#17 Could AI Have Prevented a Mistake That Led to $80 million in Post-Closing Liability?

This week, I discuss whether AI could have prevented a mistake that cost a seller in an M&A deal 80 million dollars in post-closing liability.

AI will change the world, but how will it change M&A? I want to focus on AI’s impact on M&A in this newsletter. I am not an expert on either M&A or AI, but I want to learn about both topics and how they intersect. I thought there might be others in my situation (or people who are experts in one field or the other) who would find information on M&A and AI helpful in their careers, so I created this newsletter to track and share what I learn.

Could AI Prevent 80 Million Dollars in Post-Closing Liability?

Here is a fascinating Money Stuff article about an error in a purchase agreement that caused the seller to pay the buyer over $80 million after closing.

Kingswood Capital Management LP, a private equity firm, bought SaveMart, a closely held regional grocery store. SaveMart was a partner in a joint venture operating another grocery store. SaveMart was a guarantor of the joint venture’s debt.

The purchase price for SaveMart was $245 million, net of cash and debt. The pre-closing balance sheet of SaveMart listed SaveMart’s interest in the joint venture as an asset (the value was net of debt and other equity interests in the joint venture, apparently this is how the accounting rules work). There was a “post-closing true up” of SaveMart’s balance sheet. This means if there was more debt than expected on SaveMart’s balance sheet, SaveMart would be responsible for paying it back to Kingswood. The purchase agreement’s definition of “Closing Date Indebtedness” included debts that SaveMart guaranteed—which included debt of the joint venture. SaveMart’s share of the joint venture’s debt totaled $109 million. Needless to say, SaveMart got a nasty post-closing surprise—after the post-closing true-up, they owed over $80 million to Kingswood.

I think it’s safe to say SaveMart’s lawyers made a huge mistake—their draft (not Kingswood’s) of the purchase agreement contained the definition that ultimately caused the $80 million payment to Kingswood. My question is, could AI have prevented that mistake?

As it stands today, I do not think AI could get close to catching this issue. There would be too many hoops to jump through. For example, the AI model would have to understand how the accounting rules treat an interest in a partnership on a balance sheet, then would have to recognize that the purchase agreement’s definition of debt could include guaranteed debt, and combine the joint venture’s balance sheet with the purchase agreement defintion of “debt.” It seems far-fetched to think that today’s AI could successfully navigate this complex situation. I can imagine the lawyers feeding the model all of the deal documents (which is likely impossible, considering the size of the contracts and the very limited text that AI chatbots can absorb) and asking the model, “See any problems with this?” To which the model will almost certainly say, “Nope, looks good!”

An M&A deal has vast amounts of context. Every email, every conversation with a colleague, a client, or the opposing side, and in a small way, every thought about the deal of each of person involved has some effect on the outcome of the transaction. Today’s AI models are not built to solve complex and unique issues like M&A deals because they are limited in how much context they can take in.

I think to have an AI model that is fully capable of advising on a complex transaction, it would be necessary for the AI model to “look over the shoulder” of the entire deal team so that it could understand the nuances and complexities of the transaction. Perhaps in this situation, this type of AI model (which would likely meet the requirements of artificial general intelligence or AGI) would know about the joint venture debt, the differences in accounting practices for the pre-closing balance sheet and the post-closing balance sheet, and would alert the lawyers to change the purchase agreement’s definition of debt. This might be wishful thinking, and if anything, it’s probably a long way away, but I think it’s possible!

This little thought experiment should be comforting for deal lawyers and M&A professionals. Despite the numerous claims that AI will take everyone’s job, I don’t think AI can consider everything that is necessary when advising on a transaction. Plus, AI can’t build relationships with people and act as a trusted advisor in the way humans can. It is true, however, that AI will replace some lawyers—lawyers who refuse to use AI as a tool. It will not replace the lawyers who use AI.

Unfortunately for SaveMart, their lawyer’s mistake could not have been caught by AI and they are stuck paying Kingswood to buy their business.

About me

My name is Parker Lawter, and I am a law student pursuing a career as an M&A lawyer. I am in my last semester of law school, and with some extra time on my hands, I decided to create this newsletter. I hope it is informative and helpful to anyone who reads it! I am not an expert at either M&A or AI, but I am actively pursuing knowledge in both areas, and this newsletter is a part of that pursuit. I hope you’ll join me!

Follow me on LinkedIn: www.linkedin.com/in/parker-w-lawter-58a6a41b

All views expressed are my own!