Laying the Foundation for Autonomous Private Capital. Step Two: Removing Human Bottleneck.

This is Part 2 of SecondLane’s series on how autonomous agents are rebuilding the infrastructure of private capital. In Part 1, we learned to meta-prompt. In this part, we are moving from the prompting to the autonomous infrastructure.

In 2025, the private capital world has rallied around a comforting narrative: “AI won’t replace the investor; a human using AI will.” This “co-pilot” era feels safe because it implies that the deal partner is still the captain. However, this is a dangerous delusion. We are in a transition phase, not an end state. 

Currently, firms insert humans into AI workflows simply because they provide a liability buffer. If something goes wrong, from the management point of view, it’s not AI’s fault – it’s the human overseer’s fault. Yet, this oversight effectively hinders the essence of AI innovations. Hg’s David Toms puts it like that: “Don’t put an adult in the room,” he says, “because the adult will start saying, ‘Why don’t you do this with the blocks? Why don’t you try this?'”, forcing old human workflows on AI.

The most competitive firms operate in the “dark factory” era, where human intervention is latency and digital labor produces alpha 24/7.

Even in transition phase, we see that human capital is becoming the primary drag on net IRR. The most competitive firms are moving toward the end state, or the “dark factory” era — with autonomous back offices where human intervention is viewed as latency, and digital labor produces alpha 24/7.

Eliminating the Monday Morning Bottleneck
To understand the volume of dragging, let’s look at the  cost of human latency impacting the current state of the market. According to Bain & Company’s 2025 Global Private Equity Report, the industry is choking on a record $3.2 trillion exit backlog. Partially, it is impacted by traditional investment committee cadence, where deal teams spend five days scrubbing data rooms, summarizing calls, and formatting slides to prepare for a Monday meeting. Speed of these manual processes dictates the velocity of capital deployment.

In a dark factory model, this latency evaporates. An agentic architecture can ingest a target’s data room, cross-reference it with the fund’s criteria, and generate a risk assessment in minutes. The firm shifts from a weekly decision cycle to a continuous decision cycle, which changes the game for secondary market participants, dependent on time-sensitive pricing liquidity. 

The tangible effects of the AI tools are already visible. Firms like Apollo Global Management have moved beyond pilots, creating institutionalized centers of excellence, and their portfolio companies are seeing costs down 40% in content production and 20% in lead generation.

Automating the “No”
The most immediate application of AI agents for investors is the sourcing engine. VCs usually rely on human networks to filter deal flow, which limits their scope to who they know. A dark factory approach replaces human filtering deals with autonomous agents that scan the market to flag opportunities.

More importantly, by programming the precise investment thesis into the system prompts, the architecture can autonomously reject 90% of leads that don’t match the requirements. 

In the traditional system, your $250k/year associate spends 20 hours a week saying “no” to bad deals.

In the dark factory, the associate is removed. The partner skips the noise and focuses 100% of their energy on the top 1% of opportunities. The human moves from being a filter to being a closer.

Boosting the Liquidity via “Dark Settlement”
We see the biggest impact for secondaries ecosystem in the settlement layer. Private market liquidity is stifled by notorious legal friction. A simple secondary trade often takes weeks because it sits in a human lawyer’s queue, waiting for the review of the shareholder agreement, checking the ROFR clauses, etc.

In the age of “dark settlement,” this compliance layer runs without lights. Agents parse the legal docs and execute the settlement documentation deterministically, transforming secondary equity from a slow asset class into a high-velocity financial instrument comparable to public equities. 

It’s already here. JPMorgan’s Kinexys recently demonstrated near real-time settlement for private fund flows via automated, blockchain-based process, proving that the primary friction to liquidity is administrative, and not regulatory.

Swapping Headcount for Architecture
For firm owners, this changes the fundamental unit of leverage. In the past, scaling AUM meant scaling headcount. If you raised more capital, you hired more associates. 

Future agentic ‘insurgent’ firms will scale AUM by scaling compute, not payroll.

Future agentic “insurgent” firms (as Bain calls them) will scale AUM by scaling compute, not payroll. Today, these firms are building a competitive moat through:

• Moving beyond the “rule of 40” to target a combined 60% revenue growth and margin by using AI to cut operational costs.
• Requiring every portfolio CEO to submit P&L improvements driven by AI, rather than vague “efficiency” metrics.
• Using cluster-based deployment, which means solving a bottleneck once (e.g., automated KYC or code refactoring) and immediately deploying that digital agent across all portfolio companies.
• Moving from reactive reporting to predictive action, with agentic solutions like Edwin AI identifying market threats and opportunities before they manifest.

These shifts redefine the roles of the investment professionals. If the job implies processing information like writing memos and scheduling calls, the role becomes obsolete. If the job is defined by process architecture, the role shifts from doing the diligence to designing the diligence loop.

Even at the highest levels of TradFi, this shift is undeniable. BlackRock’s Larry Fink recently signaled a fundamental ‘decoupling’ of AUM growth from headcount. The firm swiftly proved it: as of early 2026, BlackRock’s AUM surged to a record $14 trillion amid three waves of job cuts in 2025-2026, each pruning ~1% of BlackRock’s workforce.

Conclusion
The firms that win in the next decade will be the ones that identify where biological limits are slowing down capital velocity and automate those bottlenecks via a dark factory approach. At SecondLane, we are turning the lights out on the busywork so we can turn the lights on for the alpha.

Nick Cote, CEO & Co-Founder, SecondLane