Can Machine Learning Support Strategy Development?

It seems likely that the most successful deployments of AI in the years ahead will be in helping us to predict more effectively. Research from INSEAD argues that this could have useful benefits in the boardroom and help leaders with a range of corporate decisions, especially when complexity and uncertainty are strong.

The researchers focused on the private equity industry, where deal syndicates are commonplace. These often involve banks, companies, and institutional investors who pool their expertise and capital together. They analyzed whether so-called “add-on deals” change the nature of these syndicates.

Changing strategy

Add-on deals are when the private equity firm acquires a company and then merges it with one already in its portfolio. It marks a change from traditional practice that sees money borrowed to buy a standalone firm.

The researchers analyzed around 60,000 deals from over 4,500 PE firms between 1990 and 2016. Machine learning was used to analyze these deals and found a couple of clear patterns. Firstly, add-on deals often feature a new type of co-investor, which is often a company with expertise in such deals.

The second pattern is that co-investors are often new participants in the pool of partners and are not among those that usually participate in leveraged buyouts. The researchers believe that new partners would be required to pursue add-on deals, which would make PE firms with a lot of existing partners slower to adopt add-on deals.

This was put to the test on an additional dataset and strong support for the hypothesis was found. In other words, an existing alliance network can impede the adoption of new practices, especially if the new strategies require a strong alliance of partners.

The researchers believe that their finding shows how AI can be used to think of strategies in a new light and devise new ways of working that might bypass boards and their traditional decision-making processes.

Facebooktwitterredditpinterestlinkedinmail