AI Hiring - From Start to Finish

Every episode so far has been about donors. This one turns to your own team. Sean walks through a real search he ran for a client — filling a Development and Donor Relations Coordinator role — covering exactly where AI helped, where he deliberately kept it out, and where this sector needs to be careful.

AI handled volume: organizing a large applicant pool against clearly defined, objective criteria. Every actual judgment call — who advanced, who got interviewed, who got hired — stayed entirely human. The episode closes with concrete guidance on where to draw that line in your own hiring, and why transparency with candidates matters.

Sean is the founder of Magnolia Philanthropic Services, advising nonprofits nationwide on development strategy, with 25+ years in the field.

Transcript

SPEAKER_00 0:09

Every episode so far has been about donors, segmentation, adoption, what's actually working. Today I want to point this show at a different part of your organization entirely, your own team, and specifically how you hire for it. I recently ran a search for a client filling a development and donor relations coordinator role. And I want to walk you through exactly where AI was able to help, where I made a point of keeping it out of the process entirely, and where I think this sector needs to be genuinely careful. This is navigating the maze. Let's talk about hiring. I want to start here because I think it matters more than the workflow itself. Hiring is not like donor segmentation. When AI gets a donor tier wrong, you send someone a slightly mistimed email. When AI gets hiring wrong, you can end up excluding a genuinely great candidate. Or worse, systematically excluding a whole category of candidates without ever realizing that it happened. This is a well-documented problem outside the nonprofit sector already. AI tools trained on historical hiring data can learn and repeat the biases baked into who got hired in the past. If your sector or your organization specifically has a history of hiring a narrow type of person for development roles, a tool that's trained naively on that pattern will tend to reproduce it invisibly without ever flagging that it's doing so. So before I even get into what I actually did, I want to name the principle that shaped the whole process. AI touched the parts of this search that were about volume and organization. It never touched the parts that were about judgment on a human being. Now here's how the actual search worked. Step one, defining the role clearly before anything else. I know this sounds obvious, but it's the step that most searches skip. Before any candidate is evaluated, I work with the client to get genuinely specific about what this development and donor relations coordinator role needed. Not a generic job description pulled from a template, but the actual skills and experience that would predict success in this organization at this stage. Step two, using AI to help manage volume, not to judge people. So once the search was open, we had a real volume problem. We had more resumes and applications than any one person could carefully read from start to finish in a reasonable time frame. This is where AI actually helped. I want to be precise about what that means and doesn't mean. Step three. Every candidate who advanced was reviewed by a person and not a score. This is the line that I didn't cross. No candidate got automatically rejected by a tool without a human ever looking at their materials. No AI-generated score determined who got an interview. The narrowing from a large applicant pool down to a short list was AI-assisted in terms of organization, but the actual judgment calls at every stage were made by people who understand the role and the organization. For this search, that shortlist came down to four candidates who moved forward to interviews. Every one of those four was a decision made by a person reading their actual materials, not a number generated by software. Step four, interviews and final decisions stayed entirely human. I want to be direct about this. AI had no role in the interviews themselves and no role in the final hiring decision. Culture fit, how someone talks about donor relationships, how they actually behave in this specific team, none of that is something that I would trust an AI tool to evaluate. And I don't think this sector should either. So here's the actual guidance. If you're thinking about using AI in your own hiring, use it for volume and organization, sorting large applicant pools against clearly defined objective criteria you set before you look at a single application. That's a legitimate, genuinely useful application of these tools, and it can free up real time for the part of the hiring that actually requires a person. Don't use it to judge people. Don't let a tool auto-reject candidates without human review. Don't let it evaluate anything subjective, communication, style, cultural fit, how someone would actually perform in the role. Because that's exactly the kind of judgment where bias hides most easily and where the stakes of getting it wrong land on a real person's career, not just your organization's efficiency. And if you do use AI anywhere in your hiring process, tell candidates. Transparency here isn't just an ethical nicety, it's becoming now a legal requirement in some places. And it's simply the right thing to do when someone's livelihood is part of what's being evaluated. That's the search. Start to finish. AI helping with volume. People making every actual judgment call. I think that line is going to matter more and more as these tools get better at sounding confident about the things that they shouldn't be trusted to decide. Next week, I want to go back to a number that I mentioned back in episode two, and I never fully unpack that.