The State of the Stack: How to Advocate for AI at Your Firm

As Meridian’s strategy and sales director, I take a lot of demo calls. About a quarter of them go the same way…

A VP or senior associate gets on the call. They know their stack is legacy. They've read enough, used enough Claude on the side, and seen enough of what's possible that they’re curious to learn more. What they don't have is a mandate. No firmwide champion, no budget line, no partner who's said yes. Just someone who sees the gap and doesn't yet have the authority to close it.

I used to think that was a Meridian problem, a sales funnel stuck in the middle. Now I know it's not. It's the actual state of AI adoption in private markets right now. The technology isn't the bottleneck. The internal case is.

Two different objections you’ll encounter when making the case for AI for your firm

If you're the person trying to make that case, you're usually up against one of two people: the bubble skeptic and the wait-and-see peer. These are not the same person, even though they sound alike in a partner meeting.

The first is the bubble skeptic. Often senior, often on the investment committee. They've lived through cycles. They remember the dot-com crash, and they watched crypto come and go, and they hear "AI" and think "overhyped asset class," not "operational tool." Their objection isn't really about your firm's workflow. It's about capital allocation and reputational risk. They don't want to be the partner who championed the thing that got written up as a cautionary tale.

The second is the wait-and-see peer. This person usually doesn't doubt that AI works. They've probably used it themselves. Their hesitation is political, not technological. They don't want to be the one who sticks their neck out, pushes for adoption, and is wrong, or worse, looks like they chased a trend that the firm then quietly walks back. Their risk isn't capital. It's being the name attached to the initiative if it stalls.

Both objections are reasonable on their face. Both are also based on a mistake, and it's the same mistake I see time and time again coming from two different directions.

The mistake is conflation

The bubble conversation happening in public markets right now is a conversation about valuations. Nvidia's multiple. Hyperscaler capex. Circular financing arrangements between AI labs and the infrastructure providers building their data centers. That's a real conversation, and reasonable people disagree about how it resolves. But it's a conversation about whether AI infrastructure companies are correctly priced. It is not a conversation about whether using AI tools inside a deal workflow creates value for an investment team.

The valuation conversation and the AI workflow conversation present two different questions. A partner who says "let's wait for the bubble to pop" is answering the question of AI workflow with an argument about AI infrastructure valuation. The stock market's opinion of Nvidia has no bearing on whether your associates should be using AI to synthesize diligence materials faster. Short answer: They should be! And Meridian’s automatic data extraction for CIMs makes this a breeze!

And here's the part that should worry the wait-and-see camp more than the bubble camp. The industry isn't actually stuck at the "should we adopt AI" stage anymore. It's stuck in a riskier stage.

The AI gap that actually matters in a PE firm

Nearly every PE sponsor has already told their portfolio to move into the age of AI. In fact, 98% of PE sponsors have mandated AI adoption across their portfolio companies. Only about half of those portfolio companies are actually implementing it (Accordion).

What’s worse is PE shops aren’t practicing what they preach. In finance functions, specifically, the gap between mandate and practice is larger. Fewer than one in three PE-backed CFOs have meaningfully implemented AI, and more than two-thirds say they don't know where to start (Accordion).

So, the mandate exists almost everywhere, but the execution exists only in about half of those firms, and less than a third when you look at finance specifically. That gap — between what leadership has already said yes to and what's actually happening day to day — is the real state of the stack right now. It's an execution gap, and execution gaps are won by whoever moves first and moves competently, not by whoever waits longest.

‍This execution gap should reframe the conversation for both types of skeptics. 

• For the bubble skeptic: the decision to adopt AI operationally isn't a bet on a market thesis; it's catching up to a mandate that's probably already been issued somewhere in your own firm's portfolio guidance. 
• For the wait-and-see peer: waiting doesn't protect you from being wrong. Most firms with a mandate are already behind on execution. Sitting still doesn't put you in a neutral position. It puts you in the 50% that talked about it and didn't do it.

There's also a version of this that should worry the skeptics of both stripes for a completely different reason. Firms that have already moved are pulling ahead because they adopted more deliberately. High-performing firms aren't using AI at materially higher rates than everyone else, they're just far more likely to have exceeded their own business case for it. Discipline is the differentiator. Which means waiting just delays the inevitable starting point.

Of course, not every AI initiative is working. Some Ops teams are still stuck testing tools that never make it to production. That's a legitimate risk, and it's worth calling out. But it's an argument for adopting well, not an argument for not adopting. The firms that failed are the ones that bolted AI onto a broken workflow instead of redesigning the workflow around it. And that’s why we take implementation so seriously at Meridian: The quality of the tool is only as good as the infrastructure to support it. 

What actually works when you make the case for AI

If you're the VP or senior associate on that demo call with me, or in the equivalent conversation inside your own firm, here's what tends to actually move a skeptic in either camp. Follow these steps to advocate for thoughtful, meaningful AI adoption: 

1. Start with a bounded pilot, not a platform decision. Nobody needs to approve a new CRM to test whether AI-assisted memo drafting saves three hours on the next deal. Ask for the smallest possible yes.

2. Speak to the specific fear, not a generic pitch. The bubble skeptic needs to hear that this isn't a market bet, it's an operational one. The wait-and-see peer needs to hear that the risk of moving is lower than the risk of being in the 50% that didn't. Don't give both people the same slide.

3. Bring a number from a peer firm, not an abstraction. "AI is the future" doesn't move anyone in an IC meeting. "98% percent of sponsors have mandated this and half the portfolio hasn't executed" does, because it reframes the ask from speculative to overdue.

4. Frame it as workflow redesign, not tool adoption. The failure risk isn't AI. It's AI bolted onto a process that was already broken. If you're asking for a tool, ask alongside a specific process you're trying to fix.

5. Find the person who already owns the stack question, or become that person. Every firm executing well on this has someone, often not the most senior person in the room, who's asking how a new tool fits before the purchase, not after.

The firms that succeed with AI are the ones that stopped treating adoption as a debate.