The Most Useful Voice at the AI-in-M&A Summit Had Nothing to Sell

I sat through an online summit where 16 different companies pitched AI tools for M&A. Sourcing engines, diligence copilots, contract readers, full lifecycle platforms. Slick demos, real founders, genuine progress.

And the single most useful 30 minutes had nothing to sell.

It was a finance guy who built his own.

This was the "AI in M&A" summit run by Kison Patel under his M&A Science brand. Kison founded DealRoom and M&A Science, and M&A Science spun out as its own company in 2025. Worth your time if you want a fast read on where the vendor market is heading.

The lineup was 16 independent companies, no one stop all encompassing suite, sorted into five jobs to be done:

Each one is solving a real piece of the deal. If you run a process today, several of these would save your team hours. I am not knocking the tools.

But the spine of the day, the part I keep thinking about, came from the one speaker who was not selling anything.

John Palusci is VP of Strategic Finance at Schweiger Dermatology Group. Schweiger runs roughly 120 locations across nine states and does about 5 to 8 acquisitions a year. So this is a real buyer running real volume.

Over about two months of nights and weekends, Palusci said he built his own diligence setup on a HIPAA-compliant enterprise Claude account. Custom skills for P&L analysis, EMR review, and quality-of-earnings prep, wired to work together. He bought no third-party tool. He built the workflow himself.

His line was simple: "AI is for everyone; you don't need to be technical."

The insight that stuck with me was sharper. He said the smallest deals, targets with $1 to $5M in revenue, are where AI pays off most. Why? The diligence work is roughly the same regardless of deal size. On a $3M practice, that fixed effort used to be hard to justify. Build it once with AI and the floor on what is worth chasing drops. For a buyer doing 5 to 8 deals a year, that changes the funnel.

So picture the room. A dozen-plus vendors pitching workflow software, and the one operator on the agenda who wasn't selling anything quietly explaining that he assembled his own over a few weekends and bought none of it.

That does not make the vendors worthless. It makes the question different.

The lesson for a healthcare buyer is not "go buy one of these 16 tools." The lesson is to be honest about what you are actually paying for.

A capable person on your team, with an enterprise AI account and a few weekends, can build the workflow. Palusci did, and bought none of the tools that pitched that day.

Do not misread that as "it is all generic." It is not. He spent two months of nights and weekends, and the skills he built were EMR-aware and HIPAA-compliant from the first line, because the data is protected health information a generic workflow would never touch. That is real work: defining the task, engineering the context, setting the healthcare guardrails. The point is not that the workflow is worthless. It is that one capable person can now stand it up, which is exactly what stops it from being the scarce thing.

So if the workflow is not the scarce thing, what is? The inputs it runs on. A workflow can normalize EBITDA and prove out cash in any deal. What it cannot do is hand you the right numbers to put in. That part is built over years and relationships, not a weekend:

Then the judgment to read it: which signal matters, which adjustment is aggressive, which comp is a trap. The inputs are the durable asset; the judgment is how you use them. A workflow tool cannot manufacture either. And no, AI will not just make the research cheap too. It can pull a comp; it cannot tell you the EBITDA behind it was engineered by a seller you have never met. That gap is the work.

I have been on the buy side for 130-plus healthcare center acquisitions and more than a billion in deal volume, and the pattern holds. The tooling was never the hard part. The hard part was knowing which targets were worth the effort, what the numbers were hiding, and what a given signal meant in healthcare specifically. AI makes the easy part faster. It does not hand you the hard part.

Before you pay for any of these tools, ask one question: what does this give me that someone on my team could not assemble in a weekend?

If the honest answer is "a workflow," you are paying for something on its way to commodity. Someone on your team can build it the way Palusci did and own it outright. Not saying that is easy. But it is doable, and it is yours.

If the answer is real proprietary inputs, a curated universe, comps somebody actually researched, an ownership map, the healthcare read on reimbursement and regulatory risk, then you are buying something the workflow cannot manufacture for itself. That is worth paying for, and it is a different conversation than a software subscription.

One thing I should say out loud, because it is the obvious objection. I sell research, so of course I am the one telling you the research is the valuable part. Do not weigh that on my say-so. Weigh it on what the buyer in the room did: he built the whole workflow himself, bought nothing, and still could not self-build the curated universe, the real comps, and the ownership and reimbursement read a deal turns on. That is the tell.

If you are sitting on a tooling decision, reply to this email and tell me what you are looking at. No pitch. I am happy to give you my honest read on whether you should build it, buy it, or whether the research is the part worth paying for.

-Shawn

Book a call: cal.com/healthcaremai

This newsletter is for informational purposes only and does not constitute investment, legal, or financial advice.