AI and Endowment Investing: What Boards Need To Know

Artificial intelligence has moved from the periphery of institutional investing to the center of the conversation. Geopolitical developments and AI advances are rippling across financial markets simultaneously — reshaping risks, creating new opportunities, and accelerating capital allocation decisions for institutional investors in ways that are difficult to separate from one another.

Pension funds, sovereign wealth funds, and major endowments are actively deploying AI tools across portfolio construction, risk management, and operational workflows — and the pace of adoption is accelerating.¹

For boards overseeing endowment funds, the question is no longer whether AI will affect your investment strategy. It’s whether your governance framework is ready to manage it thoughtfully.

This isn’t a technology story. It’s a fiduciary story.

The AI opportunity is real — and so are the risks

The performance implications of AI in institutional investing are significant. Norway’s Norges Bank Investment Management reported that AI-related cost-reduction initiatives helped cut fund trading costs by 30% between 2023 and 2025.²

Generative and agentic AI tools are being deployed across the investment lifecycle — from generating capital market assumptions to evaluating alternative investment managers — and 58% of fund managers surveyed by Alternative Investment Management Association expect increased AI use in their investment processes over the next 12 months, up from 20% in 2023.³

At the same time, concentration risk has emerged as a legitimate concern. Several major university endowments are actively studying their exposure to AI-related equities, weighing the substantial gains captured over recent years against the risks of overvaluation and portfolio concentration.⁴ That tension — between capturing structural opportunity and managing the risks that come with it — sits squarely within the investment committee’s responsibility.

AI also introduces new governance complexity. When algorithms influence portfolio decisions, the investment committee bears responsibility for understanding how those decisions are made — even if the committee members aren’t AI engineers. As one recent governance framework put it, fiduciaries cannot oversee risks they do not understand.⁵ Delegating portfolio decisions to opaque models without clear accountability structures represents a meaningful governance vulnerability.

Three integrated strategies for boards to consider

Navigating AI in endowment management isn’t a single-discipline challenge. It requires coordinating your investment management, tax planning, and estate and philanthropic planning — each of which is affected by, and can benefit from, thoughtful AI integration.

Investment management: AI-enhanced portfolio construction

AI tools are increasingly capable of processing large datasets to identify portfolio construction opportunities across asset classes — including alternatives such as private equity, private credit, and real assets — that are central to sophisticated endowment strategies. Agentic AI systems can now coordinate multiple specialized agents to generate capital market assumptions, run competing portfolio construction methods, and evaluate outputs against your investment policy statement.⁶

For boards, this capability carries both opportunity and obligation. The opportunity lies in more rigorous, data-informed analysis of complex investment opportunities — including concentrated stock positions, private market exposures, and scenario modeling across market cycles. The obligation lies in governance: Your investment policy statement should explicitly address how AI-influenced strategies are evaluated, what human oversight is required, and how model-driven recommendations are reviewed by committee members.

As AI reshapes portfolio management, many boards are also reconsidering whether their current endowment management model — whether committee-led, internal CIO, consultant-supported, or outsourced CIO — is appropriately structured to oversee AI- influenced strategies with rigor and accountability.

Tax planning: Maximizing after-tax endowment returns

Tax planning is one of the most consequential areas where AI tools can add value for foundations and nonprofit endowments. They can address questions about managing unrelated business taxable income (UBTI) exposure generated by alternative investments, monitoring excise tax liability on net investment income, and helping ensure that portfolio activity doesn’t jeopardize the organization’s tax-exempt status.

AI can help boards monitor tax exposure across a complex portfolio by identifying potential UBTI-generating investments, modeling the impact of investment decisions, and flagging tax consequences before they become filing-time surprises. Quantifying that exposure early allows boards to make more informed allocation decisions and avoid unexpected tax burdens that erode returns.

AI tools are also well-suited for modeling the relationship between spending policy and after-tax returns — helping boards meet mandatory distribution thresholds, including the 5% minimum distribution requirement for private foundations, while preserving long-term purchasing power. For organizations subject to the excise tax on net investment income, scenario analysis can support efforts to reduce that burden across investment structures. For boards managing investments across related entities, such as private foundations, donor-advised funds, and charitable remainder trusts, AI can help align tax strategies across the entire charitable portfolio, reducing unnecessary tax drag and allowing more capital to support your mission.

Where taxable accounts exist within the broader structure, such as separately managed accounts tied to a private foundation, AI-assisted strategies can seek to improve after-tax outcomes on those specific assets.

It’s worth noting that AI-assisted tax strategies must be reviewed carefully for compliance with the fund’s tax-exempt purpose and any applicable regulatory frameworks. Technology amplifies decisions; it doesn’t replace the judgment required to make them well.

Estate and philanthropic planning: Smarter structures for long-term impact

For boards managing endowments tied to broader family or institutional legacies, AI tools are increasingly being used to model the long-term impact of different philanthropic structures. Comparing the effectiveness of a donor-advised fund versus a private foundation versus a charitable remainder trust across different return assumptions, distribution policies, and tax scenarios is precisely the kind of multivariable analysis where AI adds meaningful analytical depth.

AI-enhanced modeling also supports conversations about multigenerational wealth transfer alongside philanthropic goals. Sophisticated families and institutions managing significant endowments often face the challenge of aligning investment objectives with a stated mission — ensuring that the portfolio’s risk profile, liquidity needs, and spending policy support both long-term growth and the organization’s charitable purpose.

Boards can use AI tools to run stress tests on spending policies, evaluate the resilience of the endowment across economic scenarios, and model the impact of major distribution decisions before committing capital. This kind of forward-looking analysis strengthens board decision-making and supports better documentation of fiduciary reasoning — important both for governance purposes and for demonstrating alignment with the fund’s mission.

What your board should do now

The Intentional Endowments Network launched its Responsible Tech & AI Initiative in 2025 specifically because AI and responsible technology governance had become top-of-agenda items for investment committees and boards.⁶ The message from institutional practitioners is consistent: Boards that engage proactively with AI governance — rather than waiting for it to create a problem — are better positioned to capture the benefits while managing the risks.

Several practical steps can help your board get ahead of this:

• Review your investment policy statement for AI relevance. If your statement doesn’t address AI-influenced strategies, model oversight expectations, or data privacy requirements, consider whether it should. These don’t need to be prescriptive technology mandates, but acknowledging the category is a meaningful governance step.
• Establish clear accountability for AI oversight. Investment committees don’t need to become AI experts. They do need to know who is responsible for evaluating AI-influenced recommendations, what human review is required, and how model failures or anomalies are escalated.
• Evaluate your investment management model. As AI tools become more embedded in portfolio management, the question of whether your endowment is appropriately resourced to evaluate and oversee those tools deserves direct consideration.
• Coordinate across disciplines. The most durable approach to AI integration connects investment management decisions with tax strategy and philanthropic planning so that technology-enabled insights serve the whole financial picture, not just one part of it.

AI is actively reshaping endowment portfolio construction, tax efficiency, and philanthropic planning — and investment boards need to understand its implications today.

FAQS

What is agentic AI and how is it being used in endowment management?

Agentic AI refers to systems in which multiple specialized AI agents coordinate to complete complex, multistep tasks — such as generating capital market assumptions, constructing portfolios using competing methods, and evaluating outputs against a stated investment policy. In endowment management, agentic AI seeks to bring more rigorous, data-informed analysis to investment decision-making. It’s an emerging area,
and its adoption varies widely across institutions. The investment policy statement governs how these systems operate, which is why updating its language to reflect AI-influenced strategies is a meaningful governance step for any board.

How does AI affect the fiduciary duty of an investment committee?

AI tools don’t change the fundamental nature of fiduciary duty — they change its complexity. Investment committees remain responsible for the duty of care, the duty of loyalty, and the duty of mission, regardless of whether a recommendation comes from an internal analyst or AI. Fiduciaries are expected to understand how AI affects markets, managers, and operational risk well enough to provide meaningful
oversight. Delegating decisions to opaque algorithms without governance structures, such as clear accountability, documentation, and human review, is a governance vulnerability rather than a solution to one.

Should our board update our investment policy statement to address AI?

Most investment policy statements were written before AI became a meaningful factor in institutional investing. If yours doesn’t address AI influenced strategies, model oversight expectations, or data privacy considerations for investment managers, a review is worth
prioritizing. The update doesn’t need to be technically prescriptive. Instead, acknowledging the landscape and establishing accountability for AI oversight within your existing governance structure is a practical and meaningful starting point.

Is AI-assisted tax-loss harvesting appropriate for endowment-adjacent accounts?

For taxable accounts connected to endowment structures — including accounts in donor-advised funds or investment portfolios tied to private foundations — AI-assisted tax-loss harvesting can be a useful tool for seeking to improve after-tax
returns. The key is ensuring that any AI-driven strategy aligns with the fund’s tax-exempt purpose and is reviewed by qualified professionals. Automated tools amplify decisions, and they require skilled oversight to operate responsibly.

Who should be responsible for AI oversight in an investment committee?

Accountability for AI oversight should be explicit and documented — not assumed. The investment committee retains ultimate fiduciary responsibility, but day-to-day oversight of AI-influenced processes should be assigned to a named individual: the CIO, an outsourced chief investment officer, or a designated committee member. Many institutions are also adding AI governance language to committee charters that defines escalation procedures and minimum review requirements for model-driven recommendations.

Where can our board find guidance on responsible AI integration for institutional investors?

Several authoritative resources have emerged specifically to help institutional investors navigate AI governance, including The Intentional Endowments Network. It launched the Responsible Tech & AI Initiative for endowments and foundations in 2025. A financial advisor with endowment experience can help translate frameworks into practical policies for your institution.

What’s the difference between AI as an investment theme and AI as an investment management tool?

These are two distinct considerations that are easy to conflate. AI as an investment theme refers to portfolio exposure to companies building or benefiting from AI — a structural opportunity that many endowments are weighing alongside concentration risk. AI as an investment management tool refers to using AI to manage your portfolio more effectively — for analysis, portfolio construction, tax optimization, and reporting. Both deserve board attention, but they require different governance conversations: one about allocation strategy, the other about operational oversight.

How does AI-influenced investing differ from a traditional outsourced CIO model?

A traditional outsourced CIO provides human experts to manage portfolio strategy and implementation on behalf of the board. AI-enhanced investment management layers analytical tools into that process, and it is now being incorporated by many OCIO providers to improve analysis, reduce operational costs, and identify opportunities across larger datasets. The board’s role remains consistent in both cases: to set investment objectives, approve the IPS, and oversee the management model with appropriate rigor. The difference is that AI-informed models require governance language that specifically addresses model oversight and accountability.