Needless to say, we’re hearing a great deal about AI’s ability to revolutionize M&A due diligence. Kraken leveraging Termina.ai to accelerate their NinjaTrader acquisition is only the most recent example. Hours instead of weeks? Sounds amazing. Look, we don’t mean to sound like Chicken Little here, but before we crown AI the king of dealmaking let’s pump the brakes a bit. Are we confident that we’re not creating a bubble here, hype and buzz fueling the industry more than actual return-on-investment data.

I'm not saying AI is useless. But what I’m arguing for is not cynicism, but some serious skepticism. Remember the dot-com boom? Seriously, everyone was throwing money at anything with .com in the name. How did that turn out? Let's learn from history. AI’s promise to save time and improve processes certainly exists, but the devil really is in the data.

Here are three crucial data points to watch to tell if AI due diligence is truly transformative, or just another overhyped trend on the verge of bursting:

Data Bias Skews Deal Value?

AI algorithms are as good as data allows so long as it’s the right data. Look, data—especially this kind of intricate, context-dependent, complicated stuff—ain’t neutral. There is a potential for bias at every step, from initial data collection to the design of the algorithm. Think about it: if the training data for an AI due diligence tool primarily comes from successful tech acquisitions, how well will it perform when evaluating a traditional manufacturing company? Now, picture the accompanying public outrage when a problematic algorithm suddenly results in a $500 million write-down with no chance for recourse after a merger closes!

This isn't just theoretical. In fact, numerous studies have shown that AI systems often reinforce bias. These biases tend to exacerbate in key decisions such as hiring and loan applications. What makes us think M&A is immune? We need to push for transparency as to how these algorithms are trained and tested and undertake a full-scale, independent bias audit. Are artificial intelligence companies ready for transparency and to open their black boxes? I doubt it.

  • Key metric: Track the rate of significant post-acquisition write-downs for deals where AI was heavily relied upon for due diligence. Compare this to deals where traditional methods were used. A statistically significant difference should raise alarm bells.
  • The Unexpected Connection: Think of facial recognition algorithms consistently misidentifying people of color. This isn't a technology problem; it's a data problem. The same principle applies to AI due diligence. Garbage in, garbage out.

AI Misses The Human Element?

Termina.ai proved Kraken’s thesis for third-party developer deal with NinjaTrader. Great. Due diligence goes beyond simply crunching numbers. It's about understanding the qualitative aspects of a business: the culture, the leadership, the competitive landscape. Can an AI really evaluate the quality of a management team — or the likelihood that a key customer will jump ship? I seriously doubt it.

AI can also help spot differences between prospectuses and actual raw data — which is exactly what Termina.ai does. Can AI interpret those discrepancies? Can it grasp the story behind those numbers? Can it judge the reliability of the individuals reporting that data?

Even Robert Moore from Kraken Countless times before this, there was still dependence on advisors like PJT Partners and Ernst & Young. Why, if AI is so indispensable? Because human judgment is still crucial. Going forward, we need to be wary of over-relying on AI. Excessive reliance may reduce human oversight and increase the likelihood of mistakes or even fraud.

  • Key Metric: Monitor the frequency of "soft factors" (e.g., management quality, cultural fit) being cited as reasons for deal failures after AI-driven due diligence. If these factors are consistently overlooked, it suggests AI is not capturing the full picture.
  • The Unexpected Connection: Consider the challenges of AI in healthcare. A machine can diagnose a disease based on data, but it can't provide the empathy and human connection that a doctor can. Similarly, in M&A, AI can analyze financials, but it can't replace the nuanced judgment of an experienced dealmaker.

Unintended Consequences Loom Large?

The appeal of AI — the promise — is that it can do things faster and more efficiently. What if all that speed truly is at any cost? What if this pressure encourages the opposite — a race to the bottom? Otherwise, companies will feel increasing pressure to cut due diligence corners to remain competitive.

What if AI tools, by making these deals easier to scrutinize, inadvertently foster an environment for even more dangerous and speculative acquisitions to flourish? So is Arjun Sethi, co-CEO of Kraken, who is connected to Tribe Capital, a venture capital firm. So, yes, indeed, there is room for conflicts of interest. When AI tools make evaluations of the companies funded by the same venture capital firms that own those tools, bias can easily enter the picture.

We can’t let AI be used in ways that will harm our financial markets. This is not an attempt to stifle innovation, but rather to protect investors and ensure proper market stability. We need to ask tough questions: Who is liable when an AI-driven due diligence report is wrong? How do we ensure new tools powered by AI don’t get used to manipulate new markets or cover up fraud?

  • Key Metric: Track the overall volume of M&A activity, particularly for deals involving companies with high levels of debt or questionable financials. If AI leads to a surge in these types of deals, it could be a sign of excessive risk-taking.
  • The Unexpected Connection: Consider the impact of social media algorithms on political polarization. By optimizing for engagement, these algorithms can create echo chambers and amplify misinformation. The same thing could happen in M&A, with AI tools reinforcing pre-existing biases and leading to poor investment decisions.

AI can be used as a beneficial tool within M&A due diligence processes. While I don’t doubt the enthusiasm behind para. 1, we have to be careful. Let’s look at the data, keep our eyes wide open to the risks and unintended consequences. Let's avoid another bubble. Our wallets will thank us later.