Work Smarter, Not Harder: AI for Fundraisers

Welcome back to the Petrus Development Show! In this episode, Andrew and Rhen share their experiences with AI for fundraisers, highlighting practical tools and strategies to transform your development office.

Are you new to using AI in your work? Perfect—this is your starting point. Andrew and Rhen reveal their favorite tools and offer examples of how they're using AI to save time and boost results.

Yes, AI can feel intimidating. But here's the truth: it's not here to replace the heart of fundraising—donor relationships. Instead, it eliminates tedious tasks that steal your time, freeing you to focus on what truly matters: mission-driven work and meaningful relationships.

Tune in to discover how AI can become your secret weapon for fundraising success.

INTERVIEW TRANSCRIPT

06:31.58
Host
Well, howdy everybody. Welcome back to the Petrus Development Show. I am Rhen Hoehn with Petrus, and joining me today is our owner and president, Andrew Robison. How's it going, Andrew?

06:42.14
AROB
Going well, Rhen. How about yourself?

06:44.44
Host
I'm living the dream.

06:46.13
AROB
Right on.

06:46.30
Host
I know you guys had, a couple weeks ago, a nice kind of winter storm come through. Did it knock you out like a couple years ago?

06:51.29
AROB
Yeah. No, thankfully not. No, the Snowmageddon or Snowpocalypse or whatever that was in 2021, that was bad. In Texas, it shut the whole state down. And then through, I think politically, there are a lot of people still recovering from that, from losing the power grid and all of that. So thankfully—

07:15.54
Host
Yeah, I was supposed to be with you in Texas that week.

07:18.61
AROB
That's true.

07:18.71
Host
And you've come to Michigan a lot. I've come to Texas a lot. It's the only time that the snow in Texas has canceled one of our trips. Usually it's the other way around.

07:27.77
AROB
Yes, usually it is. No, so thankfully it wasn't that bad. We did get a good amount of rain as the temperature was dropping. So there were, in College Station, we had some pretty serious layers of ice on every car, ground, you know, everything that we had to—

07:48.53
Host
Oh.

07:52.98
AROB
It took me—oh, I don't know. This is probably nothing for you, but it took me probably about 20 minutes to just get the door open to my car after it had been icing so bad.

08:01.49
Host
Dang.

08:05.82
AROB
Dallas—my brand new neighbor, they just got married—in Dallas, and he said that they drove around a little bit and it was like a ghost town because there was just nobody out because the ice was really bad up there. So thankfully it wasn't too bad for me. Certainly no big issues, but yeah, Texas is not—it's not prepared for that. I mean, we can handle, you know, six weeks of triple digits with no problem, but you drop the temperature below freezing for, you know, two days and we don't do too well.

08:31.86
Host
Right, yeah.

08:43.77
Host
It's over. Yeah. At least it's just difficult. You can't really be prepared for that.

08:46.74
AROB
Yeah.

08:49.24
Host
You know, one of our consultants, Tara, was traveling, working with one of our clients in Canada, in Peterborough, in Ontario. She was all around the region when a snowstorm came through, and she had all kinds of issues to get through.

09:05.50
AROB
She did. She sent me a picture, and she had gone over this bridge that she had literally just driven over. It was—I mean, just snow piled on the bridge, no rails, you know, no guardrails. I was like, "You're not going to drive back across that, are you?" She goes, "Oh no, not until they plow it," which—

09:06.55
Host
One of them... Go ahead.

09:25.40
AROB
She was a little bit surprised that the Canadian snow removal infrastructure was not as robust as maybe it is in parts of the United States that get a lot of snow. She was a little bit more rural out there, but she was like, "I would think that they know how to move the snow a little bit quicker." And it wasn't the case. So she ended up getting stuck for a couple of days, but thankfully did get home eventually.

09:51.45
Host
It was rural enough that she couldn't get maps to work on her phone.

09:55.80
AROB
Oh, that's right.

09:56.82
Host
So she had to go old school, and she still had a Garmin GPS unit, the kind that you plug into the cigarette lighter and stick up on your dashboard.

10:02.10
AROB
Yes.

10:06.23
AROB
Yes.

10:06.86
Host
That's what she relied on to get her own.

10:09.50
AROB
Yes. Did you ever have one of those Garmin or TomTom or—

10:13.27
Host
I always laughed at the one that we had. It always, when you searched for something, would always take you to the closest one as the crow flies. So here in Michigan, you'd search for a restaurant, and it would show you one in Thunder Bay, Ontario, which I'm sure was like 90 miles away, but it would take like six—but probably more than that, eight hours to drive there around Lake Superior. I always got a laugh out of that.

10:39.88
AROB
That's pretty funny. When Cheryl and I did our—we did a big road trip back in 2008 because I did grad school at St. Mary's University in Minnesota, which was a distance program, but we had a graduation ceremony. We were invited to the actual graduation ceremony for the school, and so we decided to go up, but we made a road trip out of it. And how long was that? Ten days? Twelve days? Something like that. But we bought—I think we already had a Garmin, but in advance of the road trip, we went to RoadsideAmerica.com and came up with a list of like all these random sites that we could see along the way. You know, the world's largest fork and, you know, a random statue of a dinosaur in a cornfield. And it was bizarre.

11:25.05
Host
Yeah.

11:32.34
AROB
But the nice thing about that with the road trip is we would pull up the Garmin and, you know, we had to drive 10 hours that day, but we'd pull up the Garmin and plug in the next address, and it'd be like 45 minutes to your destination. And then we'd slap it on the dash and start driving. And so it definitely broke the trip up a little bit until it got into like one of those cycles where it was like, "Recalculating... recalculating... recalculating."

11:58.22
AROB
So I'm usually going through the big cities, but yeah, I haven't used one in a long time. So Tara being able to find one or buy one was definitely a blast from the past.

12:09.98
Host
Yeah, it makes you appreciate how fast, how far technology has come in just the last 10, 15 years, right? And that kind of brings us to our topic, I guess, of today with using AI tools.

12:22.52
AROB
Oh my gosh.

12:22.90
Host
And oh my goodness, yeah, and how far that's come even in just the last year or two. We did a brief episode on using AI tools in May of 2024, right? And the landscape is completely different now than it was then, even compared to September. Everything has changed. New models have come out that take things to a whole new level. And the use cases for even just kind of your standard nonprofit are getting more and more useful. And it's kind of taken off. So that's what we want to talk about a little bit today—some of our hot tips for AI hacks and the ways to use some of these tools right now.

13:03.96
AROB
Yeah, I think this is great. You know, we have our membership program, the Harbor, and as part of the Harbor, you get access to our custom Petrus-trained custom GPT, which is built out on the OpenAI platform, which is the creator of ChatGPT. And so we've made that available. And in the calls and the monthly calls that we have with the members and in some of the conversations we've had on the chat, it's pretty cool to see some people that are really unfamiliar with AI or haven't really ever used it beyond, you know, asking it funny questions or doing simple searches to actually start using the tool for fundraising purposes. And the feedback has been very positive just in terms of like, "Hey, I didn't trust the platform. I didn't want to plug too much information into a platform that I didn't know what it was going to tell me." And, you know, but knowing that you guys, that Petrus has trained this, has really made me more comfortable. And now that I'm using it, I'm finding a lot of really simple use cases that are making my ability to run the program and specifically, you know, fundraise a lot more effective and efficient, which is great. That's what we want, right? We want you to find ways to use these tools to then ultimately make you more successful in the work that you can do building relationships with your donors and offload some of the other tasks that need to be done.

14:23.42
Host
Yeah.

14:38.78
Host
Exactly. I mean, it's pretty cool in Catholic circles right now to be anti-AI from what I've seen.

14:43.74
AROB
You know.

14:44.84
Host
Just kind of a general sentiment of, "Oh, I would never touch that." And clearly anything that's people-facing shouldn't be AI, right? We're not replacing our donor visits with AI. The ministry that we do—if you're a campus ministry, AI should not be counseling your students, right? Things like that. But there's a lot of processes that are tedious and suck up a lot of time. And there's no need for a person to be doing it. If you can replace that with a computer, with AI, why wouldn't you, to free you up to go and be out doing people-to-people ministry and fundraising?

15:12.73
AROB
You know.

15:13.50
Host
Good.

15:18.58
AROB
Yeah, absolutely. I agree.

15:20.70
Host
Good.

15:20.76
AROB
So what are some of our hacks?

15:20.98
Host
So, oh, go ahead. Yeah, so I think in previous episodes, we talked about ChatGPT a little bit. We probably—I should have listened back. I think we mentioned Claude, and we might have mentioned Grok. Those were kind of the big three in our minds like a year ago. And in this episode, we're going to kind of ignore Grok, if that makes sense. I think it still has a lot of capabilities. It seems better for personal use than for professional use, just kind of the general vibe of it. The vibe of it.

15:48.47
AROB
Well, Grok—yeah, Grok, I would say the strength of Grok is that it's a little bit more entertaining than some of the other AI tools, meaning it'll joke around with you a lot more. It'll turn your pictures into silly videos, which, you know, there's value to that, right? From a professional standpoint, I think the value that Grok brings over some of the others is all the other platforms—or all the other tools, ChatGPT, Claude, Gemini—they're trained, but like the training goes up through a certain date, right? They're loaded in, they're taught, they're trained up to a date, and then that may be where their knowledge base ends. You know, like ChatGPT was—

16:38.23
AROB
For a long time, it was famous for, you know, its training was eight months old or 14 months old or whatever it was. And so if you asked it a current events question, it just didn't know. Grok, because it's connected to Twitter, brings in information that's, you know, real-time through Twitter. And so it can be helpful from a, "What are—I want to understand how this might affect current events," kind of a situation. That would be its strength. But I agree from a workflow standpoint, I pretty much stick to ChatGPT, Claude, and Gemini. You use Gemini a lot more than I do, but I definitely use it just because we have Gmail and we use Google Docs, and it'll make recommendations as I'm typing and all of that. So certainly there is—it's integrated whether I'm actively using it or not.

17:30.52
Host
Yeah, I think Gemini was kind of seen as a joke up until six, eight months ago. And they're probably the fastest ascending model, Google's Gemini. So we'll get to that. All the models are very confusing to keep up with, right? So we'll give you some general tips. And a month from now, it'll probably all be obsolete anyway though. The pace things have been going.

17:46.74
AROB
Can't wait to hear it.

17:48.54
Host
But if you aren't familiar with AI tools, we'll give you a sense of where to start. And so I'll talk about some of our use cases for those big three models—ChatGPT, Claude, and Gemini—and I have another tool I want to mention as well that I've been playing around with that has changed my work pretty dramatically just this week because I've been toying around with it. So...

18:06.95
AROB
Can't wait to hear.

18:07.83
Host
Yeah. So let's start with ChatGPT. I think everybody knows chat. They've all heard of it and probably we've all, you know, toyed around with it a little bit. Maybe not for work. Maybe you just kind of, you know, made it build coloring pages, had it write something for you, asked it some questions with mixed results. Where is ChatGPT most useful? How do you use it? How do you approach that?

18:33.24
AROB
Yeah, so I've got—just so the users know, I've got Pro subscriptions to both ChatGPT and Claude. So, you know, whatever features are available that are beyond the free is kind of what I make use of. Really, that just comes down to I can have a lot more conversations, and I can create a lot more images at that Pro level than the free. Not that I do that too much, but I don't get hit by those limits, those token limits from uploading documents that are too long or anything like that.

19:06.23
Host
Yeah, and I'll pitch in. One of the benefits of that is you get the newest model. And I think the Plus account—I think there's Plus and Pro, and I think Pro is like 200 a month.

19:10.49
AROB
Yeah.

19:14.04
Host
That's pretty—I think the Plus is 20 bucks a month, so it's not too bad.

19:14.49
AROB
Okay. Yeah, no, I don't have that.

19:17.64
Host
But historically, the ChatGPT models have been very difficult to keep up with because they all have strange names and they all have different kind of focuses, and it's not obvious. It was like, "Oh, 3.5 mini," "O3 whatever"—like they're all different, numbered names. Their newest models, ChatGPT-5, pick the right model to run based on your question. So it says, "Okay, this is a math question. I'm going to focus more on our model that's good at computing things," etc., etc. So it kind of picks based on this. You don't worry about that so much with ChatGPT anymore.

19:51.67
AROB
Yeah, so what do I use—what was your question? What do I use ChatGPT for? What's it—

19:59.22
Host
Yeah, where do you find it the most useful?

20:01.18
AROB
So between ChatGPT and Claude, which are the two—and then you can chime in on either of those and then also add Gemini to it—I find that ChatGPT, I feel more comfortable asking ChatGPT questions that it's going to give me an answer to. Maybe, you know, I want to understand history. I want to understand, you know, something about a theological question that I'm, you know, sort of—or, you know, a biblical question I'm asking. You know, maybe I want to include something in a document. So I ask it a lot of questions like that. And then I also use ChatGPT for brainstorming. So, you know, if you're writing, sending out an email, you want to come up with a list of good subject lines, that's a really good use for ChatGPT. "Hey, here's my email. Can you suggest 10 subject lines for it?" So brainstorming is kind of where I use ChatGPT. Claude, I much prefer the writing style of Claude. I just—I think that it's—they both write well. I think ChatGPT has some tendencies in kind of how it writes that I just don't—I like the way that Claude is better. You know, it's like, you know, to some extent, it's, you know, I read a lot of books. I like this author over this author because I appreciate the writing style a little bit better. There's probably somebody that says, "Oh no, you're an idiot. The writing style of this person is better. How could you say that?" So personally, I prefer Claude as the writer and scripter and kind of author of whatever I'm kind of having it write, where I might then also at the same time pull up ChatGPT and ask ChatGPT to help me brainstorm ideas to then have Claude write. So it's kind of a team approach for me.

21:51.22
Host
Yeah, I think just for general use, ChatGPT is really good. And I really like that you can add projects in there and add certain files to the project folder that are then references. That's one of the difficulties—it starts hallucinating results the more and more information you give it. When you create a project and upload like a reference file, it keeps referencing that file over and over again when you ask questions. So we make a lot of—build a lot of courses, right? And I like to create a project for each course, upload the script for that in there, and then I can ask the questions of, "Hey, what lesson was this in?" Or, "Give me a brief summary of this one lesson." Saves me a lot of time on simple tasks like that when I'm working with large kind of quantities of data. I've also been enjoying the—it has an agent mode where I can kind of pull up some different capabilities on your computer screen. So I had it do some spreadsheet work for me where I had a list last week of 1,500 addresses, and they were all mixed format. So some of them were city and state. Some of them were full address. Some of them, the state was spelled out instead of the abbreviation. It was a mess, and I needed to pull the state out of the addresses without losing the order of the spreadsheet, and it was a pain. I could not find a way to do it just with a formula because there were so many different formats all mixed in there. So I dropped it into ChatGPT and said, "Hey, can you give me just the state from each of these addresses and keep it all in the same order?" Did an agent mode, and it was one or two minutes, and it did it all almost perfectly. The only thing that it messed up was that when an address had "Northeast" in it, the street address, it read that as Nebraska. So I did have to go back and look for any Nebraskas and check that it was actually Nebraska. But that was, you know, that took me five more minutes. Something that would have probably taken hours, you know, compressed down to a few minutes. And so tasks like that that are going to save you a bunch of time and effort—there would have been no benefit to me brute-forcing that process, right? Super, super handy for that.

24:00.30
AROB
That's great. And you were using agent mode for that in ChatGPT?

24:03.16
Host
I was. It did seem to be more accurate doing it that way. So—

24:10.01
AROB
Yeah. One thing that I had—an incident recently where I used Claude to help me edit a report where we were analyzing a ton of data for a client. And, you know, so I would write it, and then I would throw that copy into Claude and say, "How would you edit this text to make it, you know, cleaner or flow better?" whatever. And then over the course of that process, you know, and then I would have a section done, and I would throw the section in there into Claude, and I would say, "Are there any gaps in this that you can see that I need to go back in and fill in?" So it was kind of acting as my editor. Well, over the course of the conversation, eventually it had seen—Claude had seen my entire report, right? Well, about a month later—and again, this was a lot of data—about a month later, that client asked me to analyze some additional information and then kind of do an addendum to that report.

25:00.47
Host
Yeah.

25:15.22
AROB
Well, it had been—you know, a month may not sound like a long time to some people, but my brain was not remembering all of the information, the specific details from that report a month later.

25:25.47
Host
Hmm.

25:27.68
AROB
And so I went in and I did the, you know, analyzed the new data, did the addendum, and threw that into the—and I pulled up the same conversation that I had been having with Claude, you know, a month ago or six weeks ago, and said—I asked that same question: "Do you see any gaps in this report that maybe I missed from this addendum from the original report?"

25:51.93
AROB
And Claude had no problem remembering very specific information that had been in the report, you know, six weeks ago, that was able to say, "You forgot about this aspect of the original report, which I think should be included in your addendum."

26:00.63
Host
Thank you.

26:08.03
AROB
And so it was like, "Oh my gosh." Yeah. So that memory of AI is much better than the memory that I have for, you know, detailed information like that. And so it helped me to kind of catch some things that maybe I had forgotten. So, you know, I could see that being really—that was useful for that case and analyzing that data. But also, eventually, I could see that being really helpful, you know, if you're building out a donor cultivation plan, right? You pull up maybe your action items from the last 10 years.

26:35.71
Host
Yeah.

26:39.42
AROB
And, you know, the AI can analyze that information in a way that it remembers, "Oh, you know, this is something that they commented on a long time ago that maybe you have forgotten. Here's a way to, you know, work that into the conversation now."

26:57.78
Host
100%. I love that. Another little task that I did using ChatGPT specifically, again, this time was—actually, this was my Christmas cards. I got a whole bunch of Christmas cards. I wanted to make a mailing list out of them. And so I took pictures of all the return addresses on the envelopes, and I just uploaded them to ChatGPT. I was like, "Hey, just make a mailing list from this for me." And bang, just like that, right? Thirty seconds instead of 30 minutes of typing them all out. It was done. It made one mistake, and it was on some writing that I had a hard time reading as well. So I think back to my fundraising days—send out a mailing. Even when we did NCOA updates, we'd always get dozens of mailings back with the return service requested of, "Oh, this person moved to this address." You have to go in, type it into the database. And, "You know, this person's deceased," like all these changes. If I could have just taken a picture, uploaded it all, and said, "Hey, give me a spreadsheet with these changes," it would have saved me hours of time every mailing, working in our fundraising office. So little tasks like that that, again, there's no real value to a person doing it. It's super nice to hand it off to the computer.

28:07.58
AROB
That's awesome. Now you've been using AI for a little bit more image and video stuff, right?

28:15.48
Host
Yeah, mostly image so far, but the video has definitely taken off pretty fast here.

28:17.11
AROB
Image. Yeah.

28:22.30
Host
On my LinkedIn pretty recently, I put a comparison of some video generation from 2023, one from late—there's a prompt that's kind of popular that went around of Will Smith eating spaghetti. Did you—have you seen this?

28:38.62
Host
Yeah.

28:38.68
AROB
Yeah. Yeah.

28:39.58
Host
Yeah.

28:39.66
AROB
Yeah.

28:39.82
Host
And in 2023, it looks cartoonish.

28:39.94
AROB
100%.

28:41.86
Host
It looks ridiculous. It's clearly—it's trying. It's getting close, but it's not.

28:45.19
AROB
Yep.

28:45.98
Host
And people make fun of it and, you know, the extra fingers and the strange faces. And people still point to that. When you look at 2025, if you didn't know that that was AI-generated, you would think that was a video of Will Smith eating spaghetti.

28:58.46
AROB
Hundred percent. Yep.

28:58.71
Host
Yeah. And so things have come really far, really fast with that. And again, we haven't had much use for the video generation. My kids and I enjoyed—there was a series when OpenAI's new video generation, Sora, came out in the fall. A lot of people were making videos of cats playing instruments in the middle of the night. I don't know if you saw this whole series or—we had a lot of giggles out of watching those. But it looks like, you know, a cat playing the guitar on somebody's front porch at midnight. And it looks real, right?

29:27.16
AROB
Crazy.

29:28.86
Host
So it's getting really good. I have been using Gemini, Google Gemini. They have a—there's an actual name for it, but it's more well-known as Nano Banana is the model.

29:41.02
AROB
Okay.

29:41.14
Host
Whenever they create a new model, they give it like a fake name and put it into this kind of testing area for people to try out and work out new bugs. And so they all have, you know, pseudonyms, and Nano Banana is the name. And everybody loved it so much that they kept it when they actually announced it. "Hey, this is actually our model."

29:57.00
AROB
That's funny.

29:57.18
Host
Yeah.

29:57.20
AROB
I didn't know that story.

29:58.49
Host
Yep. So that's where it comes from. But the images that you can create with Nano Banana—phenomenal. And the ability to edit them and fine-tune details way better than anything else. Every so often, you and I and some of our other consultants will text, you know, comparisons of, "Hey, I tried to make this image with ChatGPT and this one with Google," and such, and showing the comparisons of them. And Nano Banana is by far the best, right? We had some people asking, you know, not every parish or organization has kind of a real logo. They don't always have need for one. And then it comes time that they need a logo, and like, "Hey, let's use this stained glass window that everybody knows that comes from the parish," right?

30:41.21
AROB
Yeah.

30:41.42
Host
And like, "Oh, turn this into a logo." Okay. This came up recently—upload a picture of the stained glass window into Gemini and the Nano Banana and say, "Hey, create a cartoon kind of style logo of this stained glass window, but leave out the frame, leave out these plants that are sitting in front of it." And it just nailed it, right? So it's a handful of seconds to generate that. Pretty awesome, pretty useful.

31:06.32
AROB
That's awesome. Now tell me about this tip or this tool that you've been using this week that is revolutionary.

31:15.86
Host
Yeah, I've heard people saying that typing is one of the big bottlenecks when it comes to work and that the future is voice. And I rolled my eyes like, "Who wants to sit and talk into a computer all day?" And I gave in and have been trying out—there's a tool called Whisper Flow. And Whisper is, I think, W-H-I-S-P-E-R.

31:39.26
AROB
Who makes Whisper Flow?

31:39.32
Host
Whisper Flow. I think Whisper.

31:41.21
AROB
Where is it? Is it? Okay. So it's not part of OpenAI or Claude or Gemini?

31:42.94
Host
I think—I think—no, it's its own tool, I believe.

31:46.28
AROB
Okay.

31:47.30
Host
I haven't looked super closely, but I'm pretty sure it is. And it's a voice-to-text. And it's not as simple as the one on your phone. So it uses AI. So when you—it's designed to understand whispers. Like if you're in an office setting, you don't want everybody talking, yelling into their computer.

32:01.02
Host
So it's meant to pick up if you just whisper into the microphone.

32:03.83
AROB
Oh, yeah.

32:04.46
Host
And I know people hate listening to people whisper on podcasts. I apologize. But it's meant to pick up that whispering and do a really good job of translating that into text. It also is good at using the AI to clean up your speech as you go. So you start writing an email with it and say, "Hey, you know, oh yes, let's meet at 3 p.m. on Thursday. Oh no, actually at 4 p.m." It'll just say, "Hey, let's meet at 4 p.m. on Thursday," right? That'll come out as the text. So if you kind of make changes as you go, it'll use the AI to clean that up and spit it out in a pretty formatted block of text and so forth. So I've mostly used it on my computer so far. I don't know—I have a PC. I hold down the Control and Windows buttons, which are the two bottom left corner buttons. So they're easy just to hit without even looking. When I hold those down, it starts recording.

32:57.82
Host
When I let go—and there's another button you can push to, you can let go the buttons and just talk for as long as you want. I think for six minutes actually is the cutoff. But then you let go, and then it pops up—it kind of generates quickly, formats it, and pops it up on your screen. And man, it is so much faster to write emails, to do prompts on AI, going back and forth with the model—just to speak it. And then they have an app for the phone, which I just put on my phone but I haven't used much yet, but I'm intending to try that out as well. But man, it does change how much you can accomplish without having that bottleneck of typing.

33:14.78
AROB
Mm-hmm.

33:36.03
AROB
That's really interesting. Yeah. So I could see that being really helpful from a fundraising standpoint as you're drafting out a, you know, an appeal letter or, you know, thank-you notes, or even, you know, for a pastor, maybe, um, you know, talking through a homily, right? You can like talk through it, change what you're going to say, go back, correct it, and it can—it can real-time kind of take out a lot of those edits and make a, you know, final draft that's cleaner.

34:10.10
Host
Definitely. I didn't really realize—maybe I did, but I didn't really realize how much time I spent responding to emails until I started kind of cutting it way down by using this. But I don't really trust, you know, the auto AI-generated email responses. Like there's always something that, you know, there's like a personal detail or something that is never quite in there. But if I can just hold a button and speak for 20 seconds instead of typing for three or four minutes, it's pretty nice.

34:28.60
AROB
Right.

34:34.67
Host
So I like that.

34:36.02
AROB
Fascinating.

34:37.11
Host
So hot tip.

34:37.14
AROB
Yeah, well, those are great. You got anything else?

34:41.78
Host
One more I'm going to throw out. I know we're just throwing them out, but I think it's fun to kind of hear some of these use cases.

34:43.09
AROB
Yeah.

34:45.86
Host
So Claude, which is made by a company called Anthropic, if you've heard of that—Claude is their model. They have something called skills for Claude, which is you kind of create a file. It's called a markdown file, which is a kind of a specific format. But you can actually have Claude generate this for you. Once you're going to work out a process and say, "Hey, here's this process we did. Turn it into a skill." And then you save that skill to your profile. And then you say, "Hey, next time you're going to do that thing, run this particular skill to do this process." And it remembers all the steps that you gave it.

35:14.76
AROB
Mm-hmm.

35:16.58
Host
So we use a project management tool called Asana. And I do a lot of planning out our communications and who's doing what, you know, we have this blog post coming out, who needs to write it, by what date, who's going to set up all these different pieces, right? And it just takes a bunch of time. We do it all in one kind of combined calendar for several of us. You can make separate projects in Asana and repeat them over and over, but then too many projects gets to be a pain too. So I had—I worked with Claude with my Whisper Flow. We went back and forth on, "Hey, help me set up—if I give you the spreadsheet of our plan for communications, set up all the tasks and when they need to be completed by so I can just import them straight into Asana." And it did. And that's a process that I usually sit down once a quarter and do for three or four hours. And I did about 20 minutes the other day after kind of figuring it out with Claude.

36:03.94
AROB
Fantastic.

36:07.10
Host
And I just tested it. I said, "Okay, turn this into one of your skills and save it." So I saved it and then tried it again. "Hey, now prepare this next month." And it spit it right out, and I was done, right?

36:18.17
Host
Right.

36:18.26
AROB
Fantastic.

36:18.49
Host
So you can do that with lots of different things, and they're adding more and more kind of integrations with Claude and different programs. You can communicate, you know—they had just announced recently an integration with Asana. One thing that I can't do yet is create that task for me in Asana, but I think that's coming. So right now it gives me a CSV file, and I upload that, which, you know, is an extra step.

36:36.58
AROB
That's crazy.

36:40.06
Host
But I think in a few months, I'm going to be able to say, "Hey, generate this next month of communications tasks," and it's just going to show up for me. And it's pretty cool where we're going with that.

36:48.66
AROB
Crazy.

36:52.02
AROB
Yeah, again, it's about helping you to be efficient in the sort of the prep or the legwork or the tasks that you would otherwise spend a lot of time doing, so that from a fundraising standpoint, you can be more either present to your donors, you can be more engaged in strategic conversations with leadership or with your team, and not be bogged down with all of the busy work that just comes with really any kind of work, but definitely with fundraising.

37:33.50
Host
Yeah, I think back to my fundraising days—as my time at Campus Ministry, as that came to a close, it was me working full-time as a development director. And we had one person doing communications 10 hours a week. And I was spending way too much of my time in the office doing all the tasks that need to be done and not enough time out with donors because it was just tough to fit in with all the other things. So when I left, they ended up hiring a second full-time person to work on the database and communications, right? So that the development director could be out with donors more. I think that's where this is heading—it's taking over more of those tasks in the office for us that we can be out face-to-face with the donors, building the actual relationships. That's where we want to be going. Or if you're a, you know, the director of a ministry, a pastor, and you need to build fundraising up to a point where you can hire somebody, these tools are going to make it much more doable to get to that point more quickly with the limited time that you have to give to fundraising.

38:31.42
AROB
I love it.

38:33.34
Host
Excellent.

38:35.56
AROB
Great. What else you got?

38:37.78
Host
Those are my hot tips for this week.

38:40.74
AROB
I think it's great.

38:41.17
Host
Yeah, we could—we can play a little game if you want to. Do you want to?

38:45.77
AROB
Let's do it.

38:46.71
Host
Okay, let's do it. This is just for this particular episode—I had AI generate a little game for me.

38:54.01
AROB
Okay.

38:54.39
Host
So these are all from ChatGPT.

38:56.73
AROB
Okay.

38:57.48
Host
We're going to play underrated, overrated, or properly rated. And I got a list of kind of AI tasks or things you might give it to do or might do with it. You're going to tell me whether that thing is overrated, underrated, or properly rated—whatever, however you perceive that to be, if that makes sense.

39:11.90
AROB
Perfect.

39:16.01
AROB
Great.

39:16.98
Host
We'll make this just an all AI, all day episode. All right, great. So are you ready?

39:24.21
AROB
Yes, sir.

39:24.92
Host
The first one is custom GPT is trained on your own documents.

39:30.01
AROB
Custom GPT. Uh, definite—wait, what are my options?

39:34.24
Host
Overrated, underrated, properly rated.

39:37.98
AROB
I think that's underrated. I want to train—definitely train, use custom, use training materials, use content to train custom GPTs for sure.

39:39.83
Host
Yeah.

39:51.58
Host
Love it. Next one: AI meeting summaries instead of manual notes, such as on Zoom calls.

39:59.00
AROB
Hmm, I use—we use Otter AI for meeting notes. And it's awesome. It's really, really helpful. So underrated.

40:08.89
Host
It is nice when you miss a meeting to be able to just catch up. Sometimes you see—I remember there was one meeting that I missed, and I looked, and they were talking about, "Oh, I found this great deal on a car over at this place." And I go, "Did anybody work on this meeting? Wait a second. I guess I should have been there to keep them on task, huh?"

40:21.39
Host
I guess I should have been there to keep them on task, huh?

40:23.35
AROB
Have I told you my story that got me in really big trouble with the meeting notes?

40:30.13
Host
No, do I want to hear this?

40:33.27
AROB
So I was doing a client call. It was—when was it? It was around Thanksgiving, I think. No, around Labor Day. And so we started the call talking about what everybody did for Labor Day. And we had talked—my family and I had talked about doing hot dogs on Labor Day and just didn't get around to it. And so we're talking on the call, and in a very joking way, I said, "You know, I was really bummed that we didn't get hot dogs for Labor Day. I love a good hot dog, and it's, you know, great—you need good occasions to eat hot dogs and everything."

41:16.34
AROB
And so they're like, "Yeah, okay." So again, not angry, not, you know, no tone to it. Just joking around, not getting hot dogs. Well, this particular meeting notes I had to share with Cheryl, my wife, because she was doing some work for this client. And so she pulled it up, and the summary note says, "Andrew is disappointed with meal planning for Labor Day." And I was like—and I didn't catch that before I sent it to her.

41:39.64
Host
I remember this now.

41:46.81
AROB
Otherwise, I don't know that I would have sent it to her, or definitely would have edited it. And so she was like, "Oh, I'm happy with my meal planning, huh? Well, maybe you should do the meal planning then." And I was like, "What have I done? AI!"

42:01.82
AROB
AI!

42:02.84
Host
Yeah.

42:02.92
AROB
Yeah.

42:04.12
Host
And—

42:04.86
AROB
Yeah, so there you go.

42:05.62
Host
And that's how Andrew ended up on the couch.

42:07.42
AROB
And that's how Andrew ended up on the couch making all of his own meals for the next month.

42:08.09
Host
Oh.

42:12.02
AROB
Yes.

42:13.05
Host
Respect your AI overlords or they will—

42:14.76
AROB
Yes. That's good.

42:17.72
Host
Oh, oh my goodness. Yeah, I remember that. That was great.

42:20.95
AROB
That's good.

42:21.08
Host
All right. The next one: overrated, underrated, or properly rated. Using AI to modify or shorten or improve text.

42:31.51
AROB
I think properly rated. I have it clean up—I told that story about, you know, the report I was working on. So I think that that's a good use, and I think enough people use it for that that it's properly rated.

42:47.11
Host
Yeah, I love it for shortening text. When I write something and say, "Hey, make this 30% shorter without losing context." It's so good.

42:52.74
AROB
Yeah.

42:53.01
Host
I love that.

42:53.42
AROB
Yeah, yeah.

42:53.97
Host
Next one: AI-generated social media posts.

42:58.70
AROB
Hmm.

43:04.36
AROB
Going overrated, not because I necessarily think that it's a bad use of AI, but I think a lot of people think, "I don't have to write my own content for social media. I'll use AI to do it." And then you just end up with kind of like very generic, bland—everybody's putting up the same stuff on social media. So I think that there are ways to use it, but I think that it gets overused by people who don't feel like—you have to then use a human touch to make it better.

43:28.80
Host
Yeah.

43:39.38
AROB
And that's my overrated point.

43:40.47
Host
Yeah, people are using it for more instead of using it to make better content.

43:44.73
AROB
There you go.

43:45.03
Host
I think that's the distinction there.

43:45.53
AROB
Yeah. Yeah.

43:48.41
Host
All right. Well, we talked about this one. Where do you fall on talking to AI instead of typing your prompts?

43:56.23
AROB
Ooh, I'm going to go with underrated because I don't do it enough, but it sounds like from your experience, there's a lot of potential.

44:04.92
Host
Loving it. Yep. All right. How about AI as a thinking partner, quote-unquote?

44:11.82
AROB
Hmm. I'd say properly rated. Again, kind of—people use it. I think this is a big use for AI. It's appropriate. You know, you can kind of process through, talk through different things. And I think that people are using it that way a lot. So there you go.

44:30.97
Host
Yeah, I love when we're writing a course lesson or something like that and saying, "Hey, what am I missing here? What concept isn't fully developed?" And it gives you some ideas, and sometimes four out of five of them are trash, but one of them might be like, "Ah, that's a really good point."

44:31.00
AROB
There you go.

44:44.68
AROB
Right.

44:44.79
Host
So, all right, let's see.

44:44.88
AROB
Yeah. Hmm.

44:49.56
Host
Where do you fall on AI outlining before you write anything yourself?

44:57.59
AROB
I would say also properly rated because I think that's a good use. Because then, you know, you're applying that human check to make sure, "Okay, yes, this outline is good," or—like I find I do that all the time, right? I say, "Hey, help me start an outline." And then I go through it and I'm like, "Oh, yeah, this is good, but this is not," or "This is developed, or we need to add more here." So I think it's a good use. And I think that a lot of people use it for that. So it's properly rated. Am I playing this game right?

45:29.69
AROB
Am I playing this game right?

45:34.20
Host
Yeah, I think so.

45:34.78
AROB
Okay, good.

45:35.80
Host
Yeah, we're just kind of doing a speed round of uses, right? So I think it's helpful.

45:38.97
AROB
Yeah, that's good.

45:41.05
Host
Good. Let's—how do you feel about using AI tools inside tools you already have, such as email or CRM?

45:51.19
AROB
That's a good question. My experience is I'm not using it a lot. So maybe that means that I'm underrating it. Maybe it means that the people that are using it are too reliant on it and it's overrated. Or maybe it means my experience is unique to me or is representative of—so it's properly rated. I don't know. That's a good question. I don't personally find that I use those features a lot, like in Gmail, like in HubSpot and things like that. So maybe if I used it more, I would have more of an opinion.

46:33.30
Host
Yeah, it's tough to say if those tools just aren't there yet or if they're just not helpful for that context, but I definitely find myself using the outside tools a lot more frequently.

46:38.15
AROB
Right.

46:41.98
AROB
Yeah.

46:42.52
Host
All right, and then last one: using AI for idea generation volume—add volume, if that makes sense.

46:51.93
AROB
Idea generation volume. So like, "I need 10 ideas about this."

46:57.53
Host
"I need a subject line for this email. Give me 30 options."

47:00.18
AROB
Yeah. Yeah, I'd say that's properly rated as well. It feels like a lot of people that I know use AI for that. It seems it's a very appropriate use. And so I would go properly rated.

47:10.74
AROB
And so I would go properly rated.

47:14.10
Host
Excellent. That was my last one that I had.

47:16.79
AROB
Well, good.

47:17.11
Host
Thanks to ChatGPT for that game.

47:17.31
AROB
There you go.

47:20.12
AROB
Right, how do you feel about ChatGPT creating podcast games?

47:28.41
Host
I do occasionally ask it when we're doing, say, a blind ranking. "What are the most popular restaurants in the country?" That type of thing. And when I don't know, and I want to give you an option that's only available in the Midwest and not in Texas, right?

47:43.86
Host
Um—

47:43.99
AROB
Yeah. Yeah.

47:45.30
Host
But in terms of generating, I had to—I had to get a list of 30 options to get it down to that, whatever that was, 10 or 12. So, yeah, yeah. I'd like to think I'm a little bit better at it than chat, but maybe not. The listeners decide.

47:57.73
Host
I'd like to think I'm a little bit better at it than chat, but maybe not. The listeners decide.

48:03.79
AROB
I don't know. I think the games are great.

48:04.88
Host
We'll see. Yeah, that was fine.

48:09.15
AROB
Good.

48:09.37
Host
Excellent. Good. Well, that's what we got for today. We'll wrap it up, leave it there, and six months from now, we'll come back and talk about how everything in AI is different than what we talked about today.

48:17.88
AROB
Exactly.

48:19.90
Host
It's amazing how fast it's moving. So just try to keep up.

48:22.62
AROB
Yeah, for sure.

48:25.98
AROB
Right, cool. Well, thanks. All right, sounds good.

48:27.42
Host
All right, we'll see you again in a couple weeks.

48:30.39
AROB
Bye.