How Technology Is Rewriting The M&A Playbook -- Again - Forbes

Artificial intelligence (AI) is often billed as the latest superhero or the latest boogeyman, depending on the industry it is disrupting. But while the advent of advanced technologies such as AI invites concerns and misinformation, the mergers and acquisitions (M&A) industry is already experiencing how AI is improving the due diligence workflow for dealmakers. When time is money, tools that speed up the M&A process are critical.

Technological advances have already revolutionized M&A, helping to create the virtual data room (VDR) and advancing security and efficiency exponentially from the days of the physical data room. AI is set to transform and speed up the process once again by automating several low-value tasks.

But will AI replace the need for M&A professionals entirely? No, just like digital printing presses didn’t wipe out publishing (although no one has the title of “typesetter” anymore). However, it will require a shift in the way dealmakers think about and approach due diligence and the broader M&A life cycle.

At the macro level, the potential economic payoff of AI is immense: McKinsey calculates (via ZDNet) that the implementation of AI could generate $13 trillion in additional economic activity by 2030 and boost global GDP by more than 1% a year.

But there is a long way to go from 2019 to 2030, and we are still in the early stages of AI implementation.

The proliferation of data is a gift and a curse for M&A

Like most things in today’s economy, the need for AI all comes down to data.

We now produce more than 2.5 quintillion (that’s 18 zeroes!) bytes of unstructured data in a day. Making sense of all this data has been difficult. But AI — and more specifically, machine learning, which features statistical methods that allow a system to learn from data — is changing this.

There are two types of machine learning: unsupervised, where there is no designated outcome from data analysis and the technology solely determines the outcome, and supervised, where there are designated problems or goals and inputs, and the technology is able to find the best solution given the situation. Unsupervised machine learning is the stuff of Terminator movies, while the majority of today’s machine learning (also known as augmented learning) is supervised with oversight, direction and human-driven decisions.

For example, in private equity, if a company’s P/E ratio is greater than x, EPS is greater than y and the price valuation is greater than v, then a basic buy/sell recommendation can be reached. Of course, the data inputs can come from any range of sources and be wildly complex.

A better M&A toolbox means less time on low-skill tasks

So, what does all this mean for the rest of us who aren’t data scientists? Are headcount reductions imminent?

No.

Critical thinking is still a precious commodity and central to deal success. Yes, machine learning can help us arrive at a valuation, but it can’t negotiate financing or make subjective judgment calls based on a company’s culture and management. Those kinds of actionable, critical decisions cannot be replaced by machines and likely never will be.

The key point here is that recent advances in machine learning and data analytics mean that we will be able to replace low-value work across the M&A life cycle, such as the mind-numbing work of document naming conventions, metadata and categorization.

In fact, our company is researching and testing the use of AI and natural language processing (NLP) to cut out the time-consuming yet necessary tasks associated with the due diligence process. We’re testing the use of AI in our integrated redaction tool, which allows users of our due diligence app to bulk redact and unredact words, phrases and images in documents, right from within the virtual data room. Others are using AI to review contracts and to reduce the time and administrative work involved with the hidden obstacles of almost any M&A. And Deloitte's M&A analytics platform iDeal is “currently using AI and machine-learning on the front end of the analytics process to organize and tag massive amounts of data as part of the workbook creation process.”

With any one M&A deal requiring analyses of years of financials, hundreds of contracts and tens of thousands of documents, this kind automation is very attractive. What used to take endless hours can now be done at the click of a button. Now you can put all those talented MBAs to work doing tasks that support deal success, not deal prep (which is obviously still valuable).

How AI is expanding the spectrum of analysis

The buy/sell example from PE is painfully rudimentary. Good AI models can leverage deep learning to process hundreds of variables in multiple dimensions and identify patterns that a human analyst may not see. They are then able to learn from these analyses and subsequent outcomes to sharpen their models, altering variables and weights as the economy and markets change, which is fundamentally different from traditional rules-based computer models.

All of this helps inform management decisions and recommendations.

Does it guarantee success? Not quite. The very nature of AI models requires some failures to happen for learning to occur (as is the case for us mortal humans). If anything, the failures may come faster with AI modeling. However, with budgets and investor capital at risk, this kind of learning is far more efficient and helps to ensure companies and advisors use their resources intelligently.

https://ift.tt/33LzDDD