global management consulting firm McKinsey & Company launched its own generative AI platform, Lilli. Named after Lillian Dombrowski – who had joined McKinsey in 1945, became the first woman to earn her MBA while at the firm, and played a pivotal role organizing its historical records – Lilli activates nearly a century of institutional knowledge for the firm’s 45,000 employees.
Lilli is now one of McKinsey’s most used digital tools. The reason is simple. “It is tuned to McKinsey and our client service, meaning that the way it recognizes the intent of questions, and the way it fine-tunes answers to make them useful for our colleagues to bring to clients, is very ‘McKinsey-esque’,” explained McKinsey Senior Partner Erik Roth.
Unlike popular general-purpose AI options, which are open to the public and capable of handling a broad range of tasks, Lilli is designed exclusively by and for McKinsey staff. Responding to prompts and questions, it delivers recommendations grounded in proprietary data from the firm’s extensive knowledge repositories. At the time of its launch, McKinsey reported that Lilli had access to 100,000 documents and interview transcripts. Today, that number is likely many times greater.
Enterprising historians should feel empowered to use these tools to venture into uncharted territory.
Custom AI tools can streamline workflows, automate repetitive or low-skill tasks, and provide data-driven insights in a matter of seconds, freeing up staff to devote more time to high-ticket items – or, more cynically, giving leadership a convenient excuse to downsize. Some of the initial investment can also be recouped by selling lessons learned during the design, implementation, and maintenance of internal AI systems to AI-hungry clients.
For historians, long sidelined and uncertain about how the AI revolution might shape their professional futures, the rise of advanced in-house chatbots trained on proprietary research presents a golden opportunity to reclaim relevance.
The Arts and Humanities in the digital age
Voices in and outside the academy have greeted the arrival of AI with a mixture of fear, scepticism, and enthusiasm. Some have shown how historians can make use of AI to accelerate research and gain a better understanding of the past. Others have pushed the boundaries of the digital humanities field with AI-driven approaches, or lent historical perspectives to discussions about responsible AI governance.
Historians have occasionally been invited to participate in multidisciplinary research initiatives examining the impacts of technology. A prominent example was the 2018-2023 Living with Machines project, a collaboration between the Alan Turing Institute, the British Library, and several UK universities, that brought together historians, data scientists, geographers, computational linguists, librarians, and curators.
The value of soft skills, such as creativity, critical thinking, communication, emotional intelligence, and context awareness, is increasingly being recognized by employers. However, there remains an urgent need for more specific use cases, which would help historians apply their expertise more purposefully and with less skepticism.
Organizational History 2.0
The success of McKinsey’s Lilli has validated a strategic management theory that gained prominence in academic circles during the 1980s and ‘90s. In his 1984 paper, ‘A Resource-Based View of the Firm,’ Birger Wernerfelt argued that a firm’s internal resources, and management thereof, could be the key to creating a competitive advantage. Examining organizations in terms of their resources, rather than their products, later gave rise to the VRIS classification framework – V(alue), R(areness), I(mitability), and S(ubstitutability).
An organization’s institutional knowledge may be the purest manifestation of a valuable, rare, inimitable, and non-substitutable/strategic resource. It is also a cumulative asset, meaning the longer the organization has existed, the more institutional knowledge it has likely accumulated, thereby further strengthening its competitive advantage.
Historians that specialize in institutional history, and do so supporting private, public, and social sector clients, assess and leverage an organization’s knowledge and memory – their unstructured data – in a myriad of ways: to craft narratives, improve decision-making, recognize patterns, develop strategies, ensure smooth leadership transitions, foster cultural cohesion and creativity, identify opportunities and threats, design tailored learning solutions, and much more.
The results from these engagements are typically presented in the form of books, articles, case studies, research reports, or policy briefs. At other times, the consulting historian may be called upon to deliver lectures, facilitate workshops, or provide one-on-one guidance to senior leaders. Knowledge mobilization in the 21st century looks not unlike what medieval scholars and advisors may have experienced after the invention of the printing press in the 15th century – though now, the dissemination and exchange of knowledge is done largely electronically.
But what if the outcome of a legacy project is not a book, but an AI chatbot fluent in the organization’s history? What if historians expanded their traditional roles as knowledge creators and synthesizers to become trainers for both large and small language models?
I say, why not? If generative AI enables STEM majors to sound like seasoned writers, then enterprising historians should feel equally empowered to use these tools to venture into uncharted territory. They may have to forgo a byline on a publication that may or may not be read – a small price to pay considering the vast reach of today’s AI platforms. As of February 2025, ChatGPT alone recorded 400 million weekly users, while proprietary platforms like Lilli engage thousands of employees each week.
Future opportunities
Worldwide spending on AI is projected to more than double by 2028, reaching $632 billion, according to a new forecast from the International Data Corporation. This investment frenzy requires guardrails. Involving professional historians in the machine training and data ingestion process yields many strategic advantages and quality assurance checks for organizations.
Historians can prevent the perpetuation of biases and anachronisms. They can spot false, missing, confidential or illegal information. They can produce different writing styles for different audiences. They can furnish users with multiple interpretative scenarios. They can highlight possible path dependencies. Most importantly, they can address the ‘black box’ issue by attaching clear explanations to the system’s outputs.
In the future, as AI becomes more affordable for SMEs and nonprofits, it is entirely possible that every organization will have access to their personalized version of Lilli. To ensure optimal performance and prevent knowledge dilution, organizations must supply their internal AI platforms with high-quality information – ideally created or curated by those equipped with the historian’s toolkit.
Dr Tim Mueller, founder and managing director of Chester & Fourth.
__
Sign up for our newsletter to get the latest innovation news and more sent to your inbox.
Events innovators won’t want to miss in 2025
Dr Tim Mueller is the founder and Managing Director of Chester & Fourth Inc., a boutique advisory firm based in Canada. He helps mission-driven organizations and leaders with their strategy and innovation needs by tapping into their most underutilized resource: their own people and institutional knowledge. He is currently co-authoring a history of West Coast philanthropy funded by a prominent Silicon Valley private foundation. Earlier, he directed multi-disciplinary research projects for the Ford Foundation and BMO Bank of Montreal.
