How Nonprofits Can Use AI Effectively

Howard Adam Levy on the Nonprofit MBA Podcast

What You Will Learn from This Episode

Transcript

The nonprofit MBA purpose is to provide new business insights and fresh creative ideas for executive directors and their teams that will help them improve their organization. Here is your host, Stephen Halastnik. Hello everyone, my name is Stephen Halastnik. 

I will be your host for today’s nonprofit MBA podcast as I have been for the last six years. For those of you who don’t know me, I am the co-founder of Financing Solutions and for the last 15 years we have been the leading, the largest provider of lines of credit for small nonprofits in the United States. It’s a great product to have. 

If you think you’re unique in having ups and downs in your cash flow, you’re not, it happens to every single, not just a nonprofit, but every single business. So lines of credit are specifically geared for that type of issue and if you’re interested in learning more, please visit our website at non-profit MBApodcast.com. On this sunny but snowy day, for those of you who don’t know this, I’m based out of New Jersey. Our guest today is actually based out of New York City, which is only 50 minutes away from me. 

I live and do the podcast from the mountain and lake region of New Jersey. It’s a beautiful, beautiful area. Most people are like, wow, New Jersey’s got lakes and mountains, yep. 

And it’s really a beautiful area and I go to New York all the time. So today I’m very excited to be speaking with our guest, Howard Levy from Red Rooster Group who is a returning guest. He spoke on our podcast once before and today we’re going to be talking about how nonprofits can use AI effectively and just a little bit of a background on Howard. 

As president of Red Rooster Group, Howard Levy has helped hundreds of nonprofit organizations wake up their brands, raise more money and reach a higher level of success. He specializes in harnessing behavioral psychology, effective storytelling, strategies and action design principles to motivate people to donate and take action. For the past two years, he’s been exploring how nonprofits can harness the power of AI or visual intelligence. 

He’s here to share some of his insights and inspiration about that subject. Howard, welcome back again to the nonprofit MBA podcast. Glad to be here, Stephen. 

So a great topic, you know, I know you do your own podcast as well and, you know, once you’ve done 300 podcasts and I’m sure you’re used to this too, you know, not all subject matters are like interesting, you know, but AI, you know, I think everybody is curious and I’m sure this is going to get a lot of downloads. So, tell us a little bit about first, let’s just do a definition of AI and what is AI? Yeah, sure. And when you say curious, I think it’s between excited about the opportunities and scared as hell about the implications and probably where a lot of groups talking to a lot of people and that’s kind of where the mood is, the marketing people are very

excited, but also like how do we evolve our business to help our clients and nonprofits, like how do we, you know, tap into this so that we can do two more, you know, that businesses have, you know, a lot of these AI tools at their disposal and, you know, how do we take advantage of that? So, yeah, and I think, you know, at the end of the day too is, you know, it’s coming. 

So, I mean, it’s just coming, you know, it’s here, it’s coming, there’s a lot, you know, really on an AI and honestly, I think in our lifetime Howard, AI is going to be about productivity, right? It’s going to be just like when computers first came out, you know, it’s going to be, and those who take advantage of that are going to be better organizations because of it. So, hopefully we can give some people a little bit more insight as to how they can get more productive using AI. I certainly have noticed it myself. 

So go on, go on with the definition. Yeah, yeah, but there was an interesting comment I heard us like this is really interesting that AI is not going to take your job. Someone who knows how to use AI is going to take your job. 

There you go. Yeah. I mean, it’s the same thing like you’re saying, you know, with when the desktop computers came out or social media, it’s like, you know, it’s harnessing those technologies that are at your disposal. 

So, AI, the different kinds of AI, I’m going to talk about LLM’s large language models, which essentially are programmed to understand human language, interpret, and generate information based on the large language model as data set. So essentially, chat GPT and Bard and Gemini and Claude and these other AI tools are scraping the internet or they’re fed, I should say, you know, information from various sources so they could understand what the world is like. They could understand information and they can understand how humans talk to the world and how to reflect like human language. 

So there are great tools for this generative AI, which is what we’ll focus on today, which is like generating information, generating content or generating images. And so we’ll talk about their capacity to do that. But there’s a lot of other kinds of AI use for other things, like data analytics. 

And you know, these things have been in use for a long time. I mean, Amazon is one of the pioneers you see recommended books for you or Netflix recommended movies, right? Those are all AI, you know, algorithms, they’re learning based on data that’s fed in and they’re able to interpret that and then, you know, perform some function based on what their program, you know, to do. So when it comes to nonprofits or other organizations, it’s like, how do you get donors and how do you target them? How do you understand who a good donor would be based on, you know, different analytics? And then how can you target them appropriately? How can you craft messages that are mapped to those needs? So those are all things that are either in the works or coming. 

And I think that there’s like a gap between what AI tools are available and large companies that have big budgets and big data sets. You know, if you look at, you know, large companies, you

know, whether, you know, target or American Express or, you know, companies that are scraping millions of data points on various sources can tell a lot about you or other consumers based on all the data points and then they can target you based on that information. Hey, you know, hey, Steven, you bought XYZ, you know, in store through your credit card, through magazine subscriptions, through donations to nonprofits, you know, through all these different interactions you’ve had that they’re gathering on you. 

And now that you know, you know, what you may, what you may like, and they can use that information to target you and of course, sell to other advertisers. So, but nonprofits don’t have that capacity yet. A lot of the tools being built, sales force or HubSpot, you know, a lot of the donations systems are building in those tools to help you analyze data, but you have to have the data. 

If you have a small, you know, many organizations, they don’t have a lot of donors and they don’t have a lot of information on them. They know maybe they’re donor history, but they don’t really have a lot of other kinds of information about them, about their preferences and their interests outside of the nonprofit. So, you know, how do you use that? So, how do you use AI to target them effectively and create more effective messages that are going to resonate and drive them to donate? So, I’ve been exploring that, how to do that. 

It’s not easy when you don’t have the information, but there are ways to do that. And so, that’s what we’ll be talking about a little bit today. Yeah. 

And what I’d really like to do for our listeners too, I think the key with AI is to stay away from the technical aspect of it and just think of the application end of it. And, you know, I think a lot of times, you either can think of the application itself on that you want to automate and I want to give you an example and start in that direction or hear stories of what other people are using AI for because just think of AI as an application. They’re thinking of, you know, a technology, let’s say that. 

Just like when we were all started using CRM, customer relationship management, right? It was, you know, that was like, okay, we have all these things and maybe an Excel spreadsheet or a hard copy and files, you know, how can we get better, quicker access to that? And someone said, hey, there’s this thing called CRM, customer relationship management software. And then, you know, act came out or gold mine came out and those are the precursors. So, like, so what I hope to do today is Howard is for you to give us an idea of what nonprofits the application really is. 

And, you know, that would be some good insight as well. Is that possible? I’m not sure I understand your question. We use chat GPT a lot. 

The paid subscription, but the free one does work as well. So, I was going to talk about how you can use chat to do different things. Like, for example, a common use for nonprofits is writing a fundraising letter, right? And so, you can just ask it, hey, write a fundraising letter for me, andcyou’ll get a fundraising letter that’s kind of generic. 

It’s based on all the templates that it scrapes and says, here’s what a good fundraising letter should be. But it doesn’t have any further insight about your organization. It actually does know what your organization does, but it doesn’t have any further specificity. 

And so, I was going to talk about how to elicit better responses from that, just as, like, one narrow, you know, example. Well, that’s an application, and that’s exactly what I’m talking about. You know, like, I’ll give you one that I haven’t done yet. 

I’m about to do it. You know, these podcasts that we do, after we’re finished with them, they get written up into a 1500 word article, and then that gets posted to the website. And I have somebody who does that for me, you know, someone who writes it up and posts it. 

And so, there is an AI application out there. It’s not Chat GBT. I don’t have the name of it in front of me. 

Luckily, my son’s girlfriend is in AI. That’s what she does for a living. And so, she told me the application. 

I told her what I was trying to do. And I said, I want Chat GBT to take my podcast. And write up a 1500 word summary of it with, you know, with paragraphs and topic sentences and topic titles, and stuff like that. 

And I know it’s going to be able to do it. I haven’t gotten around to kind of doing it yet. And that will allow me to replace the writer I have. 

You know, that person doesn’t work for me. It’s a contractor. So, that’s an application, right? And with the ease of doing that, I might do more podcasts because it’s easy to do the back end. 

And doing that writing is actually very expensive. So, that’s the application. And I thought, I didn’t think about AI. 

I thought about, hey, how can I eliminate this writer who’s costing the X amount of money and get faster productivity out of this, you know? I’ll give you an example of that. I mean, we did. So, there are applications like Fathom and others that can record online conversations. 

I have it integrated with Zoom. As many people do, these days you go into Zoom room and you see other Fathom note takers coming in. There are other companies that do this kind of AI recording as well. 

And so, what they’ll do is they’ll record the whole meeting. They’ll give you a summary of it with the key points broken out. And they’ll give you the full transcript. 

What you can do is you can copy and paste the whole transcript. I did this the other day, actually. Paste it into chat and say, write an article out of this.

In fact, I have a three-unit conversation with someone. And I asked it to write three different articles based on that conversation. I gave some parameters. 

But it gave nice articles. Now, of course, you want to edit them. You want to check that what it’s saying is reflective of what the person said is accurate and all that kind of stuff. 

But within minutes now, you have content to work with as opposed to the other way, either hiring someone or going through yourself and writing. Yeah. So, those tools are out there and available already. 

And that’s what I’m talking about. Like people using that stuff. So, you want to be taking advantage of all those kinds of things to speed up your workflow, produce more. 

And I’ll talk about how to be more effective. So, it’s not just about producing more, but it’s about those things being more effective. So, like, for example, with a fundraising letter, it’s not just like writing the letter, but it’s how do you connect with your audience more? How do you make it resonate? Anyone could just write a fundraising letter. 

And most of the time, it’s from the organizational perspective. And that’s what chat GPT usually does. Here’s all the great things we do. 

If you ask it, write a fundraising letter for the XYZ organization. It’s going to go find XYZ organization’s website. So, it’s going to scan that and know what the organization does. 

You know, we’re a dog rescue. So, it’s going to help save more dogs. And here’s what we’re doing this year. 

Our initiative. And here’s how you could help save more dogs and, you know, and donate to us. It’s going to be kind of formed away. 

But that doesn’t resonate with specifically the mindset of the donor and how they see, you know, their dog, right, or the whole issue in general. So, you want to kind of flip the paradigm and have the tool help you be more constructive in seeing things from the audience perspective. So, I’ll talk to you. 

We’ll get in. Let’s get into that. Yeah, please. 

So, the first thing to do is to understand how to use the tool and to think of the tool as your collaborator. Not necessarily as a sleeve. And I kind of think we sometimes see that, you know, sleeve, do this. 

And then you get something back, right? And then you’re like, you know, thank you for that. I will use it or not use it. But when you see it as a collaborator, two things happen. 

One is that it becomes a back and forth process where you’re not just asking it for something, but you’re responding. It’s responding to you. You’re responding back.

It’s responding to you. And so you’re helping to refine the response, you know, that you get. And then it will be better. 

And you do that through a series of prompts. And I’ll kind of, you know, take you through that. And now here’s something very interesting. 

There was a research paper study that said, when you ask chat GPT to do something, and when you tell it that there is a lot writing on this assignment, that this is very important, that it’s spend time to get it right. It’s time meaning very relative because it’s like literally seconds to do things. It actually will generate a better response. 

It might take another couple seconds or minute, you know, to do it. But it will be a better response. So thinking about it as a collaborative partner rather than just a function of performing a task will actually help you to get some better results. 

And it’ll help you to think more strategically about it as well. So, you know, for example, you can ask for some contextual information. Like, help me to understand, you know, this issue. 

Help me understand how people think about this. So before diving in, you’re trying to set the stage. And there are several different levels of setting that stage that will help you to get a better result. 

So first is kind of understanding the issue. Asking like, what are ways to approach this? You know, don’t just ask it to do something or don’t even make an assumption about that. But you can ask it for different kinds of approaches to a problem. 

And then you can choose from those different problems and you can ask like, how did you come up with that? Why did you come up with that? What evidence do you have for that approach? So you can start to understand the thinking that goes into it. So you’re going to get smarter because now you’re using it as a strategic partner to understand how to approach a problem, not just to get a result, not just to get a fundraising letter, but to understand how to frame an issue, what are the key issues surrounding, you know, my cause. How to frame those things now for the audience. 

And asking it for, as a collaborator, how it came up with those. So you can trace the thinking there. And it’s interesting to see, you know, what it might say. 

You can even, you know, ask what sources are. If it doesn’t provide them, now ChatGPT4 does provide sources. And you can say, you know, is there anything else that I haven’t asked that would be valuable to know? So not like making the assumption that actually it knows everything or that’s reflecting all that knowledge or that you know everything, but having that dialogue helps you to progress further than you would if you were just to treat it as a sleeve and like assign it a task. 

Yeah, it’s interesting. It’s like hearing what you’re saying, John. I’m not John, I’m sorry, Howard.

I have a call coming from this guy, John. Sorry. Howard. 

Well, the first thing I would say is, and I don’t mean to insult anybody by saying this, which is always a scary thing to say. But you always know how smart somebody is by the questions they ask, not by what they tell you. It’s their curious mind. 

And so by hearing you say what you’re telling me now, my mind is already starting to say, oh, I just need to think of creative questions and let my mind go and just wander about, well, why is this happening? And how does it happen? And what makes it happen? You know, so, you know, I got it’s funny because as you were talking, I googled because you made me think of something and I typed in Google and I just said, tell me about the different types of AI tools that are available. Because everybody thinks, you know, we’ve all heard about chat GPT, right? Not everybody, but a lot of people have, right? But what I didn’t know my son’s girlfriend told me was, you know, how many different types of AI tools there are. And when I typed it in, it came up with 30 different types of AI tools. 

And, you know, now I’ll kind of go through, not, well, I’m on this podcast, but now I’ll go through each one of these and say, oh, this one specializes in this type of function and this special, like one specializes in video editing. There’s a lot of kinds of, yeah, I mean, there’s hundreds of different applications for these. Yeah, you know, video, audio, creating avatar, video games, predictive analysis, you know, so understanding how your donors, how much they’ll give. 

It’s actually some, some, some donor software now they’re incorporating that. So instead of just having a generic donation page that has five different options, $25, $50, $100, well, maybe the donor can give $1,000. So if you just start asking for the $50, you’re underplaying that. 

So use predictive analysis based on different things, your ISP, what location you’re in and the home values, you know, in that geography, but your Thanksgiving history. People like you so forth and it’ll present you with different options that are tailored, you know, to you. So that kind of predictive analysis, there’s so many different ways that AI can be used. 

Yeah, they were talking about the narrow idea of generative of creating content. And even within then, like you’re saying, there’s language, there’s still images, there’s animated images, there’s video, audio, composite of these different kinds of things, gaming, all kinds of things. And they’re getting much more sophisticated where you can, I use mid-journey a lot for creating images. 

There’s new apps that now create based on a specific visual style that you can upload and then it will create imagery all based, you know, in that same style. It’s now embedded into programs like Adobe Suite, Photoshop and Illustrator can generate images right within there or free or more public access tools like Canva. So, yeah, it’s out there. 

It’s been integrated into all these things. And it’s learning how to use them in a way that helps

to not just generate stuff for you, but help you to be a better strategic marketer and fundraiser and thinker by asking it and prompting these questions. So, I was trying to think of the three areas, the three most important areas that a nonprofit functions in, I don’t think it’s the right way to put it, that you would then go out and do some research and see how AI could help you in an area, one of you ready to mention, which is fundraising. 

Right. So, like, give me some examples of questions that run through your mind right now that you would ask AI to help educate you on, not to produce any results yet, but to help educate, tell me something that you think you would ask AI. Yeah, sure. 

Well, the other day, I was doing some work with an organization on job training in Northern California. So, I asked it to provide an analysis of the landscape of that issue. So, I said perform a competitive analysis of that issue, break it down by different types of entities that provide job training, understand the demographics of who’s looking for a job, find reports and information that studies have been conducted in that area that define that. 

Do, you know, analyze the different offerings, you know, relative to each other. So, there’s nonprofits, businesses offering courses, there’s training education from educational organizations, you know, and other entities. So, those, there were, you know, other questions there, and it, you know, it generated a great landscape review for me. 

So, you could be educated, you know, really quickly. Then you can say, you know, what are the issues that are kind of most resonant to this organization, you know, that we’re working with. So, now you’re starting to use the intelligence that you’ve gathered to be more specific about how that could be applied. 

What are some of the issues and how are they framed and how could we understand, you know, how they could be used, you know, for, you know, for this organization. Give me three strategies, you know, give me some profiles of donors that might appeal, you know, find this organization, this cause, you know, appealing. You know, help define some strategies, you know, for reaching, you know, those constituents. 

So, you’re using it as a collaborative tool to help you accomplish an objective. So, you’ve learned about a landscape, you can get insight about the audience. Oh, I want to talk to you a little more about the insight about the audience because I’ve been trying to do this more because these organizations are smaller. 

Don’t have like huge databases where, you know, like I mentioned earlier, American Express and Target, you know, they have millions of data points so they can see their buying patterns and so forth. These small nonprofits, you know, don’t have that. So, how do they harness AI to do that outside of their database? Well, you can ask it for information about people who might care about your cause. 

So, have there been any studies done, you know, on the issue of, you know, workforce, you

know, development or homelessness or on environmental, you know, issue or whatever your issue is, and you can see what studies have been done. Then you can ask it to quickly summarize the studies, you know, and give a perspective and ask about the relative importance and, you know, of those things. So, you can very easily, you know, within, you could sit for a matter of 15 minutes, you could even spend an hour. 

And you can learn so much by asking questions. If you could find studies that have verbatim, you know, that are surveys or verbatim responses about your issue from your constituents, you can then start to go into a deeper level. We’ll call emotionally intelligent writing to reflect the sentiment, the understanding, the values and the motivations and the drivers of your audience. 

Based on their responses to surveys or focus groups or any other data that you could find out there that AI could help you find it. And then you can say, you know, what are the emotions connected with that? How are they feeling about this issue? Are they excited? Are they pushing back? Because this stuff’s happening in our backyard. Are they thrilled about what’s happening? Are they scared about, you know, what’s going on? So, your fundraising letter then is going to be very different, right? Because now you have this insight that’s the landscape of what’s going on and how they’re thinking about the issue, all the other competitors that are out there and not just from the focal point of your organization. 

You understand how donors think about the issue, the other options they have. And you can use AI then to write a fundraising letter based on this information, based on this understanding. And then continually use prompts to refine that, you know, find evidence, find stories. 

Do we have stories in our database that exemplify this idea of self-sufficiency, making our way out or fighting the establishment or whatever the larger themes are that you’ve just uncovered. However, what happens, what used to happen when you would ask those questions in a Google search engine versus the results that you’re getting from an AI tool? Well, Google search is just getting a list of websites. Now they’ve incorporated the AI, which is a summary. 

So that’s closer to what Chad is doing is kind of aggregating and synthesizing this information in the ideal world with all truthful information and reflecting that. So when you ask about something factual, what is the cause of this or the disease or something that has a factual basis, hopefully you’re getting that, you should always check. But it’s more of an analytical. 

So first, at the base level, it’s just synthesizing lots of information for you. So you’re not looking through all these websites to find the answer to my problem, the question that I have. But as I’ve been suggesting, it’s really more a strategic tool. 

You can use it in a more strategic tool by prompting it through these different questions to ask the why, not just the what information, and continually dig deeper and create more relevance for your particular use case. The other question I wanted to ask you, is you’re using chat GPT when you’re doing some of this research? Yeah. Yeah.

There’s Jasper. I’ve been on lots of webinars about that. I haven’t actually opened the account to do that yet, but that’s very powerful because you can upload or create brand databases for your clients in it. 

So you can upload and input your mission and your vision and your messages and your brand voice and your strategic plan and all the things that inform how you want to talk about your organization. So now when you ask it a question, it’s informed by that. It’s going to use the language that you use, it’s going to use the tonality, it’s going to have the focus on the priorities and all those things. 

And you’re not starting from scratch each time to prompt it to say, you know, here’s all those things as the context for it. So I know people are using that very successfully. Hey, have you seen something that’s been on my mind and that was, have you seen any of the nonprofits that you’ve worked with? Have you seen any of them taking their raw data that they have in their fundraising, accounting software, or QuickBooks, or whatever.