Early Thoughts on AI in M&A

Artificial Intelligence is developing at a crazy fast pace with thousands of tools being developed to capitalize on the technology. Subcategories have already emerged including AI Content Detection, AI Enablement, AI Predictive Software, AI Speech Recognition, Data Labeling, Generative Ai Images, Generative AI Text, Generative Ai Video, Image Recognition, and Machine Learning (check out CabinetM’s guide here).

AI can mean a lot of different things. How do you define it?
One way of defining it is through product categorization. Is AI core to what the product does? i.e. Could the product exist without AI? If the answer is no then it belongs in an AI-oriented category e.g., Generative AI. If the answer is yes but is using AI to crunch large amounts of data or to enhance the product in some way (e.g., Hubspot with Chatspot.ai) then it should exist in its traditional category. Further, ChatGPT states that AI can be categorized as either weak or strong. Weak AI, also known as narrow AI, is an AI system that is designed and trained for a specific task. Virtual personal assistants, like Apple’s Siri, are a form of weak AI. Strong AI, also known as artificial general intelligence, is an AI system with generalized human cognitive abilities. When presented with an unfamiliar task, a strong AI system can use its intelligence to find a solution independently. Unlike weak AI, strong AI can understand, learn, adapt, and implement knowledge from one domain into another domain.

How do we see AI evolving in the field of M&A.
In the M&A world, it’s about leveraging AI to get deals done faster, better, and with less risk—all while enhancing value creation at each stage of the M&A process. If we look at the M&A process, it’s clear that AI could act as a very powerful tool to speed up analysis, assist decision making and identify risks. Eg looking at TPP’s six main stages we can see significant value from AI type tools:

Strategy: Accelerate the analysis of trends to support an M&A thesis, identify that targets exist that match the acquisition profile, and validate growth assumptions supporting plans.
Identify & Assess Targets: Enhance the ability to list and prioritize targets against a set of key criteria. Accelerate the collation of facts and figures to enhance the potential fit of each target.
Target Meetings & Valuation: Ability to garner a deep understanding of the target’s strategy by linking public domain information with answers from face to face meetings, early validation of the assumed integration model becomes possible. The value of acquiring the target becomes clearer much quicker including up to date comparables from the industry.
Negotiation of Price & Structure: If data is captured from all previous meetings and historical facts, AI could enable key patterns of behavior and remind deal leaders what matters most to the sellers, beyond just price.
Due Diligence & Legal Agreements: AI will summarize large tranches of data to validate information supplied and also importantly assist in stress testing post-acquisition strategies. AI will generate in minutes the detailed disclosure schedules that accompany an M&A agreement, including by integrating seamlessly with a target company’s internal data sources. What took an army of lawyers weeks will be done in hours.
Post-acquisition integration: AI will link key conclusions from due diligence to help shape the execution of integration strategies. It will suggest checklists to ensure the M&A playbook is not omitting key actions.

However, capturing these benefits will require these new generative AI models to ingest new data and insights garnered from target meetings, internal reports, external proprietary databases, up-to-date comparables of deals done from the market as well the public domain information readily available. Conclusions drawn from a limited data set will be dangerous.
Dominik Dellermann founded Vencortex which uses a Decision Optimization System (DeOS) to optimize the decision-making process from data to actions by combining human intuition and AI. I had a great discussion with Dominik about his new platform in the context of M&A, and there is no doubt that this type of application will bring efficiencies as highlighted above from strategy to integration. However, it will rely heavily on quality data input which will require a well-trained M&A team to extract that data from targets and from the marketplace.

M&A Legal Work: We know these models are capable of hallucination. Case in point from the airline industry. Roberto Mata sued Avianca airlines for injuries he says he sustained from a serving cart while on the airline in 2019, claiming negligence by an employee. Steven Schwartz, an attorney with Levidow, Levidow & Oberman and licensed in New York for over three decades, handled Mata’s representation. But at least six of the submitted cases by Schwartz as research for a brief “appear to be bogus judicial decisions with bogus quotes and bogus internal citations,” said Judge Kevin Castel of the Southern District of New York in an order. The fake source was ChatGPT.
In a recent WSJ article, How Can Companies Use Generative AI, Sara Castellanos interviewed Allen & Overy, a global law firm employing 3500 lawyers about their early adopter usage.
The main outcomes achieved so far seem to be around small productivity gains. An example cited required the model to create a first draft summarizing an area of law X and comparing and contrasting the U.S. and Europe. But they were keen to point out, that there is always an expert in the loop because of the hallucination factor. An interesting development cited by partner David Wakeling is the ability to point the model at specific databases potentially minimizing errors.

David Edgar at K&L Gates which employs around 1700 attorneys sees great potential for deploying what he calls M&Ai. As David nicely states, “For me, using “M&Ai” evokes the concept of “consilience” so powerfully advocated by the late Harvard University biologist E.O. Wilson in his book of the same name. Consilience connects seemingly disconnected things and derives its power from the force multiplier effect embedded in those connections. In Wilson’s world, consilience was connecting the humanities with the physical sciences and demonstrating the power of viewing the world through radically different lenses. In the M&Ai world, it’s about finding new ways to get deals done faster, better, and with less risk—all while enhancing value creation at each stage of the M&A process”.

Conclusion
We are at the start of an exciting era of technology. As always care will be required as we look for ways of deploying this powerful new wave of AI. Validation of conclusions generated by these models will be essential. We will continually review and highlight new tools we come across that improve the effectiveness of M&A programs.

TPP is buy-side investment banking reimagined. We seamlessly become an extension of your team and integrate at all levels to add deep mergers & acquisitions expertise into your business. Always ready for a conversation – 978 395 1155 or Ian@TPPBoston.com