B (3:27)

So at the beginning of my career, before graduate school, I did an AmeriCorps program and I was at a nonprofit in Chicago. And it was in the earlier days of social media when there was Google and a Facebook page actually mattered and.

A (3:44)

Yeah, totally forgot about that.

B (3:47)

Remember Klot scores?

A (3:48)

Oh my gosh, I do remember cloud scores. I would check them obsessively.

B (3:52)

So I was an AmeriCorps. I had an AmeriCorps position and they just kind of handed me their social. Because that's what you did back then, right?

A (4:00)

Right.

B (4:01)

Gave it to the young person on staff. And I found I really liked it. I also happened to be creative. I like to. To paint and write. So I think the writing plus the digital content creation side of things suited me. And then ever since then, I had content related roles at nonprofits.

A (4:21)

Oh yeah, Cloud scores, Google. I don't even know. There's so many that have sort of gone by the wayside and so much transition and so many changes in the digital marketing space. I do love what you said. It was like sort of hand the young person the keys to the social media and they'll figure it out. And as we know now, I mean, there really is so much strategy involved and real intention. You know, when you're creating a digital marketing strategy, there's so much talk around AI right now. I've had other people on the podcast talking about AI and I recently, just because I don't really pat myself on the back very much, but I was recently a contributor to the recently published book AI for Nonprofits and I was talking about social listening. But why do you think that AI is especially relevant for small nonprofits sort of in this moment?

B (5:19)

Yeah, I love that. Congratulations. I think social listening is.

A (5:23)

Thank you.

B (5:24)

I define a small organization. To preface, I define it as small staff. So really it's. Most nonprofits have staff. Staff members who are stretched thin and doing a lot of different jobs.

A (5:36)

Yes, yes.

B (5:37)

This marketing department of one multiple times. And I've worked with nonprofits who consider me their marketing department. So I really get it. And I think AI and AI and automation, it's shines at helping people do more with less, which is exactly what the nonprofit sector needs. So once AI and LLMs became more ubiquitous. I knew I needed to learn it because as an outsider working with nonprofits, I consider part of my job to learn things they don't have the time to learn and then bring it back to them. So, yes, right, like ads and automation. AI like any kind of tool because as like we said, with clout, they come and go. And so you don't have the time when you're trying to expand, execute a mission to learn about things that might stick around, might not. So I consider that part of my job.

A (6:31)

I love that looking at that. So in my work, a lot of my work with clients, I do the same thing. I do an audit and I kind of have that bigger picture. Look at some of their strategies, their website, email and social media, because they don't have time to stay on top of every single trend and what is working and what are other nonprofits doing, doing? And like, what are some case studies in the sector? And are Instagram reels really popular now? And like, what's going on with TikTok? So that I love that. I really see that, yeah, as our job to be that sort of filter and that lens for the trends and for the tech. So I know that a lot of nonprofits are really overwhelmed by the tech, especially by AI. What would you say to someone like, what's your advice to give to a nonprofit leader that's maybe hesitant to dive into it?

B (7:27)

First of all, I get that. I think hesitation is fair. It's a new tool and I think you can have hesitation and excitement and decide what works for you over time. I would say start using it even just in your personal life to start to get confident, comfortable. I've used it to build my packing lists, ask AI to find what I missed. I've used it to plan vacations. I've used it to menu plan. I think just using it helps you get familiar with like what it can do and see and brainstorm other ways to use it. So that's one thing I would start with. Just one tool, one of the main players is more than sufficient. And you can just start with the free version until you think you're going to use it more.

A (8:14)

I completely agree with you. I think one of my favorite examples, I'm not sure if I have shared this on the podcast, so if you've already heard this listeners, you're going to hear it again. But just asking questions. So my sister brought over some chicken that she had purchased at like the farmer's market and she wanted to air fry it. So instead of googling a Recipe. I asked Chachi pt. I said, here's how much chicken, here's the weight. And then I took a picture of my air fryer and I said, how can we air fry it so it'll be delicious? And it gave us specific instructions for my specific air fryer. It was amazing. So it's really like the intern and the assistant that you don't have. I love the idea of a packing list. I love the idea of planning trips. I know a lot of my friends have used it to plan trips because you can get very, very specific like we don't want to walk this much or you know, we have know an elderly parent with us. What can we do if we're in Italy for xyz? So I think using it in your personal life, diving into it is going to make you a lot more comfortable. You talk specifically about creating donor Personas with AI. This is not something that I've heard of. I think this is really interesting. Can you walk us through what that is and how you approach that?

B (9:35)

So donor Personas and audience research in general are one of my personal favorite use cases. And the way I do it is the way I do a lot of my workflows with AI, which is that I like to start with a bot that helps me craft my prompts. So for basic tasks you don't need to do this. LLMs have come a long way. By LLM I mean large language model which includes things like ChatGPT and they're meant to talk to you like a person, but we know they're not thinking. It's pattern detection, it's probabilities, but I like to use AI. I start with a bot that helps me build a prompt because part of AI's work is to me to help me lift the mental load. So while I can memorize the basics of a good prompt, I don't want to have to remember it every time. So I'll start with a prompt and I'll start with developing a prompt that helps me develop Personas based on my use case. But to give you like to take a step back and give an overview of this specific case, what I'll do is I use an AI notetaker in my day to day work. I let people know I'm using it. I only use it to take notes for my personal use. There are audio only ones available as well if you don't want video ones. But a lot of them have free versions like Fathom and Fireflies. So it's really important to keep those going and you can use the free version of those. Then I use something like Zapier in the free version. You can do this to have an automation that saves all my transcripts to for example a Google Drive folder. So I just keep them all stored there and I don't have to think about it then what I do, it's a multi step process.

A (11:20)

Sure.

B (11:20)

So follow with me here. You're recording with notes, you're saving your files somewhere. You can take a whole, whole bunch of transcripts with your supporters and attach it to an AI conversation. And now you're going to ask it to detect patterns. So you can ask it, given all these conversations, what can you tell me about my donor's hopes, fears, aspirations, goals, values? Right, all the things you need to know that motivate people. That's a brief overview. One of my favorite tools for doing this, the end result is Google has a product that's also free called NotebookLM. You can open a notebook and each notebook in the free version you can upload up to 50 documents and then you use the AI chat to like to interact with your source material. And what's really beneficial about this is that you avoid something called hallucination where AI makes things things up to a square of AI making things up. And I like NotebookLM because it sticks to what you provide and it brings things up for you and will find the pattern. So that's where I start with my audience research. You can also use Deep Research mode in any given AI tool. For example, Perplexity gives you three free Deep Research searches a day and ask it for your primary donor that you're building a Persona for. So let's say you're building a Persona for. I'll take myself a woman in her mid-30s, lives in the state of New Jersey. Right. So you ask it to deep research this Persona, including their values, including their hopes and fears, including everything you might want to know about what motivates person. I know this is getting confusing, but.

A (13:15)

No, I love this.

B (13:16)

Not least, I go to a new chat and I say I attach my synthesis of donor conversations and I attach my deep research and then I ask it to build a Persona so you can build. I have prompts that help me build Personas so I'm happy to share them after, you know, in the show notes. But you can also just ask AI hey, take all of this and build a Persona for me. You can ask AI to generate it into a document, make a one page document and you can literally copy paste it.

A (13:52)

And I think it's Important for listeners to know. I mean, a lot of my listeners, I think, are pretty tech savvy. But for people that don't know, I mean, this is all confidential. This is all you. This is your database, you know, or this is your own private chat that you are having, and these are your Personas. But it's helpful to have that information that you can then explain, maybe in a presentation or talk to your board, create marketing materials, create fundraising materials around. The other thing I want to pull out of this is what I cannot do, is have those conversations with donors. Right? That is where the nonprofit sector. That is where development professionals and nonprofit leaders get. That is where we shine. That is where we will never be replaced, is making those personal connections and having those conversations. But once you have those conversations, it's almost like a brainstorming partner thinking through, like, what are some questions I should be asking my donor? What are some ways I can improve my conversations? It's almost like having another development director to kind of brainstorm with you. I really like that. So do you use AI in other ways for fundraising, like grants, annual appeals? Can you talk about how you might use AI for those kind of materials?

B (15:13)

So, to be honest, I use AI in almost everything I do. It's my co pilot. That's always.

A (15:18)

Yeah, it's like a co pilot. Yes.

B (15:20)

Yeah. And I love what you said. That it's not. It shouldn't replace people. I think of it as creating space and time to do what people are best at. So if you aren't bogged down by the middle 80% of tasks that AI can do, then you have more time to build relationships, to say thank you, to make phone calls, to have conversations, to go out for coffee. Because we're already trying to do too much with too little time. So I don't think of this at all as a replacement. It's really just to do things on a budget and faster. Because we have to.

A (15:58)

Exactly. I love that. Do more to do more with less.

B (16:01)

Do more with less. So, for example, for grants.

A (16:04)

Yeah, I love examples.

B (16:05)

What I would do is if you have successful grants applications, right. That have led to getting money, you could attach a bunch of those to a conversation. And again, we're looking for patterns. Ask AI. What are the commonalities between these grant applications? What made them strong applications? What do they have in common? Right. You want to pull those things out, and so you're not starting. You're not doing guesswork every time you start an application. And then you can even attach to a conversation those Patterns, the basic data and stories you've gathered to include in your application, the application guidelines from the provider. And you don't have to ask it to write it for you. You could ask it for an outline, for example. You could say, how should I do this? Given the instructions, given what's worked for us in the past, can you help me outline what I should put into a draft? It helps with the blank page problem, and that's a lot of the work.

A (17:07)

No, I love that. And I really, I also think what I used to do when I was a grant writer millions of years ago, I'd have to adapt the grant to the specific format or the specific template that the grant funder wanted me to use. And it was my grant that I wrote. But I spent so much time doing that kind of busy work. And I do think like, ChatGPT could really help either trim it down or format it in the way that they want it formatted or, you know, pull out the things that, that, that the specific funder wants versus, like what another specific funder wants. So really just using it to, like you said, take the, the busy work out of work and help us really focus on what we're good at and like what we should be doing. Can you share maybe some of your favorite workflows or prompts that you've been working on or that you've taught recently?

B (18:04)

So I have a short webinar and prompt prompt bank available on my website. But part of what I teach is that there's a few elements that makes a really strong. The first is I heard it from another thought leader and I'm forgetting their name at the moment. Who says you need to tell AI who it is? Which is why you start with you are a grant writer, you are an audience researcher, you are a copywriter. Right. So you start with giving it context for who it is and then you say who you are and what you need, like I'm your client or I'm your patient. Right. Let's say you shouldn't do it for medical information, but let's say for a nutritionist. Right, You're a nutritionist. I'm a woman in my 30s and I'm a patient. Or you're an audience researcher and I'm a nonprofit executive director and I need you to research this up and coming demographic of supporters. Then you give it context so you.

A (19:04)

Context so important.

B (19:06)

Yeah. Tell it more about you or tell it more about your organization. You can attach documentation to a conversation. So I really love doing that as a shortcut so like I've attached our brand guide, I've attached a slide deck that tells you everything you need to know about us. Right. That's all you need. Then you can ask it for specific kinds of output. So I like to ask for tables, for example, because it'll fill it out really interestingly in the intersection of rows and columns. Or ask it for something long and detailed. Ask it for short bullet points. Tell it specifically what you want. Tell it what its goal is. Like, here's what I need the end result to be.

A (19:48)

Yes.

B (19:48)

Tell it to ask you questions if it needs more information. You never have to answer all the questions. Sometimes I've seen a few questions and I'm like, I don't want to answer that. And I'm like, just keep going. And it keeps going. But sometimes it helps you give it more information to get a better result. Last but not least, you want to ask it for reasoning and to share its reasoning. Now you could say, think step by step, think logically. Tell me why any sort of variation of that? Because it actually helps you get better results. It's not even about what you read. I read this fantastic book by someone who helped develop AI at Google and he says it helps it to actually give you a better result. I think it forces it almost to slow down.

A (20:33)

Yeah.

B (20:34)

Things more methodically. So those are the basics of a good prompt in general. And I have a bot that implements all of that into any prop I ask for, for example, so I don't have to write it myself every time. That's kind of the background. And then from there I really love to use AI box. I mean I use it for so much. But anything pattern recognition related is especially, especially great. Which is why I think people have found that it's great for making up silly songs or poems. Right. You'll always. You'll go to an event and someone get. Gets up and they wrote a speech or they wrote a poem or they wrote a funny song and they used it. They did it using AI because AI, that was our. So that's part of what makes it so great for research, but also for writing like you. You can get AI to write like you. That's. That's another fantastic use case.

A (21:29)

Exactly. The. It's all about the prompts and the prompts for people that are saying, well, maybe what is a prompt? It really is just what you're putting in, what you're typing in, what you're asking it to do. So it's sort of like a Google search. If you are doing an incredibly vague Google search, then you're not going to get the results that you want. But the more specific that you get, the better the results are. So it's the same with an LLM like Claude or with ChatGPT. And the analogy I always use is, it's like when I was working with interns and I would be training interns. And you have to be very specific with context, with what you want, with what the goal is. You can't just say write a press release, because if you just say write a press release, you could get any number of things. You get a three page press release, or you could get five bullet points. I mean, you have to be very specific. What I really enjoy most about using ChatGPT to brainstorm, you know, using it to do research, is that the more that it learns about you and what you want, then the more it can create sort of a really great result for you. So I just think that I just look, lean on it a lot as a research tool and sort of an editing and copywriting tool because it's just, it's just something, you know, once it sort of gets to know your style and your audience and your goals and what you want, it works really well. Are there any ethical considerations that we should be keeping in mind? I think this is top of mind for a lot of nonprofits. I think especially around maybe image creation and video creation. We're seeing a lot of images created. I'm thinking of, I don't know if you saw the most recent south park episode, but there it's an almost entirely AI generated images of the president. And I know they're using it for comedic purposes, but it's incredibly realistic. And I think that really worries a lot of nonprofits where our integrity is so imperative, our reputation is so important. Trust is also decimated across institutions. Like how can we use these tools but also continue to maintain our integrity and, and consider maybe some of the ethical implications.

B (23:53)

Yeah, I think that's a really great question. So baseline thing I always think of is actually privacy, because once you're digging into things and you're putting in data, I would recommend most paid tools are about $20 a month. And when you pay for a tool in the settings, you can exclude anything you do from teaching the model and keep it private. So depending on what you're going to be using AI for, that's a simple way to address that concern. But of course, also, if you're a doctor, don't start uploading patient charts, like anonymize things Use a little common sense for the larger ethical concerns. I completely agree, and some of it is a personal question. So, for example, I'm in the middle of developing an automated content flywheel and I do this custom for clients where we develop a system using AI agents and a base like Airtable to help them ideate and develop content consistently to help lift the load. Now there are ways to do this where you can use a video platform to have an avatar that looks exactly like you and says a script. I've decided those are not happening in my workflows. But someone could, right? I'm not going to name specific tools, but there are very popular tools out there. Or instead of generating an image, which I think there are ethical and non ethical ways to do it, I've seen people generate like things that look like a doodle or a watercolor, right? When you're not imitating someone else's likeness or imitating someone else's art and you're saying, I want an icon of a duck that looks like a doodle, right? Like, I don't think that's actually ripping someone off, that that's a fair use. But creating a portrait of someone else's exact likeness and then using it for things that create could earn money. That's unethical to me. But in automated workflows, for example, you could have templates that you use and then you put in your own image and you put in your own text and it could be free image you get from unsplash, it could be an image you took at an event, but you could use the systems in ways that feel aligned with, with your values and you don't. This is why earlier in the episode we were talking about hesitation. And I think, I think it's still important to listen to your inner voice. And so for me, I don't, I don't want videos of me that are not me. I don't want comments that sound like me but aren't me. And I think, I think it's important to keep your own compass even, even with AI.

A (26:35)

That's so important. And when we're teaching technology, I always say to nonprofits, you know your organization, you know your gut reaction, you know your mission, you know what is aligned, what is not aligned. And just always go with your gut and don't do something that you are not comfortable with. Like don't put something out to the public that you are just not comfortable with. People think that they have to be doing these things and they have to be, you know, constantly posting this content and how are they going to keep up with all of the different brands and celebrities and creators. But I really hope that nonprofits will trust their instincts like you said, and understand that there are ways to use it. But it's sort of like when Google Images came around. You know, there's ethical ways to use Google Images. There's ethical ways to use images that you find on the Internet. There's ethical ways to source different information and to attribute information to people. So just the same rules that you would use for anything like that, I would transfer over to AI.

B (27:39)

So.

A (27:39)

So how do you recommend that sort of we all stay current without getting really overwhelmed by every new tool, every new update. How can we really just sort of stay a little bit sharp and a little bit in the know without getting completely swamped?

B (27:57)

I think just making sure to use it really would be my number one, my number one recommendation. Making sure repeatedly to use it, even just a little bit and don't just let it go and not go back. Because this field, I think is going to change very rapidly and being completely out of touch with it will probably do you a disservice. But I don't think it takes a lot. I think just using it every once in a while, if it's been more than a week, try to use it for something. That would be the first thing. I like to keep up with the news on it, but that's because that's part of what I do. And I don't think you have to. I think you should be using one of the big players. There's a number of them. Gemini chatgpt, Claude Perplexity. There's quite a few out there. I don't like the aftermarket tools like the ones that tell you we're going to give you access to 20 different AI platforms, you know, $15 a month. Those I would shy away from in most cases. I'm sure there are good use cases for them. First of all is overkill. Second of all, a lot of times the way they're saving money is by shortening the context of it knows. And so what that means is the memory of your conversation is shorter and that makes it harder to iterate and refine and go back and forth in the conversation for as long as you might want to. So really you don't have to reinvent the wheel. Just use one of the bigger platforms and just continue to experiment.

A (29:24)

Yep, I love it. If someone's listening right now, they want to get started. Do you have a favorite tool? I mean, for me I love ChatGPT. I know we talk about it all the time on my podcast. It's just the one that I just kind of got started with and have not really moved on from it. What is the first simple step that you recommend?

B (29:44)

So I've bounced around a lot between tools because I find tools fun.

A (29:49)

Yes.

B (29:51)

Personally for marketing tasks. I really like Claude. I do also use ChatGPT because that's what most of my clients use and that's what's most people use. Gemini has also really come a long way. Gemini is really great for voice mode. Actually. I think voice mode is an underutilized way of getting familiar with AI. There's like a little button and it'll talk back to you and in the beginning of the conversation for anything, you just use it for things you want to learn about and then tell it to keep asking you questions because that keeps the conversation going. So for example, I've used voice mode to help define brand voices, right?

A (30:34)

Oh yeah.

B (30:34)

Or your personal brand voice and say, keep asking me questions until you get a sense of what my voice is like and then document it for me. So those are some. I think that answers the question. I tangented a little there.

A (30:48)

I love the idea of having it just ask questions, just saying, you know, you know, I would even type into ChatGPT. I am a non profit executive director. I'm really interested in using your tool, you know, to help with my marketing and my fundraising. What are some questions you might have for me and how can I get started? I would really just put that as a prompt because you'd be surprised at some of the insights, some of the ideas that it's going to give you. So yeah, just kind of like, you know, dip your toes and get started. I think most people are using these tools maybe in their personal life, but I love the idea of donor Personas. I think that is so interesting and using it as that sort of high level analysis tool for your fundraising, really understanding your donors better because just always thinking about the goal at the end of the day. We want to understand our donors, we want to accomplish our mission, we want to connect with our community, we want to build awareness. Like what are our main, main, main goals and how can all of these tools, social media, your website, everything, how can we use this technology to accomplish our goals? So focusing on the goal at the.

B (31:57)

End of the day and speaking of that goal, because we're trying to make time to connect with people and what we're used case I forgot to mention is that as AI attach A big chunky piece of content to AI, such as sharing a video you made a blog post, post an annual report, ask it to help you repurpose it. Because AI is really good at repurposing your content and say these are the channels we have. Here's what I need. And it will really help you do so much more with what you already have. And that's also a big part of saving time. Love.

A (32:33)

I love that. Yes. Your gala. Here's our big gala program. Here's all the content we've created. Like how else can we get the word out and how can we repurpose this giant document into smaller bite sized tidbits for social media for our email? I think that's a just a fantastic use case. I use it for that as well. So where can people connect with you faggy and learn more about you and your company?

B (33:00)

Sure. So I'm available a lot of the time on LinkedIn. I got very lucky with my vanity URL, so it's just/fagy. And my website is smallorg comms.com and if you go to smallorgcomms.com nation I've included some of the quick links to some of the GPTs we've talked about today, such as the one to create your prompts or the one to build your Personas or the one to repurpose your content. And you could ask me for others if there are some missing that aren't there, but I've tried to corral them all into one place.

A (33:38)

Okay, definitely go smallor comms.com nation or.

B (33:45)

Comms it's not I I hope to change it one day. S M A L L O R G C O M M S okay.

A (33:53)

I'll put that in the show notes for everyone so that you can make sure that you get those resources. Thank you so much for being on the podcast. This was really wonderful, really tactical, actionable and practical information for my listeners.

B (34:08)

So appreciate it. Thank you so much for having me.

A (34:17)

Well, hey there. I wanted to say thank you for tuning into my show and for listening all the way to the end. If you really enjoyed today's conversation, make sure to subscribe to the show in your favorite podcast app and you'll get new episodes downloaded as soon as they come out. I would love if you left me a rating or a review because this tells other people that my podcast is worth listening to and then me and my guests can reach even more earbuds and create even more impact. So that's pretty much it. I'll be back soon with a brand new episode. But until then, you can find me on Instagram. Juliacampbell77 Keep Keep changing the world, you nonprofit unicorn.