What AI Can't Automate in Dealmaking (And Why the Phone Is Back)

Every serious dealmaking shop now runs on the same sourcing databases, the same AI drafting tools, and increasingly the same playbook. When everyone has the same inputs, the inputs stop being the edge. What AI cannot automate in dealmaking is the part that was never really about tools in the first place: getting a person to trust you enough to take the call, and being there, in person, when it counts.

That is not a nostalgic argument for putting the tools down. It is a practical one about where the value in sourcing actually sits, and why the firms getting the most from AI right now are the ones being deliberate about what they refuse to automate.

Why has deal identification stopped being a source of edge?

Because everyone is drawing from the same pool. The sourcing databases have spent the better part of a decade building comprehensive, searchable records of private companies, and most serious firms now subscribe to at least one of them. If two firms searching for the same thesis pull up largely the same list of targets, the identification step alone tells you very little about who is going to win the deal.

That was already true before AI changed anything. It just means the differentiation has to come from somewhere else now: what you do once you have the list, not how you built it.

Why is cold outreach losing its edge?

Because AI pushed the cost of writing a convincing, personalised-sounding email to nearly zero, and every firm with the same database now has the same capability. Sourcing tools advertise, as a headline feature, the ability to draft a tailored outreach message for every company in a list in a single prompt. That capability used to take an analyst an afternoon per company. Now it takes a few seconds for a thousand.

A business owner who was already fielding twenty or thirty unsolicited buyer emails a week before this shift is, almost by definition, fielding several times that now. What lands in the inbox looks like personal attention. What the owner experiences is noise. And the moment every firm is pulling from the same database, running the same kind of sequence, and using the same AI to write the same “came across your business in my research” opener, the list itself stops being a differentiated asset and becomes a commodity, indistinguishable from anyone else’s.

AI is a multiplier on whatever direction you point it. The value was always in the bit that couldn’t be automated.

Why is the phone call back?

Because it is one of the few channels AI cannot touch, which makes it more valuable exactly as digital outreach gets cheaper. Some of the more effective sourcing teams have made a deliberate pivot: dropping database subscriptions that were used mainly for automated email sequences, keeping only what is needed for market-mapping, and moving the bulk of origination effort back to the phone. A short, low-effort email still gets sent as a multi-touch anchor, mostly so the prospect recognises the name later, but the actual work of origination happens through a live conversation, often run by some of the best-paid junior seats in the business. Teams making this shift have reported the phone becoming their single most productive origination channel in a given period, ahead of database subscriptions or email sequences.

That is the general pattern: the cheaper and more commoditised digital outreach becomes, the more valuable the channels AI cannot replicate become in relative terms. The phone call. The industry conference. The operator dinner. The trade show. None of these scale. That is precisely why they still work.

Does personal brand actually matter for sourcing?

Yes, because the first thing a prospect does after any cold approach is search your name. The common objection is that the owners a firm wants to reach are not spending time on LinkedIn, so building a public profile there feels like effort spent on the wrong audience. Two things are worth holding at once here. More owners and operators are paying attention than most dealmakers assume. And even for the ones who genuinely are not, the profile still does work, because the moment someone gets a cold call or cold email, they go and look the sender up.

What they find in that search either helps or hurts. A profile with thoughtful, specific commentary on the market and how it is changing gives a prospect a reason to take the second call. An empty profile, or none at all, gives them nothing. The brand does not need to reach the owner directly through their feed. It needs to be there, credible and specific, the moment they go looking, which they will.

What’s the deeper pattern behind all of this?

The instinct with any new AI capability is to do more of whatever you were already doing. More outreach emails, faster memos, more analysis run against more targets. An investment committee memo that used to take five hours and now takes five minutes tends not to get five times better. It tends to get written five times, against five times as many deals, most of which were never that promising to begin with. The same logic applies to outreach: an email that used to take an analyst an hour gets fired at ten thousand companies before lunch, and the signal that used to exist in a smaller, more considered list disappears into the noise of everyone else doing the same thing.

The memo that actually gets read by an investment committee is usually the short, sharp one, not the longest. The email that gets a reply is usually the one that references a warm introduction, not the best-worded cold opener. The deal that closes is often the one where a partner met the owner in person at an industry event a year earlier, or picked up the phone directly, or showed up credibly on the owner’s radar well before the process started. AI is a multiplier on whatever you point it at. Pointed at more volume of low-signal activity, it produces more low-signal activity, just faster. Pointed at the parts of the job that free up time for a call, a coffee, or a decade-long relationship, it compounds in your favour instead.

The value in sourcing was always concentrated in the parts that do not scale: the conference conversation, the phone call, the relationship built over years that means you hear about a deal before a banker is ever engaged. That was true well before AI arrived and it remains true now. Getting real value from AI in dealmaking is as much a question of what you deliberately choose not to automate as it is a question of what you do automate.

If you are rethinking where AI fits into your own sourcing process, talk to us about how the platform is built to handle the identification and tracking work, precisely so your team’s time goes back into the calls and relationships that actually close deals.