Before We Fund AI, We Need to Fund Readiness

 A distinct energy fills the room whenever AI enters philanthropy discussions.

That energy is part of 99% of my AI culture work.

It’s a mix of curiosity, pressure, ambition, and fear of being left behind.

Sometimes it’s about a board that asks whether the organization has an AI strategy. Or a funder who says they are interested in supporting innovation but isn't sure how. Or, my favorite – a leadership team hears that peer institutions are already experimenting with AI for fundraising, communications, evaluation, and note-taking, and wants to ramp up their efficiency with AI.

And before I know it, the conversation has moved on to pilots, licenses, tools, working groups, or worse, even more working sub-groups.

And I understand the tension beneath that energy. I really do – both as a human and as a practitioner interested in creating more dream-time before execution talks.

I mean, for many of our organizations, this is no longer a distant issue.

·        Staff are already testing tools quietly (“did she use ChatGPT to create that deck?”.

·        Vendors have always been making promises (“with only our local organic LLM can you change the fundraising future”).

·        Leaders are already being asked by boards (and perhaps donors) what their organization is doing in response.

The question is not whether AI is present in the social sector—it is. The real challenge, and my focus today, is whether the support structures around AI match the significance and urgency of its impact.

To assess true commitment to AI in philanthropy, we must track funding. Ultimately, money reveals priorities and exposes whether institutions are equipping organizations for genuine transformation or just superficial engagement.

Money tells us what the funder believes matters. What funders believe can wait. What labor counts as real infrastructure, and what labor is expected to happen on the side, for free, on top of everything else. Funding patterns reveal whether AI is being treated as a shiny add-on, a short-term experiment, or what it actually is – a structural shift that requires deeper institutional readiness.

And increasingly, I think the way we are funding AI in this sector is one of the clearest tests of what philanthropy, boards, and nonprofit leaders believe organizations deserve as they navigate this transition.

This is what I want to explore with you today – the cost of funding only the visible, quick, and short-term outcomes-enabled parts.

Too much funding imagination fixates on what is easiest to see: buying a tool, launching a pilot, hiring a short-term consultant, delivering a staff session, or producing a strategy memo.

These things are photographed well in reports. They look like movements to funders, and sometimes they are.

What often remains underfunded is the slow, quiet work essential for readiness: internal literacy, governance, implementation support, policy interpretation, data review, staff dialogue, cross-functional trust, and judgment development.

None of that fits neatly into a 12-month grant cycle. None of it looks like innovation in the market sense. And yet, this is exactly the work that determines whether the tool a nonprofit just bought becomes useful infrastructure or quiet harm.

AI cannot be named as a tool question. It is a governance question – especially if you and I choose to prioritize equity in all of that AI work.

Especially when AI does not sit at the edge of organizational life. It reaches into the middle of it. It (AI) touches questions like - how decisions get made, how much trust is placed in automated systems, what data is allowed to move through those systems, who is accountable when outputs are wrong, how staff is expected to exercise judgment, and what kinds of risks communities may be asked to absorb without ever being consulted.

Let’s see some examples of what exactly is happening when a nonprofit decides to use AI:

a. A community health nonprofit piloting an AI triage assistant is not running a productivity experiment. Through their AI use, they are making decisions about who gets seen sooner and who waits.

b. A workforce development program using AI to match clients to opportunities  - that organization is making decisions that shape people’s economic futures – often for people who have already been written off by other systems.

c. A youth-serving organization building an AI chatbot for crisis navigation is operating in a space where a poorly written policy is not beyond inconvenience; it is harm.

These are exactly the kinds of organizations most drawn to free or fast or subsidized AI support – and the ones with the least margin for error.

This is why we must focus on what it takes to achieve true readiness for AI in nonprofits—and direct funding accordingly.

Let us start with - what does (AI) readiness actually look like (and what funders keep skipping)?

Just to be clear, when I say nonprofits need readiness funding, I do not mean another workshop. I mean the conditions under which responsible adoption can actually emerge.

Let me name what those readiness conditions look like, for that’s where you and I should seek dollars moving:

• Time to learn before you are expected to implement. One workshop rarely settles anything. One pilot does not create institutional clarity. Teams need time to ask better questions, compare use cases, notice where discomfort sits, and figure out which work should remain fully human. That kind of learning does not look flashy on a grant report. It is still foundational.

• Investments into real governance, not just enthusiastic implementation. A leadership team excited about AI is not the same as a leadership team prepared. Excitement does not answer who approves use cases, what data may be entered into external tools, who reviews outputs, or what happens when a staff member leans too hard on AI in a high-trust context. These are governance questions, and they deserve to be treated as central, not as paperwork.

• Building capacity; not just access. Our sector often talks about access as if it is the main barrier. Access to tools, access to discounts, access to a webinar, and access to a pilot. But access is not capacity. An organization can have access to five tools and still have no meaningful internal ability to evaluate them, question them, adapt them, or build clear practices around them.

• Building room to pause, revise, and say no. Responsible AI use does not only depend on knowing how to adopt. It depends on recognizing when adoption is premature, misaligned, or simply not worth the risk. When every funding incentive rewards visible movement, organizations have very little room to exercise caution without looking behind.

Those conditions need to be part of AI readiness because that’s what ensures “human in the loop”

And I think boards deserve a separate mention in this conversation, because boards are often the first place AI urgency enters an organization (should you and I write another whole piece on this?)

Board members are hearing about AI through a language of speed, competitive pressure, and inevitability. They are told this is the next “major shift” and that strong organizations need to be ready. There is some truth in that. But readiness is not the same as adoption.

A board’s job is not to simply ask whether the organization uses AI, but to ensure the right conditions for good AI decisions—whether leadership understands the implications for different departments, and whether staff are being asked to take on new responsibilities that are properly resourced.

Those are different questions, and they lead to very different conversations.

I suppose more reasons for us to follow the dollars and understand equitable funding in this conversation.

So, what can good AI funding look like?

When I think about what serious AI funding looks like, I am not mainly thinking about more tool grants.

I am thinking about a fuller ecosystem of support – one where foundations treat AI readiness as strategic infrastructure rather than a trendy side investment, where boards understand that governance is not administrative drag, and where leaders are given enough room to build internal judgment rather than perform innovation.

A stronger funding response, to me, includes several layers at once:

• Assessment before adoption. So organizations can understand where AI may fit, where it does not, and what internal gaps need attention first – before a license gets signed.

• Multi-month learning support. So teams can build literacy, practice discernment, and make decisions together rather than absorb fragmented information in isolation.

• Policy and governance support as living practice. Not one-off documents that get filed away, but practices that evolve alongside real use.

• Implementation guidance tied to context. Especially for nonprofits working in high-trust, high-sensitivity, or community-facing environments where the stakes do not allow for generic playbooks.

• Peer learning and shared infrastructure. So organizations are not each left to reinvent the wheel privately, expensively, and unevenly.

Now we have to understand that some of the most important outcomes of AI funding may not look dramatic in the first year.

An organization might gain stronger language, clearer policies, slower deployment, a better sense of where AI belongs, a team more capable of challenging vendors, leadership confident to decline certain uses, and a board that links technological and ethical readiness.

None of that will be quantified in the traditional capitalistic sense.

But it may be exactly what allows an organization to remain in the right relationship with its mission and its community.

***************************************

If funders, leaders, and boards say they want ethical, community-aware, trustworthy uses of AI, then they have to fund the conditions from which those uses can actually emerge.

It is not enough to fund the exciting part and leave organizations to absorb the rest through overstretched staff, vague good intentions, or institutional hope.

AI will ask nonprofits to make decisions about trust, judgment, privacy, labor, consent, and accountability under rapidly changing conditions. Those decisions deserve more than enthusiasm. They demand attention for infrastructure.

How we resource this moment will shape far more than who gets to experiment with AI.

It will shape whether organizations can build real judgment. It will shape whether those organizations’ communities are protected by stronger governance. Most importantly (and collectively), it will shape our sector (that we care for so deeply) if it is pushed to move before it is adequately held.

I believe for you, me, and those of us who care about justice, trust, and joy, that is no longer a side question.

(And I refuse to pretend that funding only the visible parts of AI is enough.)

***************************************

*** So, what do I want from you today (my readers)?

Share with us: What is the most underfunded part of AI readiness in your organization right now – the part that no one is writing checks for, but everyone is quietly being asked to do?