Last month, I spent an afternoon with Claude building a funder impact brief and a supporting methodology document for a prospective funder. We’d met to discuss our proposal, and the funder needed more information quickly to move forward.
The brief had to answer three specific questions about projected impact across six states: how many people would be reached, how much money would be unlocked, and how many new opportunities would be generated. What would have taken a senior researcher 35 to 60 hours — and cost between $5,000 and $12,000 at consultant rates — I completed alone in four hours. Once I was done, I asked Claude how long it would have taken to build this without AI assistance. The answer reframed how I’ve thought about this work ever since.
What the Work Actually Was
The prospective funder wanted state-by-state project impact data: people reached, funds unlocked, and opportunities generated. Getting those answers meant tracking down and cross-referencing primary sources across six states. State annual reports, program websites, federal announcements, press releases, and academic fact sheets, then determining which figures were directly published versus derived, identifying errors in a large data set, tracking a rapidly moving federal program across six states. The audience was a sophisticated funder who would scrutinize every number.
Without AI, I would have needed a senior research or policy analyst with graduate-level training, deep familiarity with the field, strong data literacy, and professional writing skills. Then, this person would need to be able to drop other projects and start immediately. I did it all myself in four hours, using three AI research tools.
The Economics of Working with AI Are Disrupting Mid-Level Hiring
We often frame AI’s impact on the workforce as a threat to lower-skilled, entry-level workers. That framing is incomplete. The roles most directly affected are mid-career: the research associate two years out of grad school, the freelance policy analyst with a sustainable practice in grant writing or impact reporting, and the mid-level non-profit staffer whose value to the organization was producing exactly this kind of funder-ready documentation. These aren’t marginal positions — they’re the knowledge worker pipeline. The compression is happening because the time required to do this work has dropped dramatically, and the people who commission it know it.
Completing this project felt unreal — thrilling to produce work of this quality and depth, and hard to ignore what that means for the colleagues these tools displace. When I imagine other people gaining this proficiency, there’s no way it won’t change how they hire and retain skilled people — just as it did for me.
What AI Cannot Substitute
Someone still had to structure the work. I attended the meeting with the funder, understood what the funder needed, and built a plan to get solid answers. Without my direction, the AI wouldn’t know which of the hundreds of data points mattered or how to derive meaning from them — and it couldn’t supply the editorial judgment to decide what to emphasize or cut as the draft evolved.
What the AI did was close the gap between knowing what data and analysis we needed and having a finished product ready to share. It did in minutes what used to take days: pulling sources, checking figures against each other, rebuilding calculations when inputs changed, formatting a complex multi-state table consistently across multiple documents, and flagging its own mistakes as we validated line by line.
I supplied the direction. The AI compressed the timeline–and handled much of the research and analysis.
What This Means for Small Nonprofits
For smaller nonprofits — organizations without a research department or policy analysts, operating with a small leadership team, a couple of staffers, and trusted consultants — strong AI skills expand organizational capacity. A smaller, AI-competent organization can now produce funder-grade analysis that would formerly have required hiring another consultant. It can move at the pace of a funder relationship rather than the pace of a grant-funded position. It can generate a higher volume of high-quality work more quickly.
That’s a real gain for chronically under-resourced organizations.
But it almost certainly means the same organization will not hire the research associate it might have brought on three years ago, because it no longer needs to. A real person who went to school for this, has held previous jobs, and has been building a career in the social sector, now has one less opportunity. There are fewer seats at the table.
The Costs That Don’t Get Counted
The $5,000 to $12,000 we saved is a legible line item. So is the faster turnaround, the funder who got a detailed answer quickly, and the momentum that didn’t stall waiting for a consultant to become available. What’s harder to name is the work that didn’t get contracted out and the people who would have done it — one more economic pressure on nonprofit knowledge workers in a sector already running thin. The ease of reaching for AI to get the work done doesn’t require you to account for that.
Working with AI tools in your nonprofit organization
If you’re interested in improving your use of AI tools and workflows in your nonprofit organization, I offer short-term project work — typically five to thirty hours over a month or two — to help organizations set up and test these workflows, learn new skills and/or build custom AI spaces and projects.
Feel free to reach out if this is of interest to you: get in touch
Last month, I spent an afternoon with Claude building a funder impact brief and a supporting methodology document for a prospective funder. We’d met to discuss our proposal, and the funder needed more information quickly to move forward.
The brief had to answer three specific questions about projected impact across six states: how many people would be reached, how much money would be unlocked, and how many new opportunities would be generated. What would have taken a senior researcher 35 to 60 hours — and cost between $5,000 and $12,000 at consultant rates — I completed alone in four hours. Once I was done, I asked Claude how long it would have taken to build this without AI assistance. The answer reframed how I’ve thought about this work ever since.
What the Work Actually Was
The prospective funder wanted state-by-state project impact data: people reached, funds unlocked, and opportunities generated. Getting those answers meant tracking down and cross-referencing primary sources across six states. State annual reports, program websites, federal announcements, press releases, and academic fact sheets, then determining which figures were directly published versus derived, identifying errors in a large data set, tracking a rapidly moving federal program across six states. The audience was a sophisticated funder who would scrutinize every number.
Without AI, I would have needed a senior research or policy analyst with graduate-level training, deep familiarity with the field, strong data literacy, and professional writing skills. Then, this person would need to be able to drop other projects and start immediately. I did it all myself in four hours, using three AI research tools.
The Economics of Working with AI Are Disrupting Mid-Level Hiring
We often frame AI’s impact on the workforce as a threat to lower-skilled, entry-level workers. That framing is incomplete. The roles most directly affected are mid-career: the research associate two years out of grad school, the freelance policy analyst with a sustainable practice in grant writing or impact reporting, and the mid-level non-profit staffer whose value to the organization was producing exactly this kind of funder-ready documentation. These aren’t marginal positions — they’re the knowledge worker pipeline. The compression is happening because the time required to do this work has dropped dramatically, and the people who commission it know it.
Completing this project felt unreal — thrilling to produce work of this quality and depth, and hard to ignore what that means for the colleagues these tools displace. When I imagine other people gaining this proficiency, there’s no way it won’t change how they hire and retain skilled people — just as it did for me.
What AI Cannot Substitute
Someone still had to structure the work. I attended the meeting with the funder, understood what the funder needed, and built a plan to get solid answers. Without my direction, the AI wouldn’t know which of the hundreds of data points mattered or how to derive meaning from them — and it couldn’t supply the editorial judgment to decide what to emphasize or cut as the draft evolved.
What the AI did was close the gap between knowing what data and analysis we needed and having a finished product ready to share. It did in minutes what used to take days: pulling sources, checking figures against each other, rebuilding calculations when inputs changed, formatting a complex multi-state table consistently across multiple documents, and flagging its own mistakes as we validated line by line.
I supplied the direction. The AI compressed the timeline–and handled much of the research and analysis.
What This Means for Small Nonprofits
For smaller nonprofits — organizations without a research department or policy analysts, operating with a small leadership team, a couple of staffers, and trusted consultants — strong AI skills expand organizational capacity. A smaller, AI-competent organization can now produce funder-grade analysis that would formerly have required hiring another consultant. It can move at the pace of a funder relationship rather than the pace of a grant-funded position. It can generate a higher volume of high-quality work more quickly.
That’s a real gain for chronically under-resourced organizations.
But it almost certainly means the same organization will not hire the research associate it might have brought on three years ago, because it no longer needs to. A real person who went to school for this, has held previous jobs, and has been building a career in the social sector, now has one less opportunity. There are fewer seats at the table.
The Costs That Don’t Get Counted
The $5,000 to $12,000 we saved is a legible line item. So is the faster turnaround, the funder who got a detailed answer quickly, and the momentum that didn’t stall waiting for a consultant to become available. What’s harder to name is the work that didn’t get contracted out and the people who would have done it — one more economic pressure on nonprofit knowledge workers in a sector already running thin. The ease of reaching for AI to get the work done doesn’t require you to account for that.
Working with AI tools in your nonprofit organization
If you’re interested in improving your use of AI tools and workflows in your nonprofit organization, I offer short-term project work — typically five to thirty hours over a month or two — to help organizations set up and test these workflows, learn new skills and/or build custom AI spaces and projects.
Feel free to reach out if this is of interest to you: get in touch