Executive Summary

• AI allows our investment process to be deeper, faster, and more consistent.
• Our team can shift resources away from mechanical tasks and toward higher-value analytical and judgement-driven work, aiming to improve our qualitative research, meeting preparation, thematic analysis, and performance attribution.
• While we are adding these capabilities into existing workflows iteratively, decisions still require meaningful human judgement, interpretation, and oversight.
• Next steps include exploring the use of AI for forecasting manager outcomes.

How AI Enhances our Investment Process

Over the past two years TIFF has integrated AI into our investment process, and it has already proven to be a worthwhile investment. At a high level, AI enables us to move faster without cutting corners, analyze more information without losing rigor, and apply a more consistent analytical framework across managers, strategies, and time periods. The result is a deeper, more scalable research process that supports better-informed decisions.

More Structured and Auditable Qualitative Analysis

One of AI’s most powerful contributions has been transforming how we handle qualitative information. Investment research is inherently text-heavy, encompassing manager letters, pitch decks, due diligence questionnaires, meeting notes, and internal memos. Historically, synthesizing this information has been time-consuming.

• Dynamic Summary of a Manager’s Entire Investment History with TIFF: Prior to implementing our AI system, an analyst seeking to understand the investment thesis for a manager would have needed to navigate our research management system, locate the investment memo, review subsequent update notes, synthesize the information, and then form a view—a process that could take hours. Today, we can query a large language model that has already incorporated our entire research library to summarize a manager’s investment history with TIFF in under a minute. Importantly, all outputs are fully auditable and directly tied to source underlying documents.
• Ingestion of Incoming Manager Information: AI also allows us to systematically ingest, structure, and analyze incoming text through a process as simple as forwarding an email. Previously, tagging, uploading, and filing materials required significant manual effort. Now, we use built-in tools that automatically detect, summarize, and email quarterly letter summaries to the investment team.
• Consistent and Robust Manager Comparison Matrix: A particularly valuable capability is matrix-style analysis, which allows us to compare and contrast managers using multiple documents or time horizons. This approach enables us to identify common themes, points of divergence, recurring risks, and key differentiators far more quickly than traditional manual review. For example, we can construct a matrix that queries the most recent quarterly letters from all public markets managers and then directly interact with the results to understand areas of agreement and disagreement across our portfolio. What once required a week of effort can now be accomplished in under an hour, freeing time for deeper interpretation and discussion.
• Consistent “First-Pass” Review for New Managers: When evaluating new managers, we use a matrix to compare each underwriting criterion at the sub-strategy level (e.g., applying our pre-defined manager-ranking criteria, conducting a scoring exercise, and supporting an answer in less than 100 words). Where AI’s perspective differs from our own, it suggests areas for further investigation—augmenting, rather than replacing, human judgment.1

Improved and Consistent Meeting Preparation

AI has also become central to how we prepare for manager meetings. Using structured, in-depth research workflows, we now produce concise overviews of a manager’s strategy, history, public reputation, strengths, weaknesses, and potential areas of concern prior to both initial and follow-up meetings. Importantly, AI also helps identify gaps, inconsistencies, or areas where information is sparse—often among the more productive areas to explore in conversation. Where risks or uncertainties warrant external validation, we use AI to help structure diligence plans, including suggested data requests and lines of inquiry for in-person meetings.2

Faster, More Targeted, and Deeper Thematic and Industry Research

AI materially improves our ability to capitalize on new strategies, industries, and market themes. When exploring unfamiliar areas, speed matters—but so does breadth. AI allows us to quickly synthesize large bodies of third-party research, expert commentary, and historical context to build a foundational understanding before engaging in deeper primary diligence.

This capability is particularly valuable in early-stage thematic work, where the goal is not precision forecasting, but rather understanding the landscape: how a strategy works, what risks tend to matter, where returns come from, and how different approaches compare.3

Clearer Performance Attribution and Risk Understanding

On the quantitative side, AI-enhanced tools improve how we analyze portfolio performance and risk exposures. Traditional multi-factor regressions remain useful, but machine learning techniques allow us to go further by identifying which factors truly matter statistically and offer better ways of isolating idiosyncratic returns (skill) from systematic returns. For example, we use techniques such as lasso regressions to determine which among the dozens of equity style factors are most closely related to a manager’s results. This leads to clearer attribution and more informative conversations about portfolio construction, diversification, and risk management.4

Next Step, Forecasting

While our systematic managers are using AI methods to directly forecast asset prices, TIFF is not currently using AI methods to forecast manager-level outcomes. Over the next year, we aim to examine this area further, as we believe that layering our unstructured text data with our structured numerical data could enhance our forecasting capabilities and support better decision-making.

Conclusion

AI is enhancing TIFF’s investment process by enabling a deeper, faster, and more consistent approach to research while reinforcing the central role of human judgment. By shifting time away from mechanical tasks and toward thinking, discussion, and decision-making, AI helps our team operate more productively. Assessing incentives, motivation, alignment, strategy coherence, portfolio fit and sizing remain fundamentally human responsibilities. AI simply allows us to bring more informed data, improved consistency, and a broader perspective to those judgments. As AI technology capabilities continue to level up, we see opportunities to further strengthen our research process.

The materials are being provided for informational purposes only and constitute neither an offer to sell nor a solicitation of an offer to buy securities. These materials also do not constitute an offer or advertisement of TIFF’s investment advisory services or investment, legal or tax advice. Opinions expressed herein are those of TIFF and are not a recommendation to buy or sell any securities.

There can be no guarantee that the use of Artificial Intelligence (“AI”) and Large Language Models (“LLMs”) will lead to investor returns. AI tools and LLMs may contain errors or inaccuracies and should not be relied upon as a substitute for professional advice. Any references to AI tools and LLMs use and advantages should be construed accordingly.

These materials may contain forward-looking statements relating to future events. In some cases, you can identify forward-looking statements by terminology such as “may,” “will,” “should,” “expect,” “plan,” “intend,” “anticipate,” “believe,” “estimate,” “predict,” “potential,” or “continue,” the negative of such terms or other comparable terminology. Although TIFF believes the expectations reflected in the forward-looking statements are reasonable, future results cannot be guaranteed.