There’s a quiet frustration building across private equity and asset management firms right now.
The tools are there. Teams have tried ChatGPT. Microsoft Copilot may already be available. Vendors have added AI features into existing platforms. The demos are impressive. The investment has been made.
And yet… not much has changed. This is not because firms are ignoring AI. Across industries, AI usage is already widespread. McKinsey’s 2025 global AI survey found that 88% of organizations are regularly using AI in at least one business function. But the same research found that most firms are still in experimentation or pilot mode, and only 39% report EBIT impact at the enterprise level. [McKinsey Report]
In private equity, the pressure is already here. Deloitte’s 2025 GenAI in M&A survey found that 86% of corporate and PE respondents have integrated GenAI into M&A workflows, and 83% have invested $1 million or more specifically for M&A teams. The question is no longer whether firms are experimenting. The question is whether those investments are changing how work actually gets done. [Deloitte Report]
DDQs are still being assembled manually. Investor reporting still depends on spreadsheets, email threads, and shared drives. Deal teams still search through old CIMs, notes, and data room files to find the right information. Portfolio updates still require too much manual follow-up. Compliance approvals still move through informal back-and-forth.
The workflows still look the same.
So what’s going wrong?
Most firms assume the issue is the technology itself. Maybe it’s the wrong platform. Maybe it wasn’t implemented properly. Maybe the use cases weren’t ambitious enough.
But in most cases, the real issue sits somewhere else entirely.
AI doesn’t fail because of the tool. It fails because of the system around it.
AI is not a plug-and-play solution. It’s a layer that depends on the quality of what already exists beneath it—your data, your workflows, and how your team actually works day to day.
If those foundations are fragmented, inconsistent, or unclear, AI doesn’t fix the problem. It exposes it.
We’re seeing three common patterns:
1. Fragmented data environments
Critical information lives across Excel models, CRM notes, fund admin portals, shared drives, data rooms, email threads, old DDQ responses, policy documents, and individual desktops.
There’s no single source of truth, which means no clean way for AI to interact with that data.
This creates a simple problem: if your team does not know which document, system, or version is the source of truth, AI will not know either.
That is when AI starts producing answers that feel impressive but are difficult to trust.
This is one of the biggest blockers in private equity. Pictet’s analysis of AI adoption in PE found that data and output quality are viewed as the biggest barriers to adoption, with privacy and cybersecurity also creating significant concern. [Pictet Survey]
2. Undefined or inconsistent workflows
Ask five people how a process works, and you’ll get five different answers.
One person knows how DDQ responses are drafted. Another knows which compliance language needs approval. Someone else knows where the latest cybersecurity policy lives. A senior partner remembers the context behind a portfolio company update, but that context was never written down.
AI requires clarity.
At this stage, many firms have not fully documented how work actually gets done. The real process lives in people’s heads, inboxes, and habits.
That makes automation difficult because AI cannot reliably support a workflow that the firm itself has not clearly defined.
3. No clear ownership of transformation
Technology is introduced, but no one is truly accountable for embedding it into daily workflows.
IT may own the platform. Operations may own the process. IR, deal teams, finance, and compliance may all touch the workflow. But if no one owns the full transformation, adoption becomes optional.
That is when AI becomes another tool people experiment with, instead of a real operating improvement.
For AI to work, ownership cannot sit only with technology. It needs business ownership from the people who understand the workflow, the risk, the approvals, and the expected outcome.
The firms that get this right treat governance as part of the operating model, not an afterthought. EY’s 2025 Responsible AI research found that organizations with stronger monitoring and oversight practices are more likely to report revenue growth, cost savings, and productivity gains. [EY Survey]
The expectation gap
There’s also a mismatch between expectation and reality.
AI is often positioned as a leap forward. In practice, it’s a multiplier.
If your underlying systems are strong, AI accelerates them. If they’re not, AI exposes every inefficiency.
That’s why some firms are seeing real gains while others are questioning whether it’s worth the effort at all.
The cost is not just wasted software spend. The bigger cost is that teams continue losing time chasing information, senior people stay involved in low-value review work, investor responses take too long, reporting cycles remain painful, and answers are difficult to trace back to source documents.
For investment firms, that matters.
Productivity is important, but so are control, consistency, and auditability.
What successful firms are doing differently
The firms starting to see meaningful results aren’t necessarily investing more in technology.
They’re taking a different approach.
They start with workflows, not tools.
They choose one high-friction process first, such as DDQs, investor reporting, deal screening, portfolio monitoring, side-letter tracking, or compliance review.
They map how the work actually happens today.
They identify the source documents and systems of record.
They define where human review is required.
They clarify who owns the process.
They measure time saved, risk reduced, consistency improved, and adoption by the team.
Most importantly, they recognize that AI is not just a technology initiative, it’s an operational one.
The research points in the same direction. McKinsey found that AI high performers are nearly three times more likely than others to have fundamentally redesigned individual workflows. That is the important distinction: successful firms are not just buying AI tools. They are redesigning how work moves through the firm. [McKinsey Report]
A shift in mindset
We always advise clients that the question shouldn’t be: “Which AI tool should we use?”
Instead, it should be: “Are we set up to actually benefit from AI in the first place?”
Until that question is addressed, even the best tools will fall short.
We usually start with one practical question:
Where is your team still doing high-value work through low-value manual steps?
From there, we review the workflow, the data behind it, the approval points, the risks, and the people involved. The goal is not to force AI into the firm. The goal is to find where AI can safely reduce manual work, improve consistency, and give teams more control over the process.
Before choosing another AI tool, firms should assess whether their workflows, data, and ownership model are ready for AI.
That is where real AI readiness starts.