Artificial intelligence seems to be viewed as the future of everything, but what does it mean in the context of M&A and corporate finance? David Prosser gauges the mood of mid-market advisers.
Advocates of artificial intelligence (AI) may be prone to making outlandish statements about the transformative potential of the technology, but corporate financiers are convinced these new tools will support dealmaking. Research published this summer by SS&C Intralinks revealed 43% of dealmakers had already adopted at least some AI tools. And a survey from Bain & Co found that while only 16% of M&A practitioners currently work with generative AI (GenAI), 80% expect to be doing so within three years.
The challenge is for mid-market corporate finance advisers to keep up: with smaller tech budgets and less in-house expertise, there is a danger some of these firms will be left behind. “Most people don’t use software such as Word or PowerPoint to anywhere near their full potential,” warns Chand Chudasama, a strategic corporate finance partner at Price Bailey and head of the firm’s London M&A team, “so getting them to think about the applications of this new technology they don’t understand is a big ask.”
The key is to be prepared to experiment, suggests Lee Humble, UK head of corporate finance at Azets: “We’re excited about AI, but we’re in the early stages of adoption where we’re still trying things out. This will enable us to move forward with a more rounded and safe strategy for implementing new technologies; that’s something we can do later this year.”
Chudasama’s advice is to make it easy for people to explore AI. “What really changed things for us was getting access to AI tools at the click of a mouse button, rather than expecting people to launch an application or navigate their way to a specific URL,” he says. “We’ve also embedded AI functionality into software our people are already used to using, so it’s less of a leap into the unknown for them.”
Equally, suggests Anna-Louise Shipley, a corporate finance director at Buzzacott, there is merit in moving cautiously. “We know change is coming and we’re definitely starting to embrace that, but we’re also conscious that this is an emerging technology with new risks and we’re happy to learn from others’ mistakes.”
In practice, there are three key parts of the M&A process where AI is starting to gain real traction: deal origination; due diligence; and execution and integration. In each case, the hype around the potential for M&A is now beginning to translate into reality, with growing numbers of third-party products and services available, and more corporate financiers developing in-house solutions.
Take origination: the challenge for dealmakers wanting to buy or sell a business is always finding the best possible counterparty, whether that’s another corporate or a financial investor. But the universe of potential counterparties is huge and includes not only obvious contenders – companies in the same sector or investors with a specialist interest in such deals, say – but also players that might not immediately spring to mind. The latter might include counterparties in far-flung overseas markets, perhaps, or businesses in adjacent sectors.
M&A advisers are good at working their way through this universe, using established databases to come up with long lists. But AI could help them be even better, automatically scanning far larger datasets for counterparties that meet set parameters to ensure no stone is left unturned. GenAI, moreover, enables advisers to ask much more tailored and nuanced questions of both structured and unstructured data. The result should be an imaginative list of potential counterparties that offer the right strategic, financial and culture fit.
“AI such as Microsoft Copilot and, in particular, Copilot for Bing is a good way to enrich more traditional research methods,” says Sarah Belsham, a partner at RSM. “It adds a complete view that, when combined with expert review from the M&A team, can deliver enhanced output. It doesn’t necessarily speed up the process, as it’s imperative that results are reviewed and revised effectively, but it does allow us to deliver a more comprehensive output.”
Build or buy?
Larger corporate finance advisers are developing their own versions of these tools, but there are also specialist platforms and products that advisers can simply buy into. The technology company mnAi, for example, now holds 12 billion searchable datapoints on 11 million UK companies, covering everything from their financial performance to the gender diversity of their boards and shareholders. Established M&A platforms such as Beauhurst, Megabuyte and Intralinks are becoming increasingly sophisticated.
Shipley says her firm is making more use of such tools as a useful starting point. “We can use them to put lists together that we then validate through more traditional approaches,” she says. But she also counsels caution – the ‘hallucinations’ for which GenAi in particular is often criticised can cause problems: “We’ve done searches that have come up with completely fictitious companies.”
Turning to due diligence, many advisers believe the potential for AI to expedite – and improve – this stage of the deal is significant. They certainly need help – recent research found that the average number of pages in a typical virtual data room has increased from around 29,000 in 2018 to more than 50,000 today. Working through that data, without missing key information that might influence negotiations or even undermine the deal rationale, is an intimidating challenge.
AI tools, including GenAI, will do more of the heavy lifting. They enable practitioners to interrogate the data in as many specific ways as they see fit – and to ask questions with increasing subtlety. Such tools also link to external data – case law, regulatory statutes and news sources, for example – to provide additional context during diligence.
Dylan Polley, an M&A associate director at Oaklins Evelyn Partners, points to the power of AI for crunching through huge amounts of data to generate insight. “We use a number of the data analysis tools in GPT-4, which can automatically take data from inputs such as PDFs and plug it into Excel,” he says. “It instantly creates insights that we may not have initially considered, because it’s looking at data in new ways.”
Search for new business
Corporate financiers on the hunt for new clients are increasingly using AI to make initial contact. GenAI makes it much easier to produce high-quality approaches, researching the particular issues, challenges and opportunities that prospects face and crafting a bespoke message accordingly.
“AI is really valuable for outreach,” says Dylan Polley of Oaklins Evelyn Partners. “Sending generic cold emails rarely yields a particularly good response, but we’re now using AI to ensure our communications are much more individually tailored; potential clients are far more receptive.”
On the breadth of application, Price Bailey’s Chand Chudasama agrees: “AI definitely makes a difference. We can’t give away exactly how we’re using the models, as the likelihood is they will only provide competitive advantage for a couple of years and then others in our space will have caught up. That means AI has to pay back in the relatively short run through very specific applications.”
Mid-market firms naturally have limited resources for marketing, so while business development teams could potentially evolve some of these communications manually, doing so at scale without compromising on the quality is challenging. AI can automate the process, often producing materials that are far more effective than previous efforts. “New clients have complimented us on the first contact we’ve made without having any idea that a robot did all the work,” says a marketeer at one mid-market M&A firm.
In addition, GenAI is good at creating content – summaries of due diligence documents, perhaps, but also drafts of deal announcements and regulatory filings. Human interaction will still be needed – not least to police the accuracy and quality of such outputs – but there is an opportunity here to save time and resources.
In execution and integration, tech has a role in anticipating deal challenges and, ultimately, driving M&A success. “We’ve developed training models that are making very good progress on nuanced topics,” says Chudasama. “That has been really useful in speeding up the process of analysis, and the comparative difference to untrained models is stark.”
New applications are emerging at pace. In Australia, for example, technology firm Ansarada has developed a tool it calls AI Bidder Engagement Score; it claims a 97% accuracy rate in predicting bidder behaviour from day seven of the deal. The consultant McKinsey’s myIMO tool, by contrast, helps businesses post-
completion: users can ask for advice or create content critical to integration. The tool is trained on McKinsey’s database of M&A playbooks and best practices to help users make smarter and quicker decisions in integrations and separations.
Data skills
With so much going on, the learning curve is steep, so mid-market firms are investing in upskilling. “Data literacy is at the heart of effective use of AI,” says Belsham, “so over the past three years we have equipped our teams to work effectively with data – to understand the questions that need to be asked to ensure they can trust and be confident in the data they are using.” The firm has also brought new skills in, she adds: “We have invested in data analytics specialists to enhance our capabilities and complement our M&A expertise.”
Importantly, says Polley, most staff don’t need to develop further technical knowledge; rather the onus is on ensuring they can get the best out of new tools. “Probably 95% of the training we do with colleagues is about helping them to ask the right questions of AI,” he says. “That’s where the added value will come from.”
That’s what clients want, of course, but corporate financiers are conscious that some may also have concerns about new technologies. “We always ask for consent before working with AI on behalf of a client – and we always disclose our methods,” says Azets’ Humble. “We do need to be absolutely transparent about what we’re doing and, in particular, what data is being used, where it’s going and how we protect it.”
Indeed, client confidentiality is a key concern. For example, data uploaded into a non-private GenAI application will almost certainly be ingested by the underlying model, because the technology improves by learning from each new input. This potentially compromises the security of the data.
Question of trust
Recognising these challenges and reassuring clients that confidentiality will not be compromised can help them to feel more comfortable with AI. “Clients want good advice from trusted advisers,” says Belsham. “AI will never replace that, but delivering more comprehensive research that is interpreted by an experienced team to shape the best recommendations is of value to clients.”
Many clients will be experimenting with AI for themselves, which may cause its own problems. One adviser’s client challenged him on valuation. “He’d put his data into a GenAI application that used a very generalised set of discount rates and it gave him a number 300% higher than the figure we’d come up with. We’re using models that we’ve trained to be much less generic, but he took some convincing that our numbers were accurate.”
While this story raises a wry smile from the adviser in question, it also underlines the growing pressure on mid-market advisers to get up to speed with AI. When the technology is being adopted by the clients – never mind bigger rivals – firms cannot afford to stand still.
Needle in the haystack
There’s a new breed of venture capital investor who sees AI as a source of competitive advantage when it comes to identifying the best deals. London-based Moonfire Ventures is a leader, launched three years ago by former Atomico co-founder Mattias Ljungman and computer scientist Mike Arpaia.
Moonfire uses a custom-built AI model to identify potential investee companies, each week reviewing up to 50,000 companies in four key sectors – health, work, finance and gaming. Its algorithm uses parameters set by the firm, from founders’ experience and track record to potential returns on investment, to identify likely targets.
One company identified by the technology is LiveFlow, a fintech offering accounting tools for small and medium-sized businesses. Moonfire then made contact to conduct due diligence, leading to a £3.5m seed funding round at LiveFlow led by Moonfire.
Moonfire continues to evolve the technology. Last year, for example, it incorporated OpenAI’s GPT-4 GenAI platform into its models, making it easier for staff to use them. Text prompts have replaced linear algebra as the means through which the models are interrogated.
