The Cost of Progress Isn’t Technical — It’s Relational
Philanthropy doesn’t break because it’s slow. It breaks because we mistake activity for clarity. Because we implement tools without alignment, scale systems without sense-making.
Now AI promises to make us faster. But speed without strategy just gets us to the wrong answer faster.
The hard part of our work was never the “how.” The question AI can’t answer is: why.
And in our field, that’s everything.
From Legal Pads to Pattern Recognition
When I stepped into philanthropy in the early 2000s, my predecessor held the title Secretary to the Board. Her job was to sit in grantmaking meetings, yellow legal pad in hand, recording decisions and filing them in manila folders.
When I arrived, those notes had moved into GIFTS, a grants management system. The desk had a computer, not a legal pad. But the real shift wasn’t technological.
The role had become something else—even if the title didn’t say so.
The job wasn’t just documenting decisions anymore. It was translating them. Connecting dots across cycles. Seeing the silences. Interpreting what didn’t get said in the room.
It was the early work of sense-making.
The tools changed. The title didn’t. And critically—the organizational understanding of the role’s strategic value didn’t either.
Two decades later, I see foundations making the same mistake with AI. We’re updating interfaces without updating our expectations.
AI Can Hold the Whole Essay Question
I explain it this way:
Remember those high school history essay prompts?
“Compare the views of Jefferson, Madison, and Hamilton on the Federalist Papers. Analyze their perspectives on centralized government. Then explain which view you find most compelling—and why.”
AI can now hold that question. It can retrieve the facts, compare the arguments, summarize the context.
But it can’t answer the why.
That’s our job.
In practice, this means:
When your AI tool flags a grantee report as “high risk” based on sentiment analysis, someone still has to ask: High risk to whom? According to whose definition of success? And what context is the algorithm missing?
When your system auto-generates a dashboard showing “program efficiency,” someone needs to question: Are we measuring what matters, or what’s measurable?
The technology can hold the data. Leaders hold the judgment.
In philanthropy, “why” is where strategy lives:
Why we’re willing to accept lower “efficiency scores” to fund grassroots organizing that builds long-term power
Why we designed reporting systems that slow down approvals when equity is at stake—because some decisions should take longer
Why our grants management infrastructure treats human judgment as a feature, not a bug
AI can surface the context. But it can’t explain what we care about. It can hold complexity—but only we can choose what matters.
Relational Infrastructure Starts with the Job
In Reimagining Philanthropy: The Power and Responsibility of Grants Managers, I wrote:
“Grants managers are not just administrators of process. They are stewards of values.”
Yet many AI initiatives in philanthropy are launched through IT, strategy, or executive functions—without rethinking the frontline roles that actually carry our intent through workflows.
That gap matters.
The Deloitte 2026 State of AI Report found that 84% of organizations have not meaningfully redesigned any jobs in response to AI.
Read that again: we’re investing millions in new technology while leaving 20th-century job architectures untouched.
It’s the equivalent of installing high-speed internet but keeping the dial-up workflows.
The bottleneck isn’t technical. It’s organizational. We are layering technology onto roles that were never built to hold this kind of responsibility—and wondering why we feel unprepared.
As I wrote in It’s Time to Rewrite the Job Description, job design isn’t administrative. It’s strategic infrastructure.
If we want AI to support discernment, care, and equity—we have to build roles that do, too.
The Real Work Is Meaning-Making
In Collective Learning in Philanthropy, the authors write:
“Learning is most effective when it is social. Technology can support it, but cannot substitute for the relationships that make learning meaningful.”
That’s the invitation AI gives us—if we’re paying attention.
If we let it, AI can reduce administrative burden, detect blind spots, and surface patterns across time. But none of that matters unless we have the human capacity to ask: What does this mean? And what will we do with it?
That second question is harder than it sounds.
Because information creates obligation. Once you know something, you can’t unknow it. And if you choose to do nothing with what you’ve learned, that becomes a choice too—one with consequences.
When Information Creates Obligation
I learned this lesson in a deeply personal context.
During pregnancy, I was offered a series of genetic tests that could reveal potential developmental delays or diseases. I declined most of them—not because the information wasn’t valuable, but because I realized: if I had that knowledge, I would feel compelled to make a decision I wasn’t prepared to make.
The information would have created an obligation I wasn’t willing to carry.
AI in philanthropy will do the same thing. It will show us patterns we’ve been able to ignore. It will surface questions we’ve been able to avoid.
Your new AI tool might reveal that you’ve systematically underfunded BIPOC-led organizations for a decade. What will you do with that information?
It might show that your “equity focus area” has resulted in zero grassroots grantees receiving general operating support. What will you do with that information?
It might flag that your board’s definition of “program success” correlates perfectly with proximity to power and wealth. What will you do with that information?
Information without infrastructure for action becomes a burden, not a tool. And worse—it can become evidence of willful negligence if we see the pattern and choose to do nothing.
This is why the technical question (”Can AI show us this?”) is so much easier than the organizational question (”Are we prepared to act on what it shows us?”).
Before we deploy tools that surface inequity at scale, we need to ask: Do we have the governance structures, the strategic courage, and the relational trust to respond to what we’ll learn?
What a Different Approach Could Look Like
Imagine this instead:
A foundation wants to adopt an AI tool that summarizes grantee reports. But before rollout, the grants team:
Brings in grantees to test if their voice is being honored, not flattened
Asks program staff whether the summaries sharpen or blur strategy
Rewrites internal roles to include reflection, not just automation oversight
Aligns with the board on what insights are needed—and who gets to define them
Creates clear protocols for what happens when the AI surfaces uncomfortable truths
The result isn’t just a new tool. It’s organizational intelligence—the kind that compounds over time because you’ve built the infrastructure to learn, not just automate.
That’s the difference between adoption and transformation. Between buying technology and building capacity. None of this happens by accident. It requires deliberate governance design.
This is the work I do with foundation leadership teams through AI Readiness and Governance engagements — assessing organizational capacity, aligning strategy with values, and building the infrastructure to act on what AI will reveal. If you’re exploring what readiness actually requires,
Technology Doesn’t Transform Institutions. People Do.
As Alberto Romero writes in The Algorithmic Bridge, we’re at an inflection point where the skills we’ve spent our lives developing may no longer be sufficient on their own.
The question isn’t whether AI will change our work—it’s whether we’ll build the wisdom to guide that change.
Philanthropy doesn’t need to build the next large language model. But we do need to build a model for how technology can serve equity, context, and care—not just scale.
We need relational infrastructure before we need automation.
We need clear job descriptions before we need prompt engineers.
We need protocols for difficult conversations before we need dashboards.
We need space for the why.
The Foundations Getting This Right
The foundations getting this right aren’t starting with vendor selection or pilot programs. They’re starting with harder questions:
What patterns will we ask the AI to surface—and what will we do when we don’t like the answer?
Because here’s what actually happens:
You implement an AI tool to “improve grantmaking efficiency.” Six months in, a program officer runs a simple query: “Show me our funded organizations by geography, budget size, and leadership demographics over the past five years.”
The dashboard loads. And there it is: a visual map of your stated equity priorities versus your actual funding distribution. The gap is undeniable.
Or your AI-enhanced grants management system starts auto-tagging applications. You notice it’s flagging nearly every grassroots organization as “high capacity risk” while green-lighting established intermediaries—because you trained it on your historical approval patterns.
Or you ask your new tool to analyze grantee feedback sentiment over time. It shows that organizations led by people of color consistently report feeling “grateful but unheard” while predominantly white-led organizations report feeling “valued as partners.” The AI didn’t editorialize. It just counted the words.
These scenarios aren’t hypothetical. They’re inevitable—if you’re actually using AI to see patterns instead of just automating approvals.
So before you deploy the tool, ask:
When the data shows our equity commitments haven’t translated to funding patterns, are we prepared to change our strategy?
When the AI surfaces that our “capacity” criteria have systematically disadvantaged certain communities, can we rewrite our definitions?
When the dashboard makes our stated values measurable against our actual practices, do we have the governance structures to act on the gap?
These aren’t questions for your IT department. They’re questions for your leadership team.
About This Series
Article 1 → Moral obligation
This article, AI Can Hold the Question, is the first of a seven-part series exploring how foundations can build governance infrastructure before AI compounds existing asymmetries.
The Path Forward
AI isn’t coming—it’s here. And it will reshape philanthropy whether we’re intentional about it or not.
The question isn’t whether to adopt these tools. It’s whether we’ll build the relational infrastructure, strategic clarity, and institutional courage to use them well.
Because AI can hold the entire essay question.
But only we can answer why it matters—and what we’ll do with that answer.
In philanthropy, that’s not a technical challenge. It’s a leadership imperative.
Before your foundation adopts AI tools, you need infrastructure for what those tools will reveal. I work with boards and leadership teams through AI Readiness & Governance engagements and fractional CIO partnerships to build that capacity.
If you’re navigating this moment — and ready to lead, not react — complete this short intake form and let’s explore how I can support you.
Sources & Further Reading
Collective Learning in Philanthropy – The Foundation Review
It’s Time to Rewrite the Job Description – Proximate Press
Reimagining Philanthropy: The Power and Responsibility of Grants Managers – Proximate Press
The State of AI in 2026 – Deloitte
You Spent Your Whole Life Getting Good at the Wrong Things – The Algorithmic Bridge
About the Author
Rachel Kimber is a global philanthropy advisor and fractional CIO helping foundations align technology adoption with institutional values and mission. With more than 20 years leading grants operations and systems strategy across 70+ countries, Rachel helps leadership teams navigate AI readiness, data governance, and organizational change with clarity and confidence.
Rachel serves as faculty for Fund for the City of New York AI Sprints, New York Community Trust Leadership Fellows, Good Academy AI and She Leads AI, and was an ELI Mentor with the Technology Association of Grantmakers. She co-created the AI Readiness Framework and will be speaking at PEAK Grantmaking 2026.
Curious what AI readiness really looks like for your foundation? I work with boards, leadership teams, and grants management staff to turn questions into strategy. → Let’s talk.