How ChatGPT Ruins the RFP Process & How to Solve

Nonprofit RFPs vs RFCs

[00:00:00] . Following on with our favorite guest of the moment right now, Heather Yandow, of course, founder of, nonprofit ist, and consultant at third Space. We are talking about whether or not, because this has been coming up a lot, will G p t destroy RFPs? And I know you are the, like captain of the team, of folks that think RFCs request for conversations are superior and many ways to RFPs request for proposals.

[00:00:54] So we’re gonna throw a lot of acronyms out here. Where and how do you think our conversation since the last time we had it with regard to R F P? Versus R F C is in the landscape now of G P T. So I think it’s just gotten more complicated, more muddy, and I lean even more towards requests for conversations now.

[00:01:17] You know, one of the things that I have already started seeing is chat, G P t AI generated cover letters. So I just hired for a position. Hmm. And one of the cover letters was clearly generated by ai. It used exactly the same language, and it didn’t mention that the person who was applying had a personal connection to me.

[00:01:39] So to those, those were red flags. Didn’t even have to use any tool, just read it and thought, this is not right. So we’re already starting to see it. I suspect that if I was a nonprofit putting out a, a request for proposals, I would’ve already gotten some that were AI generated. So it’s becoming, RFPs in my opinion, are becoming increasingly useless, increasingly challenging when the goal is to actually find the right consultant to help you with your challenge.

[00:02:10] For sure. I think, you know, we’re using G P T synonymously with any generative AI tool that will create an output based on an input. I think there are ways of designing this, but I think there are tiers, right? Mm-hmm. I think, frankly, legally speaking, if you have government funded projects you have to solicit for, and with RFPs, request for proposals is just part of the game.

[00:02:35] You need three competitive bids. And that’s just how the government cookie crumbles. Yep. There’s a lot of white space below that though, however, where, you know, I see small projects, we’ll call them projects under 60 K or even under 30 K, where if you put out an R F P, like is disturbingly easy to generate a proposal?

[00:02:56] Like I have a proposal generator for for whole whale. I don’t use it because it’s just not how we go about it. But let’s say I was a do anything now shop. Literally, I could just go through, copy a proposal, put it in, see if I get it, and then go forward. And I think there’s real risk of having massive you know, signal to noise issues in that process.

[00:03:20] Does, how does that land for you? Yeah, I absolutely think so. I mean, you know, putting together proposals. Is super time consuming especially for small shops. You know, there are two and a half of us, so to muster the resources it takes to put together a really good proposal is a heavy lift, and I can absolutely see why using chat G P T or something else, like it would be really attractive, right?

[00:03:48] It’s certainly going to reduce the amount of time I’m gonna have to spend writing. It can be a, a good jumping off point. I think that what we’re likely to see is that those. Organizations who can take advantage of it. So particularly those organizations, those nonprofit consulting firms who have people who write proposals are gonna probably take advantage of it.

[00:04:11] And George, they might be, you know, doing it with all of the respect and within the bounds of good AI usage, which I know you think and talk a lot. But I do think it’s gonna create more proposals and not necessarily more useful proposals for nonprofits to review.

[00:04:28] Yeah, I think all roads lead to conversations, though. They do. I think at some point, you know, you’ll vet, you’ll go through, there’s just a lot more processing that ends up happening on the nonprofit side, albeit ironically, you could also use AI for synthesizing that. And we end up back full circle to just have the conversation front.

[00:04:47] Have, have the, you’re gonna have to have it anyway. So being like our, you know, like our request for information or intent could lead to a conversation round and. That would maybe filter out a lot of this because the number of proposals you are going to get is simply gonna increase over this threshold because it’s so easy.

[00:05:06] Literally. Yeah. If you are saying like, oh, he’s talking about some advanced technical thing. Here’s what I want you to do. If you’re a consultant right now or if you’re a nonprofit, I just want you to understand how simple this is. You just go on ideally to check GT four or Anthropic. If you have an account there, what you’re gonna do is prime the conversation with who you are, what you do, and the role of that ai.

[00:05:27] The next thing you’re gonna do is here is a sample of the structure of my proposals. Here’s a couple case studies now. That you’re clear, please write a and respond to the following context of this new proposal. And you’re gonna toss that in and you’re gonna end up with something that’s disturbingly good as a first draft.

[00:05:47] And frankly, if you’re lazy, just send it as mm-hmm. Whatever your first, your, your final draft. Uh, I do, as you mentioned, talk a lot about keeping human in the loop as soon as you send it out to the world, if it’s. If it is all within your control, please, please make it a first not final draft policy of using AI and keep humans in the loop for now when exposing LLMs publicly to individuals especially if you are in I’d say crisis or trauma adjacent conversation.

[00:06:19] For, for folks, what is L L M? Large language model. I feel like this is just alphabet soup. Uhhuh, you’re a generative AI thingy. It like there’s nuance, there’s fine. It’s what we’re talking about. So going back to that, what you’re gonna do is create that. Now, if you’re on the nonprofit side, here’s how you go about it.

[00:06:37] You say you are a selection criteria. AI that evaluates proposals for our organization. Here is the proposal we created. Now, evaluate this and break out based on price, competency likelihood to deliver on time elements, and shove it into the spreadsheet for me. And ta-da. You’re just gonna go back to having.

[00:07:00] A conversation. A conversation. Absolutely. Absolutely. ’cause it’s, it’s not buying the best digital camera, right? It’s not a spreadsheet able thing. Finding a consultant, most of the time you actually wanna know if you’re gonna be able to work with this person. You wanna have an understanding of who they are, of what their personality is, of what their style is, and.

[00:07:19] Certainly we talk about that in proposals but it comes back to having a conversation and really seeing where that conversation goes. And that’s something AI at least currently can’t actually do for you. No. And the truth is, you’re gonna be working with a person, not an ai. That’s right. Uh, one of the things I do and have always loved about non-profit is, is that you can just go shopping for folks and just say, Hey, I wanna have a quick conversation.

[00:07:44] One click sends them a message so that you’re like, all right, look. I have this fundraising campaign. I could put it on our feet. Lemme just talk to a few folks and see That’s right. What they think about the project. What am I missing in here and how do I go about that? Yeah. Yeah, absolutely. And that’s, that’s, we designed it that way on purpose, right?

[00:08:02] It’s people’s pictures there, not their logos. We know that you’re connecting with a person and we wanna make that as frictionless as possible. Call ’em, email ’em, send ’em a message through the site, any way you wanna get in touch with them. And then yeah, have that conversation. What does this look like?

[00:08:20] What do you think? What are the questions you have for me? What would the timeline be for this? Yeah. And I, I do think there’s some risk as well, which is why I want more foundations to do this. I’m to, to pay attention and to communicate to grantees about the, the risks here, because when you go out there with.

[00:08:40] An R F P, you can get a lot of inbound and potentially unethical inbound where folks can just sort of fake it till they make it, but they’re using language borrowed mm-hmm. Stolen from other groups or agencies, and they can mimic that style and confidence. And suddenly, you know, you’re, you’re selecting a an inferior option or someone who has no clue what they’re doing, but the jargon lines up just fine.

[00:09:04] So you’re like, oh yeah. And. Oh, they’re half the price isn’t that nice? Like Uhhuh. Uhhuh as many times as you need to. You get what you pay for. Yeah. You’ll learn that it’s an immutable truth. It is. And you know, obviously having conversations is gonna help break through that. And then always checking references.

[00:09:22] Who else have you done similar work with? Can I talk to them? Let’s have a conversation with those folks about what the experience was like working with these people. Yeah. I don’t know if there’s any other points in here. I have already just broken the entire system by literally explaining how to build a R F P generator and then R F P proposal evaluator, so that you just have this little disturbing, dystopian mm-hmm.

[00:09:47] Cyclical behavior. But you know, to come back on, you know, where this sort of started requests for proposals, RFPs are so anachronistic like they date back because in the industry of the 1880s, they needed to put this stuff in papers. So they needed to put it in a paper to be like, here you go. All come, come find my, you know, my road construction project or train my railroad construction project, Uhhuh.

[00:10:13] And it wasn’t until the 1960s, thereabouts, where the government really adopted this as a standard practice for large purchasing projects. Yes. Government size stuff. Yes. Not. Tens of thousands of dollars, right? Yes. You’re like, oh, it’s so much money. It’s not, it was created when the government’s about to spend, you know, you know, X millions, hundreds of millions of dollars.

[00:10:37] That’s what it was designed for. The government, as you remind, like I I said, is like, is requirement when you give to a nonprofit and use government dollars, like, oh, no, no, you gotta do that. R F P process. Mm-hmm. But understand that’s where that comes from. That’s where that comes from. And there’s, there’s a belief that that’s the right way to do it.

[00:10:57] That’s the professional way to do it. That’s the equitable way to do it. And I would argue for all of those, that’s, that is not necessarily true. It is not always the best way. It is always not always the most professional way, and it’s certainly not the most equitable way to find someone to work with.

[00:11:15] Yeah. I don’t know. Maybe to play the other, other side here potentially. One of the things that, when you use the word equitable, in my mind it, it means you have to have the capacity and resources and capability to go about the very lengthy process of creating a proposal. Yeah. And that process. And there’s many folks in the nonprofits network that like definitely bristle.

[00:11:38] They don’t even like go after. Yeah. They’re like, no, no, no. I won’t even bother. Which means you’ve already precluded a lot of qualified candidates from applying. That said, I just explained how you could create a proposal builder so that you could get to the conversation. May, maybe the, maybe there’s a bright side there.

[00:12:01] Maybe there’s a bright side there. I, you know, I’m one of those folks who I, I, I don’t do cold proposals. And that doesn’t mean necessarily that I have worked with a nonprofit before, but I need to at least be able to have a conversation. Very rarely. Does a two or four page R f P have all the information I need to know even what to pitch as a first option for how I might be able to support this organization.

[00:12:26] I often have lots of questions. I wanna know a little bit more about the history. I wanna know about why now. I want, you know, I wanna understand why the budget is where it is and what the board’s buy-in is things that people don’t often write down in their request for proposals. So even that first conversation again, Warming it up a little bit, having a sense of who’s really there, what the real challenges are is, is super helpful.

[00:12:52] I will say for Whole Whale that we do respond to RFPs, but only if there was a conversation first. Yep. There you go. Like everything starts with a conversation just to make sure we’re aligned. Yeah. Are we in the ballpark? Is our type of service, meet your type of need and. We do churn out a lot of proposals.

[00:13:10] However, they’re much more like project plans, like mm-hmm. We literally take that we and move that into a contract parts of it and say, this is what we’re gonna execute on. ’cause that’s what we talked about. And you know exactly what you get. So you’re already doing pragmatic work now? Yes. We, uh, we do lose a number of proposals.

[00:13:29] But that’s, you know, that’s part of the game. Yeah. And I think of them less as proposals, maybe more as like project plans to make it. More tangible. Mm-hmm. Mm-hmm. Yeah, that makes sense. That makes sense. We often put together proposals that are those project plans, and then the very, if we get the work, the very first piece of that is let’s actually dig into this project plan and figure out if it makes sense.

[00:13:51] We were basing this on an R f P in a 20 minute conversation. Right. Like, we don’t, we, we need some more information. Well, I’m excited to put this out there. I think the more we talk about it, it’s just like, it’s like this quiet secret and you’re like, oh. Mm-hmm. Like nobody knows that a hundred million plus people are using G P T tools to like write all manner of thing.

[00:14:11] Like we know Yes. We just aren’t talking about it. Yes, yes. You know, it’s like if the, if the teacher in the classroom accidentally left the entire answer key on the chalkboard while you took the test and everyone was like, Is anyone gonna tell the teacher that it’s there? Like, can we just start having this conversation?

[00:14:29] We know it’s happening, we know it’s going on. And by the way, if you’re doing it lazily, if you’re doing it in a poor way if you’re ever curious, you can go to tools like G P T Zero. Put in that text and you’re gonna get a what’s called perplexity and burness score, which is, uh, was this probably created by, uh, an AI or not a generative ai or not.

[00:14:52] Uh, and so if you are bad at prompting, if you’re doing this in a lazy way, uh, it’s very much detectable. Mm-hmm. But you don’t realize that. But people that know, know, mm-hmm. And I, I, you know, The plus side is there’s lots of ways that these tools can help us as nonprofit consultants. You’re actually doing a webinar on that very soon.

[00:15:12] There’s lots of ways that these tools can help nonprofit leaders. So there is a positive side. Just, yeah, be careful with the RFPs. Agree to agree. All right, Heather, all, thanks again. Thank you. And folks can find you at nonprofit, do IST nonprofits. That’s correct. And thanks for the community you’re building.

[00:15:30] Thanks, George.