Research and software providers

Research on AI used in M&A

In January 2024, Bain & Company released their M&A Report, where they polled 300 M&A practitioners on using Generation AI (Gen AI) in their M&A processes. According to their report:

• Gen AI use for M&A deal processes was low at 16% (at the time of the research), but it is expected to reach 80% over the next three years; and
• early adopters are using Gen AI primarily to identify targets or conduct document and data review. The image in their report effectively illustrates this point.

The report found that while Gen AI (when it is being used) is being mostly used to find the right deal and validate the deal through due diligence, there is still potential to expand its use to planning and executing the integration to derive the most value from the merger or acquisition.

Potential uses - generation types

When considering where Gen AI may be beneficial in the deal cycle, it helps to think of the different types of information the AI tool can generate (known as generation types):

• foundation model providers such as Anthropic's Claude or Google DeepMind's Gemini;
• text generators such as Microsoft’s Co-Pilot or ChatGPT;
• image generators such as OpenAI’s Dall-E or Midjourney;
• code generators such as Github Copilot or OpenAI’s Codex;
• video generators such as Runway, HeyGen and Pika Labs; or
• audio generation such as Google's AudioLM or OpenAI's Jukebox.

Software providers

After identifying the type of information you’d like to generate, you could narrow down to the software provider of that generation type and any specific configuration that you may require. Many of the providers of AI tools are specialist start-ups being acquired by or supported by large technology companies, such as Microsoft Corp, Google, Meta Platforms Inc, Amazon Web Services and IBM Corp.

An example of an AI software provider is the start-up, Harvey AI, an AI company founded in 2022 and used by professional services firms. It uses a configuration of Microsoft’s OpenAI's GPT software and states that it is able to review contracts, assist with due diligence by examining documents and support regulatory compliance.

Other proprietary tools used in M&A

• Canvass AI: for enhanced due diligence research
• Intralinks: AI-Powered M&A Platform | DealCentre AI
• Mosaic: LBO modelling software, which can predict investment returns with a few quick manual inputs (it works similarly to Blackstone’s Atlas)
• Xapien:  Due Diligence tool that is able to prepare reports in minutes

Human judgement and risk management

It's important to keep in mind that using AI for the digitisation of tasks will not negate the need for judgement. One example of professional judgement that remains very important is to decide on the amount of data that should or can be disclosed to other parties involved in a deal.

Additionally, the need for expert human review of the outputs of AI for inaccurate or incomplete conclusions or hallucinations remains of paramount importance.

Firms will have to be transparent with their clients about the abilities of any AI technology they implement and how they manage the risks of deploying AI. Refer to our AI in Corporate Finance risk page for mitigating actions to manage risks associated with applying AI.

How to prioritise which tasks to start with

With the sheer number of options available along the deal cycle to deploy AI, it can be a daunting task for those that haven’t started their AI journey. They need to figure out where and how to get started and which tasks should be automated using AI.  Refer to the Practical considerations for using AI in corporate finance for more information.

To help you prioritise which tasks to start with, we recommend asking the following questions.

• What were the original aims of deploying AI in your organisation – eg, increase the speed at which deals are initiated and completed, enable better, more successful deal outcomes, or, for advisors, opening up a new service? Compare to the benefit of each of the possible AI uses above.
• Which tasks are manual and repetitive that could be automated through simpler algorithms (and do not rely heavily on nuanced human judgement)?
• Which tasks, once automated, are easy to audit – both in terms of information source and how the findings were arrived at?
• Where can I build a competitive edge?

For a thorough outline of practical matters to consider when deploying AI in the deal lifecycle, refer to the Corporate Finance Faculty’s advice on practical considerations for using AI.