Somewhere in the prehistory of intelligent dealmaking sits a bag of paper. Not a tidy archive box or a color-coded binder. A bag. Years ago, Merlin Piscitelli described clients handing document teams assortments of old records, the stubborn raw material from which a modern transaction had to be built. The optical character recognition system could read typefaces on deeds from the 1800s. It could pull each character from the page, sharpen a bad scan, and turn disorder into something searchable. Signatures, with all their private flourishes, remained another matter.
It is a useful scene to keep nearby when the conversation turns to agentic AI. Before a machine can reason across a transaction, the transaction has to become legible. Before it becomes legible, somebody has to open the bag.
Piscitelli has spent nearly two decades watching this hidden work climb the corporate agenda. He joined the business now called Datasite in 2006, when a virtual data room was still a fairly literal proposition: a controlled digital place where buyers, sellers, bankers, and lawyers could inspect sensitive files. He went on to run international sales, lead sales across Europe, the Middle East and Africa, and serve as the region's chief revenue officer. In 2025, he moved to New York and into the chief marketing officer job, responsible for telling the global story of a company whose product was learning to do more than store the evidence.
The salesman learns the territory
There is a compact education hiding inside one recommendation on Piscitelli's LinkedIn profile. Paul Hartzell, a former colleague, recalled Piscitelli arriving in London with little knowledge of Europe, little knowledge of M&A markets, and little knowledge of the software. Piscitelli answered those deficits in the most analog fashion available: travel, meetings, and practice. Within six months, Hartzell wrote, he knew his territories, could perform live demonstrations for demanding clients, and had begun building a formidable pipeline.
The anecdote strips sales leadership of its mythology. A new continent is learned one train, airport, meeting room, and difficult question at a time. The product demo is where confidence meets the possibility of public embarrassment. It rewards the person who knows which button to press and, more importantly, why anyone in the room should care.
That apprenticeship became unusually durable. Piscitelli stayed while the company around him changed shape. Merrill Corporation had begun as a financial printing business in 1968. Its data-room application arrived in 2002. By 2018, the business was concentrating on digital tools for capital transactions; it later adopted the Datasite name. The old work of moving physical information did not vanish. It was translated into indexing, access controls, audit trails, search, and software.
One company, a widening field of view
By 2022, Piscitelli described a platform that stretched beyond due diligence into deal sourcing, preparation, acquisition, and post-merger integration. He also offered a number that explained the urgency: Datasite's revenue had risen 37 percent across the prior two years. Remote work had pushed dealmakers toward more technology, and the company had introduced three products while expanding its teams. His pitch joined growth to relief. Automate the outreach, categorization, indexing, and redaction, he argued, so professionals can return to strategy.
“Time in M&A deals really matters.”Merlin Piscitelli, discussing the foundations of deal technology in 2019
A marketer with a seller's memory
The move from chief revenue officer to chief marketing officer makes intuitive sense when viewed from the client end of the table. Piscitelli spent years hearing how deal teams describe pressure when nobody is polishing the language. He managed close to 100 EMEA sales representatives and helped open Datasite's second UK office in Manchester. He discussed London as an IPO destination, compared it with New York and financial centers in Europe and Asia, and wrote about markets that could seem quieter in public than their underlying deal launches suggested.
Sales asks whether the buyer believes the promise. Marketing decides which promise deserves the microphone. Piscitelli's public language tends toward practical verbs: categorize, index, redact, track, validate. Even his enthusiasm has an operational pulse. When the London office reopened and his sales team gathered again, he called himself the “luckest leader alive,” typo intact, then finished with “Let's Go!!!” The polish of the current C-suite role has not entirely subdued the sales floor.
His colleagues describe the same bias toward action. Another recommendation says he leads from the trenches and manages by example. Such endorsements are naturally affectionate, but together they explain the energy behind a long tenure. Remaining at one company for twenty years can look like constancy from a distance. Up close, Piscitelli changed countries, territories, jobs, customers, and the central technology story several times.
The machine may read. The human must answer.
AI has given Piscitelli a larger stage for an old argument. A deal room now can do more than wait for a keyword. On the sell side, software can help bring content in, structure and permission the room, and prepare materials for release. On the buy side, it can assist diligence, surface concerns, and help create investment-committee materials. The prize is cognitive space: fewer hours manipulating files, more hours understanding what those files mean.
The limit matters as much as the capability. In Datasite's 2026 discussions of AI-driven M&A, Piscitelli repeatedly placed judgment, relationships, and accountability with people. The system can accelerate a question. It cannot accept professional responsibility for the answer. That distinction may sound obvious until a deadline is tight, an output is fluent, and billions of dollars make convenience particularly seductive.
“Trust matters in dealmaking because we're dealing with the most sensitive, confidential information.”Merlin Piscitelli on AI inside the deal room
His preferred word is “trusted,” and he uses it insistently: trusted environment, trusted organization, trusted partner. In another business, repetition might sound like a campaign. In M&A, confidentiality is the atmosphere. The documents reveal a company before the market knows it is for sale. Permissions are strategy. A leak is not an awkward software moment; it can alter negotiations, invite regulatory trouble, or end the transaction.
That is why Piscitelli's version of the new deal team is “AI-informed, but human-led.” It avoids the theater of pitting worker against machine. The interesting question is where each creates value. Process-heavy work can move toward software. Sourcing, diligence, and analysis can gain mechanical assistance. People receive more room for skepticism, negotiation, and the social work by which a transaction becomes acceptable to those who must live with it.
A 2026 survey discussed by Datasite found that 62 percent of dealmakers believed human-only decision-making was no longer defensible in complex transactions, while seven in ten thought firms ignoring AI would struggle to compete within five years. Piscitelli's response was not to surrender the wheel. It was to design a better dashboard: permissions, citations, audit trails, source-backed answers, and validation. Intelligence becomes valuable when a professional can inspect how it arrived.
The useful kind of futurist
In August 2026, Piscitelli joined the Deal Lawyers Download podcast to discuss AI as a force multiplier, the legal risks of adoption, and the human-in-the-loop model. The phrase is technical and slightly bloodless, but the idea is personal. Keep somebody in the loop because somebody must care enough to doubt, check, persuade, and be accountable.
His career gives that position more weight than a fashionable caution attached to a new product. Piscitelli remembers the blurry scan beneath the model and the client who turned up with the bag. He has watched faster search become automated indexing, and the repository become a workflow platform. He knows that technology's clean interface often rests on somebody else's mess.
The CMO's current task is to make an AI story credible in an industry professionally suspicious of loose information. His advantage is that he has sold the less glamorous layers. The next chapter of the deal room may begin with a question typed in plain language. The answer still rests on two decades of older questions: Is the information complete? Who is allowed to see it? Where did this conclusion come from? Who will put a name beneath it?
A machine can now help open the bag, read the deed, sort the files, and point toward an inconsistency. Piscitelli's proposition is that this should give the humans more time to do the dangerous, consequential, and occasionally delightful thing machines cannot do for them: decide. The paperwork moves on. The responsibility stays seated at the table.