The AI-Reality Check: Principles for Responsible Innovation in Fundraising

So AI has absolutely dominated the nonprofit tech conversation lately, but usually in the context of very flashy, futuristic use cases.

But to ground us today, what is the most surprisingly effective yet completely unglamorous way that you've seen a nonprofit utilize AI recently that actually delivered real business value?

I think this is something that's been surprisingly effective in what we've seen in our work, uh, across four hundred and fifty organizations and really kinda digging into this AI space now.

I think the unglamorous adjective might be a bit debatable, um, particularly because for me, as someone who's worked kind of at the intersection of technology and social impact for the last eight years, I probably find some of these kind of unglamorous opportunities highly glamorous because they're really saving the, saving the sector time and headaches.

But one I would say surprisingly effective use case that we've seen support the sector and really kind of help to build upon some of the sixteen years of work that we at Vera have been doing to amplify the impact of the sector is using, uh, generative AI to build a donor report that matches a specific funder's template.

So, like the budget actuals, the qualitative assessment, the activity and milestone progress, the indicator results.

And ultimately, this is kind of work that program teams have historically spent days, weeks, sometimes months developing.

Um, and what we've seen is nonprofits using an LLM, connecting it via MCPs.

I'll unpack that because it might be a term that we use a little bit more in this, but MCP stands for model context protocols for those who are unfamiliar, and really it's just kind of an open standard that powers interactive tool support and helps kinda connect the apps and, uh, technologies that you're already using to bring that data directly inside your LLM chat.

So through these kind of connections, we've seen LLMs kind of pull together a donor report, so kind of supporting the project or program budget actuals from one system, the indicator results sitting in Excel or Qubo or Salesforce or Map Impact, and assembling kind of a first pass of the report, which is saving the sector, I would say, a, an enormous amount of time and kind of ultimately freeing people up to do a bit more kind of augmentation with AI.

So more of the critical thinking, the program analysis, the work that kind of improves outcomes.

Yeah, I think we've been in conversations with various organizations and seen this in practice and, and have a, have a cool demo as well that we're happy to share, share at any time for those that, that might find it useful.

Ultimately, I would say as, like, we unpack this further, we, we run AI peer learning groups with the sector more broadly, and as we unpack this in those discussions, I would say it's kind of the-- where AI earns and keeps its kind of trust is in repetitive low-risk tasks first, and that's kind of the place where a lot of nonprofits are getting started.

It's one of our kind of core guidelines for the sector is really, like, pinpoint where it lifts the workload or accelerates a process that's already well understood and already has the kind of right data pipelines underlying it and will save time and headaches ultimately.

Um, we've also kind of built an ROI framework around this and have a calculator that kind of helps make these decisions and help to kind of identify what those effective use cases might be that looks at how much time and money does it cost today, how much will AI help? So looking at, like, expected time savings, expected quality improvement.

Thirdly, what will it cost to implement? Um, fourthly, how will it advance your mission? Which kind of builds out a mission impact score.

Um, uh, fifth one is how will it change your team's capacity? So building out kind of that capacity impact score, and then layering on top of that kind of a risk level and applying that as a discount to really spit out an overall recommendation.

So that's also can be helpful, I would say, in identifying, like, for you as an organization or a nonprofit, what might be kind of the most, most effective at this point in time and might not be something that's kind of big and glamorous but, that you have on your radar.