AI in Action: Redefining Middle-Market M&A 

Angela Raitzin: Good morning and good afternoon, everyone. Thank you for joining us today. I’m delighted to welcome you to this timely discussion on how artificial intelligence is shaping the world of mergers and acquisitions.

AI promises speed, efficiency, and deeper insights that can transform how transactions are identified and executed. But it also raises important questions about data quality, ethics, and balancing human judgment with machine intelligence. Our goal today is to explore these dynamics — to understand not only where AI is making an immediate impact, but also how it might reshape M&A in the years ahead.

Let’s start with introductions. Kathleen, would you begin?

Kathleen Garenani: Hi, I’m Kathleen Garanani, Director of Responsible AI and AI Governance at BDO Digital. My role involves helping clients deploy technologies and systems responsibly, ensuring organizations become good stewards of AI.

Mike Silverman: I’m Mike Silverman, Chief Investment Officer for Crescent, based in Dallas. Our firm, formerly CH Investment Partners, merged with Crescent about a year ago. We’re an active private equity investor, both in funds and co-investments. I’ll share how we’re practically applying AI in our work today.

Tal Hacohen: I’m Tal Hacohen, a partner in the M&A group at Holland & Knight. We’re a full-service law firm with about 2,200 lawyers across the U.S. and South America. I’ve spent the past 20 years handling M&A deals, and with the developments in AI, it’s been an exciting period of learning and adapting.

Steve Rathbone: I’m Steve Rathbone, Vice Chairman and Managing Director in the Investment Banking group at Stout. For the past 20 years, I’ve specialized in sell-side M&A, primarily in the middle market. We’re excited to explore how AI can help streamline processes, reduce friction, and enhance efficiency.

Angela Raitzin: Thank you all. For our participants, please use the Q&A section to submit questions — we’ll address as many as possible toward the end.

Let’s begin with an internal perspective. Mike, how has AI changed your role and Crescent’s operations over the past year?

Mike Silverman: It’s moving fast — exciting but also a little daunting. We’ve rolled out our own internal ChatGPT, which keeps our data secure and gives us flexibility in how we use it. On the investment side, AI is changing how we approach analysis. It helps with quick wins — writing Excel macros, drafting letters — but also acts like an extra person at the table for brainstorming. Of course, we always approach its output skeptically; it can be wrong, so human oversight remains essential.

Angela Raitzin: Tal, how is AI impacting your role and the legal industry more broadly?

Tal Hacohen: I completely agree with Mike — always assume it might be wrong. In law, our job is to analyze and advise carefully, so while tools like ChatGPT make things faster and more efficient, accuracy is crucial. We’ve developed our own proprietary AI system that allows us to take the first pass at tasks more efficiently. It’s particularly helpful for due diligence — instead of reviewing every line of every contract, we can extract the key points we need to focus on, saving time while ensuring accuracy.

Angela Raitzin: Steve, what about investment banking?

Steve Rathbone: For us, AI has made research and content preparation far more efficient. Whether for meetings, pitch materials, or confidential memorandums, it accelerates the early stages of the process. It doesn’t replace our work but augments it, helping teams get a jumpstart. Long term, we’re breaking down every step of the M&A process — origination, onboarding, execution — to identify where AI can reduce friction. The more granular our understanding of the process, the more we can automate intelligently.

Angela Raitzin: Kathleen, from the accounting and advisory perspective?

Kathleen Garenani: At BDO, we “take our own medicine.” Like others here, we’ve deployed an internal AI sandbox that keeps proprietary data secure. It’s been invaluable for brainstorming, automating mundane tasks, and de-risking decisions throughout the M&A lifecycle. And just to clarify — LLM stands for Large Language Model. It’s a subset of AI that generates new content, as opposed to robotic process automation, which focuses on repetitive tasks.

Angela Raitzin: Perfect — thank you for the quick AI 101! Now, let’s move to the client perspective. Tal, how are your clients adapting to or embracing AI?

Tal Hacohen: It depends on the client. Many are experimenting, while others approach it cautiously. Some come to us after using ChatGPT for legal questions — sometimes it’s right, sometimes very wrong. Clients are realizing that while AI can help, it’s no substitute for professional advice. That said, forward-thinking clients are now choosing firms that use AI effectively. They see the value in efficiency and want to partner with teams who can integrate these tools into the workflow.

Steve Rathbone: I see a similar trend. For our clients — typically mid-market distributors or service providers — operational AI adoption is still early. But as use cases prove out, especially around logistics, purchasing, and data analysis, adoption will accelerate. Private equity firms are further along. Many now use AI tools, like platforms that assist with due diligence and contract review, to improve speed and accuracy. Corporates are also adopting AI for target identification and outreach, making deal sourcing more precise.

Kathleen Garenani: Our clients are at all stages — some just creating AI use policies, others developing sophisticated frameworks. We’ve built proof-of-concept AI models for private equity firms to analyze huge datasets and test investment theses. We also use AI in cybersecurity due diligence to identify data risks and recommend modernization opportunities.

Mike Silverman: Adding to that, we’re using AI to analyze data from hundreds of co-investments, benchmarking performance across sectors. It’s helping us make faster “no” decisions and improve efficiency. AI also helps summarize long investment memos — turning a 50-page internal report into a 3-page summary in minutes. The key is speed with oversight — letting AI draft, and then spending time refining.

Angela Raitzin: Let’s talk about the market impact — deal flow, multiples, and due diligence.

Steve Rathbone: AI is already affecting deal flow. It’s improving target identification and outreach, allowing more tailored, data-driven marketing. Deal sizes and multiples haven’t shifted much yet, but as AI demonstrably boosts company earnings or efficiencies, we’ll see that reflected in valuations. In due diligence, AI tools are making financial and data analysis much faster and more consistent.

Kathleen Garenani: We’ve seen that too. For example, we used AI to uncover billing anomalies in a healthcare acquisition — which led to renegotiations due to compliance issues. It’s a de-risking tool, not a magic bullet. We also help clients leverage IoT (Internet of Things) data to improve operational insights and valuation discussions.

Tal Hacohen: On the legal side, AI streamlines everything from diligence to closing. Tools like “Closing Folders” automate signature collection and document organization, saving clients both time and money. But again, humans must oversee the process — AI follows instructions but can’t ensure the instructions are correct.

Mike Silverman: For us, the goal is freeing up time for higher-value thinking. AI reduces the grunt work, letting professionals spend more time analyzing businesses, identifying risks, and developing insights.

Kathleen Garenani: Exactly — we ask clients, “What’s the worst part of your job?” Then we automate it so employees can focus on creativity and innovation instead of repetitive tasks.

Angela Raitzin: Kathleen, what “inning” are we in with AI adoption?

Kathleen Garenani: Early innings — maybe the first or second. Many organizations are just realizing how much AI they already use. Some are building internal models; others are simply creating data-protection policies. Agentic AI — where machines act autonomously — is still rare. Most companies are far from removing humans from decision-making.

Mike Silverman: I’d say we’re still stretching before kickoff! It’s very early. The tools are improving fast, but there’s still a learning curve in knowing how to use them effectively.

Tal Hacohen: Agreed. In some ways, AI has been around for decades — think of Spotify or Netflix recommendations — but in business applications, we’re still in the early stages. The technology evolves daily, and even experts admit yesterday’s presentation may already be outdated.

Steve Rathbone: Exactly. AI as a concept is mature, but as an everyday work tool, it’s new territory. I’ve made it a habit to constantly ask, “What can AI do for me right now?” That mindset helps uncover new efficiencies.

Angela Raitzin: Let’s turn to ethics and regulation. Mike, what’s your take?

Mike Silverman: We have to proceed very carefully. At Crescent, we never make investment recommendations based on ChatGPT outputs — that’s a hard line. AI is like a “friendly alien” — helpful but unpredictable, and it sometimes invents facts to please you. An example: when one of our consultants asked ChatGPT to summarize a new bill, the first version was completely wrong. A day later, it was accurate — because it had learned from legal briefings released overnight. That’s powerful, but also a reminder to be cautious.

Tal Hacohen: Right — always verify your sources. Even with internal tools, I check where the data comes from. It’s like traditional legal research — you’d still go back to the statute or case file to confirm accuracy.

Steve Rathbone: We’re mindful of compliance implications too, especially around data confidentiality. There have been cases where confidential deal information entered into ChatGPT accidentally leaked, emphasizing the need for private, internal AI systems.

Kathleen Garenani: Exactly. Responsible AI means adhering to privacy regulations like GDPR and the EU AI Act. Even U.S. companies must consider global data laws since data often crosses borders. There’s also the issue of AI manipulation — foreign entities generating fake content to mislead AI models. Governance and human oversight are essential to prevent bias and misinformation.

Angela Raitzin: Now the big question: Will AI replace us?

Tal Hacohen: I’m optimistic it won’t. It will change roles, not eliminate them. Junior professionals will need to learn how to use these tools effectively, while senior professionals will focus on strategic oversight. It’s a retooling, not a replacement.

Steve Rathbone: Agreed. AI will make teams leaner and more efficient, but not redundant. Those who learn to use it well will become even more valuable.

Kathleen Garenani: We’ll see new titles, like “AI Manager” or “Bot Supervisor.” Humans will oversee digital colleagues and ensure alignment with company goals. Agentic AI still requires human oversight — we can’t fully remove people from the loop.

Mike Silverman: Yes — there will be transitions, but I’m optimistic. The U.S. workforce is resilient. People will adapt, new industries will emerge, and innovation will continue.

Angela Raitzin: Final thoughts?

Tal Hacohen: AI is powerful, but remember: it’s still just a computer. Trust but verify.

Steve Rathbone: Automate where possible, augment everywhere else, and keep learning. If you’re not spending time understanding AI, you’ll fall behind.

Kathleen Garenani: Have internal advocates — people who champion AI adoption and celebrate small wins. And always keep humans in the loop.

Mike Silverman: Don’t rush to pick “the winner” in AI. Remember, in 2005 the top search engines were Microsoft and Ask Jeeves — not Google. The landscape changes fast, so proceed wisely.

Angela Raitzin: Perfectly said. Thank you to our incredible panel — and to everyone who joined us. Let’s go out there and keep learning new ChatGPT prompts!