Judgment, Amplified: How AI Is Changing Private Capital
- Monz Uddin
- May 2
- 4 min read
By Monz Uddin, Co-Founder, MUSU AI Advisory
Three things have happened in the last few weeks that should land on every senior desk in private capital.
SpaceX, now merged with xAI under a $1.25 trillion valuation, struck a deal with Cursor that includes the right to acquire the AI coding company for $60 billion by year-end. Cursor was valued at roughly $30 billion at the end of last year. Anthropic announced Claude Mythos, a model so capable at autonomously finding software vulnerabilities that the company restricted its release to a consortium including JPMorgan, Microsoft, AWS and Google; Anthropic is reportedly fielding investor offers at an $800 billion valuation. And NVIDIA, whose chips this entire era runs on, took a $5 billion stake in Intel, its longtime competitor, to co-design the next layer of AI infrastructure.
Step back from any one of these and the pattern is the same. Capital is no longer just chasing AI infrastructure. It is moving rapidly into the application layer: the agentic systems that read documents, build models, and execute multi-step work autonomously. The market has spent three years pricing the chips. It is now pricing the applications that run on them.
That shift matters acutely for private capital. The work that defines a GP is exactly what today’s agentic systems are built to do: synthesising proprietary diligence, monitoring a portfolio in close to real time, drafting investor communications grounded in unique datasets. According to KPMG’s 2025 M&A Pulse Survey, roughly nine in ten PE dealmakers already use GenAI or agentic AI somewhere in their workflow. The window where this is a competitive choice is closing fast.
I’ve spent twenty years moving across capital markets at Citi in New York, data engineering at Sun Life in Toronto, and AI deployment work in the Gulf. Most senior people I talk to in private capital are still underestimating how quickly the agentic leg lands inside their own firms. They look at a saturated AI conversation, conclude the story is mostly priced, and turn their attention elsewhere. The conversation is saturated. The story is not priced. Not for us.
Where This Shows Up in Private Capital
What separates today’s tools from chatbots is autonomy. They plan, call other tools, read across documents, and produce structured outputs. Here is what that looks like in practice.
In screening and origination, agents produce a baseline view on a target company, sector, or fund manager in hours rather than the week it used to take an analyst team.
In due diligence, document review compresses dramatically. Agents surface inconsistencies across a data room faster than a junior team can, and produce first-draft analyses that let senior people focus on judgment instead of retrieval.
In portfolio monitoring, agents handle the messy reality of receiving differently formatted reports from twenty portfolio companies and producing something consistent on the other side.
In LP and IR work, the time it takes to produce a tailored investor letter or a careful response to an LP query drops materially. Agents pull from fund documents, board decks, and prior correspondence to draft within minutes.
In compliance, agents speed up the work of mapping shifting regulatory expectations against where your firm actually sits. That becomes more pointed as regulators turn their attention to AI use itself.
None of this eliminates human judgement. It amplifies it.
The Part That Doesn’t Get the Honest Treatment It Deserves: Governance
Most firms I talk to already have AI in their environment. They just don’t always know it. Some of it is sanctioned. Most of it isn’t. Senior people are pasting sensitive material into public AI tools to draft memos, summarise diligence documents, and pressure-test their thinking. There’s often no audit trail and no clear institutional view on what data has just walked out the door.
The exposure is real: information leaks, IP exposure, no retention or access control, and if an investment later goes wrong, no defensible record of how AI shaped the decision.
The institutions that benefit most from AI aren’t the ones that deploy the most tools. They’re the ones that put discipline in place first: a clear policy, a controlled environment, sanctioned use cases, and a governance frame that lets the firm capture productivity without taking on risk it can’t manage. The upside gets captured through governance, not around it.
Three Things Worth Doing This Quarter
First, find out where AI is already being used inside your firm. Not the official answer. The actual answer. Fifteen minutes with five senior people will tell you more than any vendor pitch.
Second, pick two or three workflows where agents give you the highest-value, lowest-risk gain. For most private capital firms that’s somewhere in due diligence, LP communication, or internal research. Sequence from there.
Third, set the governance perimeter before you scale. This isn’t a compliance exercise. It’s a fiduciary one. The firms that put it in place early move faster later, not slower.
The firms that win the next decade in private capital won’t be the ones with the most AI tools. They’ll be the ones whose investors and decision-makers know how to use the intelligence now available to everyone, with discipline, with governance, and pointed at the work that actually matters. That’s the proposition we built MUSU around. Strategy first, technology second. Governance built in, not added afterwards. Judgment amplified, not replaced.
MUSU AI Advisory is a small, independent firm working with private capital institutions on AI assessment, governance, and adoption strategy. We are model-agnostic and have no vendor partnerships. If you’d like to compare your firm’s posture against peers, our self-assessment tool launches shortly. Write to monz@musux.ai for early access.
About the Author
Monz Uddin is Co-Founder of MUSU AI Advisory. He has held capital markets business development roles at Citi in New York, led data engineering at Sun Life in Toronto, and run AI implementation engagements in the Gulf. He holds an MBA from Cornell and a BSc in Computer Science from the University of Toronto.
Comments