TIFF's Tom Duffy on AI, Fund Manager Specialization

Artificial intelligence is speeding up market research and making data more accessible, shifting competitive advantage from information gathering to value assessment. According to Tom Duffy, director of Private Markets at TIFF Investment Management, this raises the bar for fund managers, who must now focus on specific sectors, evaluate new technologies carefully, and exercise patience with investments.

As automated analysis becomes widespread, broad sector coverage can turn into a liability. Competitors can access similar insights simultaneously. Duffy states that in a world where AI has commoditized data, focus and expertise in a specific market area can differentiate a good manager.

To identify true expertise, investors must evaluate managers against specific criteria. Managers need a clear strategy explaining why their niche offers a structural advantage. Sourcing deals through unique, proprietary networks provides better diversification than backing generalists tracking the same assets. While machines accelerate data processing, they cannot replicate human judgment for assessing founder capabilities or market shifts.

Fundamental Underwriting Over Hype

Applying specialized expertise to the competitive AI market requires frameworks that separate viable businesses from hype. Investors must identify the core customer need a company solves, rather than just its sophisticated models. They should verify product stickiness by ensuring enterprise customers see clear benefits and maintain usage long after trials. Assessing operational scalability is key to see if a business can expand rapidly without breaking under manual service demands.

This approach helps investors avoid funding unproven ventures simply because they use new software. Duffy notes that while automation accelerates research, the best general partners still must make qualitative calls on people and markets that no tool can make.

Handling Liquidity Pressures

The same discipline needed to underwrite complex technology is vital for managing delayed cash returns in mature portfolios. Slow distributions can indicate quality assets are taking longer to mature, not necessarily failing.

To avoid destroying long-term value through premature exits, investment boards should adopt patient portfolio strategies. Spreading capital across multiple years mitigates reliance on a single economic cycle. Evaluating health metrics like rising valuations and expanding profit margins gives a more accurate picture than immediate cash flow. Using specialized secondary markets can generate liquidity without forcing high-growth companies into public domains prematurely.

Duffy warns against manufacturing short-term liquidity. Forcing exits into a closed IPO window or a soft M&A environment could mean selling great companies at the wrong time.

Winning in a data-commoditized environment requires both specialized investment expertise and patience during exits. The source article, summarizing an episode of the VC10X video series, also presents counterpoints from other research. McKinsey's 2026 research suggests large language models can miss key deal data, indicating winning firms will combine specialist judgment with proprietary research. PitchBook-NVCA data shows most capital flows into AI companies and large financings, creating crowded trades. McKinsey and Bain reports note limited partners prioritize distributions and that secondary markets remain a small part of private equity, meaning delayed exits cannot fully offset fundraising pressure.