You’ve seen the headline by now. 97% of finance leaders in VC and PE-backed companies are using AI, with three-quarters reporting ROI within twelve months. Impressive, right?
No. Not really.
Because the question was never “are you using AI?” — it was always “what are you actually doing with it?”
The 97% Number Is Meaningless Without Context
Let’s be honest about what “AI adoption” means in most finance departments right now. Someone installed Copilot. An analyst is using ChatGPT to summarise board packs. The FP&A team found a plugin that formats their Excel models faster.
That’s not transformation. That’s convenience.
It’s the equivalent of calling yourself “digital” because you moved your filing cabinet to SharePoint in 2015. The tool changed. The thinking didn’t.
The 97% figure tells us that AI has become table stakes — like having a laptop or knowing how to use a pivot table. It tells us nothing about whether these teams are fundamentally rethinking how finance operates.
Copilots vs. Architecture: The Real Divide
Here’s where the split is happening, and it’s widening fast.
On one side, you’ve got finance teams using AI as a copilot. It sits alongside existing workflows, making them marginally faster. Summarise this report. Draft this email. Clean this data set. The human is still the bottleneck — AI just lubricates the process.
On the other side — and this is a much smaller group — you’ve got teams building AI into the architecture of the finance function itself. Autonomous agents that monitor cash positions in real-time. Systems that don’t just flag variance but investigate it, pull the supporting data, and draft the narrative before a human ever looks at it. Governance frameworks that are designed specifically for agentic AI, not retrofitted from your SOX compliance playbook.
The difference isn’t speed. It’s operating model.
A copilot-enhanced finance team is still batch-oriented. They still run month-end. They still produce reports on a cadence designed around human processing time. An AI-native finance team operates continuously. The concept of “closing the books” starts to dissolve when your systems are reconciling in real-time.
What AI-Native Finance Actually Looks Like
I’m not theorising here. I run an AI assistant — Saul — that operates 24/7. It monitors my email, manages my calendar, tracks my investment portfolio, executes trades, scans news, and handles routine correspondence. It doesn’t wait for me to ask. It acts, escalates when needed, and learns from the outcomes.
That’s what AI-native looks like at the individual level. Now scale that to a finance function.
Imagine a portfolio company where the finance team’s AI agents are handling bank reconciliations autonomously, flagging only genuine exceptions. Where cash flow forecasting updates continuously based on real-time revenue data, not last month’s actuals plugged into a spreadsheet. Where the CFO’s morning briefing isn’t a deck someone spent three hours building — it’s a synthesised intelligence report generated overnight from live data sources.
This isn’t science fiction. The technology exists today. The gap is in the willingness to let go of the old operating model.
PE Firms Are Asking the Wrong Question
When a PE firm conducts due diligence on a portfolio company’s finance function, the question “do you use AI?” is already obsolete. Everyone uses AI. The answer is always yes.
The right questions are harder: What’s your AI architecture? Which workflows are fully autonomous vs. human-in-the-loop? What’s your governance model for agentic systems? How does your finance function operate differently today than it did eighteen months ago — structurally, not just faster?
KKR has already flagged this concern — that AI capability gaps could create a meaningful split in exit outcomes. Portfolio companies that have genuinely integrated AI into their operations will command premium multiples. Those that bolted on a chatbot and called it transformation will not.
This is the real game-changer in PE-backed finance: not whether AI exists in the business, but whether it’s load-bearing.
The CFO Role Is Splitting in Two
The 2026 CFO agenda looks fundamentally different depending on which side of this divide you’re on.
One version of the CFO sees AI as a tool in the toolkit. Useful. Saves time. Makes the team more efficient. They’ll adopt it incrementally, bolt it onto existing processes, and measure success by how many hours it saves per month.
The other version sees AI as infrastructure — as fundamental to the finance function as the ERP system or the chart of accounts. This CFO is redesigning processes around AI capabilities, not adapting AI to fit legacy processes. They’re thinking about data architecture, agent orchestration, and continuous assurance — not just “can we automate the board pack?”
PE operating partners need to know which type of CFO they’ve got. Because the incremental adopter will deliver incremental value. The infrastructure thinker will deliver step-change capability. And in a compressed hold period, that difference matters enormously.
The Competitive Moat Isn’t Adoption — It’s Depth
When 97% of your peers have adopted the same technology, the technology itself is no longer a differentiator. The moat moves downstream — to depth of integration, quality of data architecture, sophistication of governance, and willingness to let AI operate autonomously within defined boundaries.
Most finance teams are wading in the shallows. They’ve got AI, sure. But it’s supervised, constrained, and fundamentally optional — remove it tomorrow, and the function still operates the same way, just slower.
The teams that will win are the ones where AI removal would be structural. Where the operating model has been redesigned so thoroughly that the AI isn’t an enhancement — it’s a dependency. Not because of recklessness, but because the architecture is sound, the governance is robust, and the results speak for themselves.
97% adoption is the starting line, not the finish. The race that matters hasn’t even begun for most.
Leave a Reply