In our previous post, we shared findings from Let’s Keep Learning, a research initiative led by Engage R+D and ORS Impact, with support from the Barr Foundation and others. In that post, we suggested, based on what we learned, that the challenge facing evaluation and learning in philanthropy is not primarily a value proposition problem, but a conditional, structural, and capacity problem.
Across interviews with a range of stakeholders, we heard real belief in the value of learning and evaluation. Leaders saw it as essential to understanding impact, improving strategy, and maintaining some accountability in a sector where formal accountability mechanisms are limited. But belief in the value of learning in theory is not the same as having the conditions, structures, habits, or capacity to use it well in practice.
To better understand how we might resolve this disconnect, we hosted a workshop at GEO’s 2026 National Conference with more than 100 foundation and evaluation leaders. We shared what we had been hearing through Let’s Keep Learning: foundation evaluation and learning functions appear relatively stable internally, while external capacity is under greater strain, and internal foundation staff are in some cases being asked to play broader roles without commensurate resources or clarity about scope.
We asked participants to help us think beyond diagnosis. Given current conditions, what should foundations do differently? What does learning need to become now? Where are there bright spots and examples of innovative practice that we could learn from? The ideas we heard fell into four broad categories.
Learning Needs to Move Closer to Strategic Decision-Making
Participants repeatedly named the need to orient strategy around questions, not just plans. They talked about using hypotheses, testing assumptions, embedding learning in goals and workplans, and making peace with “80 percent theses” that are directionally correct rather than waiting for perfect information. Each of these shifts, they said, greatly increases the usefulness and relevance of learning and evaluation work.
In other words, learning work is most relevant and powerful when it is part of how strategy develops and adapts, rather than something that arrives later to assess whether the original strategy worked.
Participants also named a power dimension. Several groups talked about giving more decision-making authority to people closer to the issues, creating feedback loops with grantees, and bringing more humility to how strategy is shared with the organizations expected to carry it out. Strategy, in this view, is not a finished product handed down from above. It is a set of commitments and hypotheses held lightly and with healthy skepticism that should be informed by people with direct knowledge of the work and revised as learning accumulates.
This insight also has implications for boards. Participants called for boards to see more of the ‘messy middle’ of learning: the midstream questions, tensions, adaptations, and incomplete evidence that shape real strategic work. If boards are expected to govern in complex times, they need visibility into the learning process itself, not only polished summative products which limit options for meaningful governance.
Learning Needs to Create More Value Beyond the Foundation
Too often, foundations decide what they want to know, collect data from grantees and communities, and only later ask how findings might be shared back. Participants pointed toward a different sequence: ask at the outset who the learning is for, what grantees or community partners would find useful, what information already exists, and what learning can be returned to the system rather than extracted from it.
In many foundations, evaluation questions are still organized around internal accountability, board reporting, or program strategy. Those purposes still matter. But in the current environment, where nonprofits and communities are under strain, learning efforts that only flow upward into foundations are insufficient, and also reinforce the sense that evaluation is something done to grantees rather than with or for them. Balancing inward-oriented reflection and learning questions with ones that generate learning for the field is essential in this context.
Participants also pointed toward more collective forms of learning. They shared collaborative structures that are working now: learning communities as strategy, pooled resources for shared tools, common applications for reporting and learning, and better coordination within and among funders. Increasing the value of learning may depend not only on whether a single foundation uses data well, but whether learning helps fields see patterns, coordinate action, and reduce duplicative burden.
The Field Needs to Protect and Replenish Learning Capacity
Participants surfaced real concern about the broader evaluation and learning ecosystem: A number of important field gains that have been made over the past 10-15 years are at risk. High-quality philanthropic learning depends on a broad infrastructure of consultants, facilitators, researchers, community-based knowledge holders, technical assistance providers, and field-building organizations. When that infrastructure weakens, foundations may still have internal staff, but the field loses perspective, methodological diversity, facilitation capacity, and trusted relationships with grantees and communities.
Notably, our research showed that firms led by people of color and small firms who bring valuable expertise into evaluation are under stress. In addition, pathway programs like Expanding the Bench and other field infrastructure like communities of practice are feeling the strain of the current moment. The ability to partner with an array of learning and evaluation practitioners with diverse racial, geographic, and subject matter perspectives is essential for foundations to gain meaningful insight.
So, how can we counter these stresses, preserve what we have built, and further strengthen the field? Participants suggested a range of solutions including underwriting internships and fellowships for small firms, supporting convenings and facilitation for evaluation-adjacent work, adding mentoring options at major conferences, and helping foundations see their role in sustaining the field. These are not simply professional development ideas — they are field preservation strategies.
Learning Needs a More Intentional Stance on AI
Participants did not reject AI. They saw real possibilities for supporting thinking, design, synthesis, and assumption-testing. But they were equally clear that AI should not replace human judgment. Several groups emphasized some version of “human in the loop — always.” Others urged foundations to ask whether a use of AI is strategic, not merely possible.
AI may speed up parts of learning work, but speed is not the same as insight. Used poorly, AI can decontextualize knowledge, reproduce bias, erode nuance, and further distance decision-makers from people and communities. Used well, it may help teams more quickly see and explore patterns, challenge assumptions, reduce administrative burden, and create more space for human sensemaking.
Participants called for common guidelines, spaces for foundations to compare policies, small innovation grants for nonprofits, and deeper conversation about ethics and appropriate use. The emphasis was not only on building technical skills. It was on preserving judgment, trust, relationship, and accountability as AI becomes more embedded in everyday work.
A New Phase of Philanthropic Learning and Evaluation
Taken together, these themes suggest that the field may be entering a new phase. The work is not disappearing, but it is being repositioned. That creates opportunity, but also risk.
The opportunity is that learning can become more central to how foundations navigate complexity, helping leaders make better decisions, adapt with humility, and stay accountable to the people and communities their work is meant to benefit.
The risk is that learning becomes narrower, faster, more internal, and less connected to field and community needs. Foundations may reduce burden in ways that also reduce insight, embrace speed in ways that weaken quality, or use AI in ways that increase efficiency but diminish judgment.
The challenge, then, is not simply to defend evaluation and learning as it exists now. The challenge is to redesign learning for the conditions we are actually in.
That means treating learning as a strategic practice, not a compliance function. It means building feedback loops that inform real decisions, making learning useful to grantees, communities, and fields, investing in the people and organizations that make high-quality learning possible, and using new tools, including AI, in ways that strengthen rather than substitute for human judgment.
Perhaps most importantly, it means foundations resisting the idea that learning is a luxury reserved for stable times. In moments of uncertainty, pressure, and constraint, learning is not less important. It is more necessary — but only if it is close enough to the work, the people, and the decisions to matter.
Clare Nolan is co-executive director of Engage R+D. Yvonne Belanger is director of Learning & Evaluation at the Barr Foundation. Sarah Stachowiak is CEO and owner of ORS Impact.
Let’s Keep Learning work was supported by the Barr Foundation, Blue Shield of California Foundation, The Colorado Health Foundation, Hopelab, The James Irvine Foundation, and the Walton Family Foundation.
