David Norris · Co-founder, Bold Crow AI | David Norris is a leader at three organizations that are making big things for nonprofits so they can make big things happen: Bold Crow, Proofpact, and The Nonprofit Hive all use trending technology to benefit the nonprofitsphere.
Today’s guest doesn’t just skim us across the surface of AI in the nonprofit sector; he’s taking us to the forefront of the future of fundraising.
David Norris is a leader at three organizations that are making big things for nonprofits so they can make big things happen: Bold Crow, Proofpact, and The Nonprofit Hive all use trending technology to benefit the nonprofitsphere.
As soon as we said “trending technology”, you knew AI was joining the chat, didn’t you? You were right; we can’t talk about cutting-edge tech today without it. …But what’s there to talk about that you haven’t heard? We all know there are AI tools for almost everything—if you know where to look and how to integrate them.
And that’s where David Norris comes in. David builds custom AI agents and applications that address nonprofits’ specific needs. Stuff like automations that reduce weeks of work to 10 minutes and conversational AI guides that accompany you through the digital universe. David’s also sharing info you need around AI’s ethical considerations, explaining autonomous AGI agents, and predicting which roles will be most affected by artificial intelligence.
This conversation is a must-listen if you’re looking for a deep dive into the potential that AI will bring to the nonprofit sector, who artificial intelligence is affecting the most, and technology trends to keep your eyes on.
Hello and welcome to this episode of Nonstop Nonprofit!
Hi, there Nonstop Nonprofit listeners! This is David Schwab your host and welcome to Season 4! We're starting off Season 4 with a bang by giving you the power of David squared! I am excited to introduce today's guest, David Norris, who is a leader at three organizations making big things for nonprofits so they can, in turn, make big things happen. Bold Crow, Proofpact, and The Nonprofit Hive all use trending technology to benefit the nonprofitsphere.
Now as soon as I said “trending technology”, you knew I was talking about AI. Well, you were right; because we can’t talk about cutting-edge tech today without it. …But what’s there to talk about that you haven’t heard? We all know there are AI tools for almost everything—if you know where to look and how to use them.
And that’s why I'm so excited to kick off the season with David. Because he doesn’t just skim across the surface of AI in the nonprofit sector like so often we get. No, he’s going to take us deep into AI technology for the nonprofit sector now and to the forefront of the future of fundraising.
David builds custom AI agents and applications to address nonprofits’ specific needs. Things like automations that reduce weeks of work into 10 minutes and conversational AI guides to accompany you through the digital universe. David also shares information you need around AI’s ethical considerations of AI adoption. He helps explain what autonomous AI agents are and gives us predictions on which nonprofit roles will be most affected by artificial intelligence.
If you're at all interested in fundraising or nonprofit technology or want to learn more about AI in the nonprofitsphere, this episode is a must-listen as we dive deep into the potential that AI will bring for all of us.
David Schwab Hello Nonstop Nonprofit listeners! This is your host David Schwab, Head of Growth and Marketing here at Funraise and you are about to find out the impact of David squared because we have my friend David Norris on with us today and we're about to talk about all things AI, machine learning and building momentum through technology. David, thanks for joining us today.
David Norris Thanks for having me, David. Appreciate it.
David Schwab For those in our audience who don't know who David Norris is, David is, in my opinion, one of the four front foremost experts in artificial intelligence in the nonprofit sector right now. And I have been itching to have this conversation for a long time because if you followed anything I talk about or anything, we're doing it fundraise, you know, have a heavily invested in artificial intelligence and building tools for the sector, mainly to catch the attention of people like David who are on the front end. So, David, again, thanks for joining us today. Do you want to just give us an idea of your background, how you got to the nonprofit sector, how specifically you got into the niche of artificial intelligence in the nonprofit sector?
David Norris So, yeah, again, thanks for having me. I really appreciate it. This should be a good one. And you set quite the stage. I don't know about actually being the top dog, but I definitely know that there's a lot of smart minds and a lot of smart brains that fundraise too. But how I got started in this was in 2017, I believe it was. I created an artificial intelligence chat bot and it was trained on Reddit data and using TensorFlow, and it was rather atrocious in terms of how it was responding. And I mean, it was, as you would expect, a Reddit conversation to go, but we had trained it on some subreddits that had to do with digital marketing and development and machine learning and things of that nature. So a little bit more user friendly, I suppose, but still not great. We didn't I mean, it was nowhere near the parameters that these large language models have today. How I got started with nonprofits. Well, we'll say it was a winding path. We had some nonprofit clients at my agency that I had started in 2015, but we weren't exclusive exclusively working with nonprofits. I wouldn't say that I exclusively work with nonprofits today either, but I have tools that are built for and meant for only nonprofits. Is that a full enough answer there? Yeah.
David Schwab So one thing I have to respond to there, you're telling me that A.I. chat bots existed before 2023?
David Norris That's that is a correct statement. It might not seem plausible, but.
David Schwab Yeah, so there was there was a life before chat.
David Norris GPT There was a life.
David Schwab Well, there'd be, you know, in my opinion, there's no better person to have on and have this conversation with seeing how you probably have more experience not just using but building AI and chat based tools to talk trends. You know, trends are people pay so much attention to trends and there are so many articles and so many resources, it's hard to cull through. But with your perspective, I want to dig into some of the trends of artificial intelligence and specifically AI in the nonprofit sector. But first, like what is it about applying AI to the nonprofit sector that piqued your interest?
David Norris A few different things. And I think realizing that there's an ethical use here, there's an opportunity for automation. You know, we all are familiar with things like the overhead myth and the fact that, like, you know, donor dollars are somehow expected to all make their way to the cause itself. And, you know, not seeing that that's a good reality. I'm saying that that is unfortunately one reality that is in effect right now. But I do believe that this can, that this trend, if you will, maybe not even a trend is this practical use of AI, is opportunity for nonprofits to rethink how they're operationally, operating. Digitally speaking, there's a tremendous opportunity here. And just in terms of automation, taking out the menial tasks that are, you know, taking up space but have to be done. And, you know, these digital operations can be handled by AI agents. You know, you'd mentioned trends as well. The conversational interface is a trend. It's been somewhat of a trend. But even before things like Chat GPT, right, like there were virtual assistants that you could build, you know, it was really cumbersome to build them prior to, you know, access to these large language models like we have today. And by cumbersome, I mean, you know, you were building them with, if this, then that type of statements, right. If this happened or if this utterance happens, occurs, then say this if this utterance occurs, and that's actually kind of how we were building chat assistants or virtual assistants back in the day was just with that. It was very simple, well, very simple, but got very complex quickly if this, then that type of pattern. And so with large language models given that there's a generative AI component now to this. You don't have such complex back ins and functional scopes. They're just to build a relatively usable, sensible chat assistant. So I'd say conversations or the conversational interface is definitely a trend and something that I see compounding into the future to where really we're going to be operating with websites that are more conversational like and I can give an example later, but that's just one thought.
David Schwab Yes, interesting. I've actually been hearing a lot of of thought leaders in the air space talking about this like conversational carry through that follows you as you travel online between sites and apps and tools. It's like a conversational air guide through the digital universe. I think that it's very interesting. I haven't cracked how that applies to the nonprofit sphere yet. I promise I'll tell everyone once I do. But it is an interesting trend to be paying attention to. On top of of trends, we've already mentioned a few different types of tools. There seems to be a new tool every day, right? There's someone's building something to do the next best thing. Sometimes it's hard just to get an idea of where to start. So if our listeners are sitting here going like, I don't know what a large language model is or conversational interface or any of that, what would be a good place? Like, what are some tools on the market today just to get started using AI and machine learning and maybe flip to the other end of the spectrum? What are some of the more robust or advanced tools that people could go and look to level up what they're doing?
David Norris Yeah, that's that's a good one. And this comes down to personal, somewhat personal preference and somewhat use case. But I would say for most people you can start with Chat GPT, you give it some instructions if you want, go in there. Do not put any personally identifying information about yourself or others. I mean you can about yourself I suppose if you're willing, but definitely don't, you know, take like a database scrape and put it in there and then have to do something with it, I mean, without obfuscating that data beforehand. So chat CBT is one that I would say is a good starting point. It is generative AI at the core. It is what sparked many of the ideas and I think you're aware like beginning of 2023 was what started this onslaught of generative AI applications and moving on from Chat GPT, let's say maybe getting a little bit more niche, but going into platforms like Fundraise, like appeal, AI, using those generative services to craft, you know, emails, email appeals, things of that nature, crafting content, using it to generate content, realizing that depending on the application, there's likely a draft stage before a, you know, final, let's publish this type of stage production ready content and moving from that. I would say, you know, if you're looking to have something built, like one of the things that we do is we build custom AI applications that interface with large language models that we host on our app engine and, you know, serve those for our clients. And that's I think, you know, really the the most niche that you can get to where your content is now being placed into either fine tuning data or training data. And there's things like vector databases and LLMs that get, you know, access through that. Now, by the way, a lot of them is large language model and vector database is essentially a way to store data that is accessible by what are called embeddings. And embeddings are essentially vectorized content. You can think of vectorized content as like almost like snapshots of like images that is your content in encoded form. So you have not to get all technical and nerdy there, but.
David Schwab We are here for all of the technical and nerdy deep dives we can do. So I mean, I'm learning new terms and use cases as we're talking. I think I want to bring it back a step, though, and think to application. So like there's a very clear application, you go to chat GPT and say, write me a blog post about X, Y, Z, you copy, paste it to your website, you obviously you adjust it because you're not just going to take what a machine writes and put it on your site, but you use that to start the writing process. It's a great way to do content for your website or appeals. There's more advanced features like using machine learning to customize donation experiences, which I'm a huge fan of. And then, I mean, there's even things I that and machine learning that, you know, models databases and creates audiences and lists and helps you identify major donor prospects hiding in your file that you would have never been able to find on your own. But I say that to say there there are so many things that take machine learning and artificial intelligence and are tools for fundraisers and people. Even in other departments at nonprofit organizations, sometimes it can be hard. We just talked about it can be hard to start because there's so many opportunities. So my my long winded question there is what types of tools or what things have you seen make the biggest impact for nonprofit tiers, whether it's fundraisers or programs like people delivering programs, even back office administrative support or finance teams? Like what are some of the things and teams that are seeing the biggest impact with artificial intelligence right now?
David Norris I think the the teams that aren't programmers, I'll give it a blanketed statement that I'll get really narrow, but the teams that aren't programmers are the ones that are benefiting from this the most. And that's, you know, you can think of it like a fundraiser as being able to ask in natural language some sort of question about data and that data being present in a natural language response. And I think that's the benefit of taking generative AI, hooking it into your database locally or within a non public application and then having access to that data. I mean, like you said, there's data that's being uncovered that is, you know, that would have never been touched previously, you know, But one of the fundraisers decided, hey, I'm going to ask this question of this, you know, sort of chat interface here that's hooked to our data. And I got back this really unique response. Is this correct? And, you know, it was, but that's just an example. Now, I'll say one of the things that I've seen that we've done that's changed the way people are, I think interacting with their data is hooking into, let's say, something like a database audit, something like Google slides. So like we we've built this tool that basically generates Google slides of or fills out rather Google slides that you've already built, replaces content within them based on, you know, CSP data as your data excellence data. And that's really powerful because there's these, you know, audits that would take, you know, sometimes weeks to complete that you can literally get done in 10 minutes. I mean, realistically speaking with a review, I think that's the that's the benefit or power of some of these applications that I think creatively speaking is one way to approach generative AI that I don't think a lot of people are doing. You know, most people kind of attach themselves to the look in a chat bot that, you know, has some of our data in it, or maybe it doesn't, it's just going to generate blog post, but we've instructed it to do it a certain way. That's fine, but there's so many creative uses for dinner today.
David Schwab So David, you were you shared this really interesting thing that you're able to do where you guys build an audit or build a report, populate data and visuals, and Google slides based off of combining raw data, using a large language model or an AI tool to transfer that data into a report. And it saves you not just minutes or hours, but days and days of work. I think that could be a really interesting place to go deeper and change the topic for a minute from like what can we do or what are some of the things possible? Like how do you do that? Because that seems really cool. And like if I was listening to this podcast, I'm sure some of our listeners be like, I would love to shave 5 hours off of the weekly report That takes me my whole day Friday to do like you're telling me. I could go home at noon on Friday and still get the same amount of work done sold. So help help us out. Like, give us give us the way to do that.
David Norris Yeah. And there's varying levels there, too. I mean, we can start with basic summarization. How we do it is we take essentially it's the equivalent of like email merge tags and you would build a report in Google slides or a PowerPoint converted and essentially write queries in natural language against those data or the merge tags, rather, that then get filled in with data. We have something called workflows which automates all of that. So in a sense you can pre format, you know, if you if you generate a standard report that is kind of your baseline report and you do so regularly, that is something that would be pulled in, you know, to the workflows and those queries be written so that you can essentially sit back and watch it just get filtered out on its own. We do provide a chat interface with that, so you can request specific merge tags to be edited or revised. But that is that's kind of how we do it.
David Schwab That's really interesting. It kind of reminds me. So I listen to I don't know if you listen to him as well, but I listen to the HubSpot marketing podcast where it's the HubSpot CMO and the Zapier CMO and all they talk about right now is building efficiency through AI and it's just super interesting and they talk a lot about agent. And I know you've talked a little bit, you've hinted at agents, but it seems like that probably maybe the next best thing for us to talk about and dig into here, because I'm constantly faced with the question in my own role here. But also as I talk to fundraisers everywhere is like, Well, I have to do 50% more with 100% less. Like, we have to double down on revenue this year because we've got $1,000,000 shortfall. But to do that, normally I would need five more people on my team. Obviously, when we're in a budget shortfall, I can't hire people, but I need the people to to close the gap that we're facing or any of those of those scenarios. And I'm really starting to look at like, okay, maybe maybe this the agents are an untapped resource or an untapped potential for fundraisers or, you know, program directors who have to log who they're serving or when they're serving. Can you just help us like, let's go really here, Let's go deep in this this agent land. Help us understand first what it is like, how you build them and some use cases for those.
David Norris Yeah. So early on, well, in 2023, there were these, you know, GitHub repos that were popping up with autonomous agents and it was really attractive. I, I personally really like the idea. You know, people were like, Oh, this is AGI, which is artificial general intelligence, which is, you know, the ultimate, you know, level, if you will. And it technically it's really not. I think we're still dealing with a lot of narrow intelligence here, so to speak. But these these sort of frameworks and repos sort of popping up and things like, you know, baby AGI auto GPT. LangChain is one that is more of a framework for building agents. And we actually started with long chain and then kind of diverted into our own proprietary workflow and automation system. But an agent is essentially something that an entity or an application, if you will, that that works autonomously, that completes tasks or, you know, objectives that it is given. And it does so by thinking through how or what or when it needs to do these things. So it can be relatively loose or it can be relatively rigid With our system, we've made it a little bit easier to become rigid in that, okay, do stage one, then do stage two, then do stage three. There are some frameworks that allow the agent to more autonomously think through these sort of thought processes, but without getting into the nitty gritty there. In short, an agent is capable of doing really complex tasks, multiple tasks, using multiple tools, and by tools I mean the agent will, for instance, open the web browser script, the web page. The agent will visit a URL in Google or search Google for something, and then the return. Those results can complete functions. It can execute zaps, it can do many different things. And that's that's I think the beauty of an agent is the way that we've kind of addressed it with clients is it's one step above a chat. But, you know, it's more automation than it is conversation. Although there can still be a conversational interface with this agent, we put it that way.
David Schwab It's super interesting. David understanding the ways agents work and how they work together and how you can build progressive functionality and automation without the necessity of human interaction, It's it's cool, but it's also a bit concerning because it makes me go like, okay, who, who's building the agent that's coming for my job then? So I'm going to ask you this next question is how long until, you know, we find ourselves out of a job?
David Norris That's a good question. Optimistically, I think it will be a while. And I say that because if we are truly doing this to better our relationship or for instance, we're doing we're doing these automations and workflows and handling these operations so that we can spend more time being human bonding time, building relationships, things of that nature. I look at this as an augmentation, and certainly right now the way that we are building applications is and similar to to what Fundrise is doing with Appeal AI, you know, it's it's draft mode, right? We essentially like we have this one relatively complex automation that we essentially have a staging area that before the final result is an email that goes out. Before that goes out, it has to be reviewed by a human. So rather than that person having to do all this stuff upfront, they can now just review and either approve or deny and those have functions behind them. But when approved, for instance, email gets sent. That is an example of a supporting. Well, it's an augmentation. Something that is not taking over their job, but rather making it a heck of a lot easier. And I think that's what we're going to see first.
- - -
David Schwab Well, now that we've gotten you all excited about AI, I want to pause for a minute and introduce another trending topic in the nonprofitsphere. Asset-based giving is largely an untapped tool in the fundraiser's toolbox. And until recently, it was a lever most organizations couldn't afford to pull. But as technology increases, the ability to give donors a chance to give out of their wealth to the cause that they already care so deeply about, is making it as easy for a donor to give an asset-based gift as a credit card gift. Take a quick listen to Steve Leatham, founder and CEO of Donate Stock, one of our partners here at Funraise as he explains the value of asset-based giving, how it's easier than ever to allow donors to give out of their wealth rather than their wallets, and how technology is enabling more organizations like yours to give more opportunities for donors to lean in to the causes they care so deeply about.
Steve Latham Hi, I'm Steve Latham, co-founder and CEO of DonateStock and I'm here to talk to you about how we're making stock gifting easy and accessible for all Funraise nonprofit customers. For decades, stock gifting has been a 1% solution. The wealthiest households giving stock to the largest nonprofits and educational institutions and foundations. It was historically very difficult for small nonprofits to participate in stock gifting, and it's always been a very difficult, tedious process for donors to execute stock gifts. Today it's a very different story. We've made it easy and accessible for all donors and all nonprofits. So first of all, why stock? First of all, it's the most tax-advantaged way for donors to give. When they donate stocks instead of cash. They actually can avoid the capital gains tax they would have otherwise incurred if they sold it and they can pass it on to you. So the profits get larger pretax stock gifts. It's also the most widely held financial asset and where most households have their assets concentrated. Their wealth is in their investment account. It's in their stocks, mutual funds, ETFs. It's not in a checking account. It's not in their credit cards. So this represents a massive untapped source of funding that now is actually quite accessible for all nonprofits. And like I said, we've made it really easy for Funraise customers now to activate stock gifting and inside Funraise and make it easier for your donors to make tax-advantaged stock gifts. Some of the issues that historically made stock gifting hard were for donors a lack of awareness and just a very painstaking manual process to execute stored gifts. That was my experience 15 years ago when I tried to make a stock gift. It was such a painful experience. I never did it again. And it was really that inspiration in 2020 that led us to realize that there was a big opportunity to help nonprofits by unlocking stock gifts, by making it easy for the donors as well as the nonprofits. It's also been challenging for nonprofits. Even those that can get a brokerage account, which is meaning that the larger ones, they really struggle with processing stock gifts. One, it's a it's a lot of work. It's a very manual-intensive process. Multiple people have to touch every stock gift. Number two, there's no information about the donor. That's one of the biggest challenges. Even though the donor filled out the paperwork to send their stock to that brokerage account, on behalf of the nonprofit, their information doesn't travel with stock, so nonprofits don't receive instant gifts in a brokerage account. They don't know who it came from. The donors assume they know because they fill out some paperwork. And so you have a big problem that the donors are not getting knowledge because the nonprofits don't know they're getting stock. So those are the issues that we confronted a few years ago. I realized if we could build a better way to make it easy for donors to execute the gift in minutes instead of hours or days and for nonprofits to streamline and automate the process of receiving, selling, acknowledging and processing stock gifts, then we can really unlock a lot of value for nonprofits and as well as make it easier, more tax-advantaged ways for giving for donors. So that's what we do. Think of us as PayPal, for stock gifting. It's an easy pop-up widget that allows a donor execute a gift in minutes. We then process that. We can send the stock to the brokerage account for a 501c3 that we manage through DonateStock charitable. Allows us then to convert stock to cash. Send it directly to you. So no brokerage required, no fuss, no hassle. It's really easy. And now it's in your platform as a Funraisie customer. And best of all, there's no cost to activate. It literally can activate it your giving form in minutes and then we can start processing these gifts for you. We also argue with the content to go educate your donors, your board, your communities. The assets through email, web content, a button for your site, for ways to give page, as well as social posts and other collateral and campaigns that we can army with to then educate your donors. So we can help make donors aware that stock gifting is easy, allow you to harvest a large portion of your donor's investment accounts, where their wealth is concentrated and like only source of funding for nonprofits. So it's really easy. Again, you can activate it inside a Funraise. You can come to DonateStock.com to learn more, and we look forward to helping you buck the trend and make the most of this Q4 given season.
David Schwab The DonateStock and Funraise integration makes it easier than ever for organizations like yours to empower donors to make bigger, more meaningful gifts than ever before. In the same way that technology is empowering you to create new avenues for revenue, artificial intelligence is creating new ways for you to be a more effective, more efficient and more impactful nonprofiteer. Let's dive back in with David Norris for the rest of our episode.
- - -
David Schwab So in your opinion, what kind of roles or job functions are going to feel the most impact? And I'm a, if you ask anyone on my team, I'm a ruthlessly optimistic person. So I'm going to frame it as an optimist in an optimistic way because I think it's a good thing. What roles, what job functions within a nonprofit organization or any role, any function, are going to see the most dramatic change as artificial intelligence, artificial intelligence agents and things like that are adopted and get broader adoption.
David Norris Yeah, that's a good question, David. I, too am also an optimist, and I like to think of this as more of an evolution. And if you think of this as we thought of how we use the Web Web 2.0, let's say like building websites when nonprofits decided, hey, it's going to be a good return on investment for us to spend marketing dollars on building a solid user experience within our website. Kind of the same thing goes here. Now, this does traverse into internal operations as well. And so I would say like, roles that are most affected or I believe will be most affected is almost every role. If we're doing this correctly, from a creative standpoint, from a marketing standpoint, from a development standpoint, from a fundraising standpoint, from a leadership standpoint, everybody has to evolve. And so I usually equate this to utilization of something like Chat GPT that is an individual that is using that application. So that application is an interface to a large language model. As an individual, I'm using it to do certain things. Now when it comes to like data, data security, like we're not tossing our donors data into Chat GPT, for instance, please don't do that. But that's an individual's use and I hope it wasn't too broad to say that everyone within an organization is going to be impacted by this. But I really believe that because when you adopt AI outside of something like Chat GPT, when you have these micro applications or operations or other automations built for you and for your organization, just like you would have a website built for your organization that is a team investment. Everybody is involved. Everybody is, you know, benefiting from those automations and workflows. And so I think that whether it's an AI assistant that's helping generate copy or whether it's an AI agent that is performing tasks on the back end and scraping data from the web for something like, you know, community outreach, I think those are all things that do impact the entire organization. And everyone in the organization therefore evolves.
David Schwab As we're thinking about roles impacted. What has been the most helpful use case of AI, specifically thinking about fundraising, what's been like the thing that you know, as as you dig in and you work with organizations and you work with fundraisers, what has been the most helpful application or product or or use case you've found so far and how the second half of that, how can someone capitalize on on that today?
David Norris This one is easy. It's a standout and the application itself is, well let's say data and reporting, but really it comes down to data accessibility. So I think I mentioned this previously, you know, data just became way more accessible. You no longer have to rely on people that know how to write SQL queries for the database and then translate that into, you know, essentially natural language or give this data set in some sort of like CSV format to someone else to then make sense of it and report on it or audit it or whatever it might be. Data just got a heck of a lot more accessible in that, maybe we're spending less time writing reports. Maybe these reports are being automated and written for us. Maybe these reports aren't being written at all, and we're just able to ask a simple question to where this data is now right at our fingertips and can be extracted via natural language. I think that's the biggest thing to come from generative AI for nonprofits in particular, and it also can make sense of unstructured data or data that's considered incomplete. You know, one of the biggest things that I see is how data is collected as well. And that's also maybe a close second in terms of, okay, so we have this field where it's a state field, right? But we didn't actually pre fill it with a dropdown selection. We've just left it open and people are typing in, you know, Ohio or some people are typing in OH or some people are typing in lowercase ohio or whatever it might be, or maybe they got the abbreviation wrong. And here we are now with this data set that has all these weird markings for, for the state, and it's those fields that we can then essentially connect with with A.I.. So I think that's, you know, data is the number one thing that just got cleaner, just got more accessible. And ultimately nonprofits can leverage that, take advantage of it through, you know, again, custom applications and integrations with platforms like (Funraise) yours, David, that ultimately lead to, you know, better donor retention, better donor engagement, better, you know, donor experiences all around when it comes to, you know, actually using that data.
David Schwab Yeah, I think it's so interesting that you went straight to data because I think that's really the most powerful way to use AI, too, as a nonprofit fundraiser or leader or program director or anyone working at a nonprofit is, you know, one of the hardest things to go through is the data that an organization has. So, David, I think, you know, a friend of mine, Tim Locke. He was on our podcast with me last year and he shared a stat that has stuck with me ever since. And it's that 90% of nonprofits collect data, but only 5% of nonprofits use that data to make decisions. This conversation I keep hearing about generative AI and I applications or adoption in the workplace. It really is the great equalizer. Mm hmm. It doesn't. It's not going to replace anyone. But people who know how to use it are going to level up their skills and their value faster, specifically when it comes to data and understanding data. And even like you said, that's such an interesting point, cleaning up data, that is one of the things I'm most excited about AI is that it takes away the manual work that requires very little knowledge. Doesn't take a lot of knowledge. No. Oh, means Ohio. Right. But go ahead and try and build a program that can scrub an entire column and recognize all the random ways someone's going to spell their state for every single state in the country. Now, what if you're an international organization and you've got to do that across like you've got to figure out mailing addresses across the world and translate everything. Like, it gets really interesting, that type of application, because then it frees a nonprofit tier up to be able to go actually make an impact. Like if you're a fundraiser, then you're not you're not spending 5 hours cleaning a data file before you do a data for to do a mail merge to send a fundraising letter, you get to spend those 5 hours segmenting your audience, creating content specifically for those people.
David Norris That is now meaningful data.
David Schwab Exactly. Exactly. The other thing I'm really, really excited about, and I know it's been a trend, it's been around for a while, but it's it's using machine learning and artificial intelligence to model file. So it's like you clean up your data then. Now you've got you've got that clean data. Now you can model it. And there's this really cool company that I followed for a while called Boodle. I don't know if you're familiar with with them at all, but they've got a2a system where you can ingest your file and it will help you do things you can only dream of, right? It'll say, Oh, by the way, these ten people on your file that you've never talked to because they give you $10 once a year are giving seven figure gifts to these other organizations that look just like you because you didn't know to ask it. So why don't like why don't you go ask them for a substantial gift, right?
David Norris Yeah, yeah.
David Schwab Here's a list of best prospects to convert from one time givers to sustainers. Or here's the list of people most likely to reactivate on your inactive list. Like it gives you intelligence about your segments that would have taken a small army of people to go through and decipher, if at all. Because that's I mean, that's really the dream of machine learning applied here, is it takes an aggregate of knowledge that is more than any one person could know and applies it in real time to give you that value of 1 to 1 information. That's one trend that I'm really excited and I can't wait to see adopt further. There's there's a lot of tools and platforms and companies now doing that type of audience modeling and using broad data sets and combined with machine learning to find trends and and build audiences and build segments. And I think that's going to be a great win for the sector because you're just making your fundraising dollar that much more efficient. So it's really it comes down to efficiency then too, is like, how can I make my fundraising dollar most efficient and how can I make my man hours most efficient? Because if if there's two things we're lacking in the sector, it's bandwidth and budget. So if we can make those more efficient, I think we're going to see a really, really cool formula for success.
David Norris Yeah. And, you know, as you mentioned, it doesn't take this expert level data scientist to then, you know, know how to model this data or know how to, you know, find key points within the data set or, you know, again, like these complex SQL queries that extract the data that you need to only be further fine tuned. And I think that that's the benefit of this. I mean, from a bandwidth standpoint, yes, but also from the standpoint that, you know, it won't take a small army. It can take, you know, an intermediate or even sometimes beginner level, you know, data scientists, if that's what you want to call it, to manage this data now. And I think that's the benefit because those types of people, you know, sometimes have really creative outlooks and spins on how to how they might want to leverage the data. But if they don't know how to actually interact with it now you can I think that's the beauty of of all of this. Again, I use the phrase data accessibility makes it a heck a lot more accessible to everyone.
David Schwab Mm hmm. I think this is a mean there's a lot we could unpack and continue to talk about here. But I think this is a good point for some. Also I want to bring into the conversation because I know you have a really strong opinion and good perspective here. And it's a conversation that I think is it's a good conversation. It's the right conversation, and it's really catching traction in our sector. And that's ethical application of artificial intelligence and machine learning. And and you've you've hinted at it like, hey, don't put personally identifiable data into chat. Right. Don't put your donors name. Use their donor record ID. Right. I want to dig in to like ethical AI. And, you know, very simply, first, like, what is ethical AI? Where are we at with that? Where did it how did like, how did the conversation start? And then I just want to like let you go free and and talk to us about ethical. AI What are your views on ethical adoption of AI, how it fits in the sector? What are some of the trends you're seeing in the conversation?
David Norris Yeah, this is this is a good one and it's a great topic, and I think it's also a foundational topic as you start to explore, you know, artificial intelligence or generative A.I. or certain applications assistants or agents for your organization and ethical use, I think comes down to also a phrase I like to use is responsible use. So responsible AI. And, you know, for instance, we can start with chat GPT, you know, not not placing sensitive data into the text area and then submitting it. But also I would just, you know, take one step back and just say, why are we copy pasting data anyways? That's that's kind of a recipe for disaster in and of itself because, for instance, you toss this into Chat GPT, let's say that it was uber private and there was no chance that it was going to be used as training data, which that is effectively what happens if you do not turn off that setting, which some people don't even know that that setting exists, there's better features and so on and so forth. So, you know, let's say that it was uber private. You would copy paste this data and maybe only a section of it would actually get indexed in terms of what was sent with your prompt and or that data starts to fall off as that chat history grows and all of a sudden you're now interacting with what you thought was wholesome endpoints and it's no longer wholesome. There's only maybe, you know, three quarters of your data set or the next conversation that you have, there's only half and then it completely falls out and you have to restart and do all that over again. And it just gets messy. And so, you know, there's certain applications that can be built or utilized. I mean, referencing the API here in this case, that is much safer for the data. The data can be confined or constrained to a private, let's say, vector database or even if you're using, you know SQL in this case it could be a my SQL database where this is stored again, privately, not for anyone else, only accessible by you, but again, just being responsible with how that data is handled. That's a that's a huge, you know, point for us in terms of when we build something, we tell explicitly, here's where this data is stored, here's how we're accessing it, here's how it gets sent over to the APIs are the endpoints for things like Open AI's API or Google's Palm API. So I think that the ethical AI, in my opinion, is an entire application that has been thought through previously in terms of how are people going to interact with this, you know, where's the data stored, why are people interacting with this and how is it beneficial to our team or our community? And so when we start talking about ethical use or ethical AI, I think there's a lot to be said for user experience and what portions of that user experience are powered by AI and what are not, for instance. And I guess I could bring up, you could have an AI powered application that scrapes the web for, you know, your community's sort of current sentiment and thoughts and status and then do things with that content, turn it into data, turn it into insight, turn it into something that you can then do community outreach for, or around using that as sort of the pivot point. And I think when you're doing that and you're creating this content with generative A.I., you have to realize this is a draft. This is not something we're just going to, you know, release and let do its thing. That's where the human involvement comes into play, where, you know, again, there's humans that are essentially ensuring that this is accurate. There's a review process greatly cuts down on time, but there's still a review that has to happen. And the other one, I would say is in certain situations, taking out the bias that could potentially happen with, you know, generative output. We can look at things like chat applications that are for emotional support or, you know, things like suicide prevention or things like that. There's just certain applications where this could play a role but might not be a primary role and might not even have a role involved at all. And I would say that in most cases where there should be or needs to be empathy, you're not going to get that with. I that that has to be a human. You can use generative A.I. to point a user to the correct human more quickly and more efficiently potentially. But as far as having conversations about sensitive subjects and sensitive topics where people are, for instance, having trouble, we should not be leaving that to generative. It's not there yet. It might never be there. We might think that we're at a point where it's close. But, you know, humans are going to have to be at the other end of those messages.
David Schwab Yeah, I think the human element probably is the critical piece there.
David Norris 100%.
- - -
David Schwab And now, enjoy this break from your friends at Funraise, the nonprofit industry’s most innovative fundraising platform.
Funraise Funraise provides exceptional, innovative, and friendly fundraising technology focused on nonprofit growth. Made for fundraisers by fundraisers, our state-of-the-art platform integrates every core aspect of front-and-back-end fundraising management into a convergent platform perfect for nonprofiteers like you.
Every day, we celebrate nonprofits like United Nations HCR and National Wildlife Federation as they trust Funraise to power fundraising campaigns around the world. (Warning: we go through a lot of confetti.)
Get started raising the funds and the fun at funraise.org [F-U-N-R-A-I-S-E-DOT-O-R-G]
- - -
David Schwab Now that you’ve heard how Funraise can radically change your nonprofit’s fundraising game, let’s get back to the conversation.
David Schwab In that. A lot of what I hear about AI models and and self like as teaches itself and machine learning learns from itself to create out like create the algorithm in the back end and learn and adapt and adopt. It still requires human interaction to keep it pointed in an ethical direction and apply that almost emotion to the logic that comes from A.I.. It's really interesting. As we think about ethical A.I. and and specifically the human element. I want to talk a little bit more here about humanizing A.I. And get your thoughts on what that looks like.
David Norris I think that one of the biggest things that anyone can do is just disclose in a case where A.I. is being used. I'm not talking about like watermarking every single image or things like this, but disclosing the A.I. was used to generate some sort of response in the instances where if it's like an email that's being sent, if that's an automated email, let them know that I, you know, was responsible for sending this and this is a two parter. It lets them know that if there is a lack of empathy or lack of personalization or some sort of error that hey, like, okay, this wasn't the organization, some humans sitting behind the desk just not caring. This was, you know, A.I. generated and it's likely it's plausible that that's more that they're going to accept that more easily than if they thought that it might have been a human that just didn't care. And then the other one is just building trust, transparency, and again, being honest with how you are using A.I.. Your community will appreciate that. They'll appreciate knowing.
David Schwab David is as we get ready to wrap up, I think the final really important topic we need to cover here is adoption of A.I. In the nonprofit sector. We've talked about ethical A.I., humanizing A.I., how to get started, how to go deep agents. All of these are really awesome things that can be used and should be used. But I think we we also need to pay attention to the fact that there's a lot of fear about A.I. and unknown things. And when we face fear and the unknown specifically in our sector, there are a lot of walls put up. So how can someone who wants to be an A.I. adopter and start to facilitate that change in that adoption internally? What is a good place to start?
David Norris I think yeah. And mixed in there, you know, the idea of, you know, exploration and implementation and you know where because right now we're looking at a landscape, I like to call it the future frontier, like we're literally living it. I just did an article on this today actually, where it's basically it's basically pointing at the fact that we're living the future frontier. There's still a lot of unknowns. There's so many people out there who consider themselves knowledge experts that still say things like Chat GPT is an LLM, which it is not. It's an application that interfaces with an LLM. And so I think a lot of this comes down to knowledge being transferred and just the unknown is what is scary. It's not that A.I. is going to take over the world or, you know, do these horrendous things. It's that there's a lot of unknown out there and there's I mean, even people that are well renowned or knowledge experts or whatever you might want to call them, thought leaders on, you know, platforms like LinkedIn aren't necessarily helping the situation by saying things like Powered by Chat GPT or Chat GPTq is is one of the best large language models. And it's you know, these are all things that are close to being true, I suppose, but they're just not factual. It's not true.
David Schwab I love that that idea of the future frontier, because, you know, we had the wild, Wild West was the great frontier. Then space was the final frontier. And now this really does feel like the future frontier where there's just so much possibility. It can be a little daunting to start when it almost like, I hate to use the term blue ocean, but I'm going to because it's the right thing here. Like there's just so much where could be a good entry point for someone to bring air in for the first time to their organization. Like we've talked about content and we've talked about data modeling and we've given a lot of ideas and like agents and all of these things. But like, let's say step one day one, what would be like, Hey, here's your first application, here's your first tool, low barrier to entry, low risk, something easy to can. It's everyone, like, Hey, this is a win.
David Norris Mm hmm. If I had to answer that with a specific application or a specific tool, I would say use a known service or a platform that you currently have that is offering some sort of generative a service application leverage that get to know it, but get to know how it works. Start there. That that's a small step. You know, for instance, you are using this application as a whole. You know, maybe it's a maybe it's a CRM, maybe it's a donor management system, maybe it's a fundraising platform, whatever it might be. Leverage tools within as a starter. I would also say that's my answer to if I had to answer with a specific tool, I would say taking one step back, though, this does come down to like foundational knowledge. Now, that foundational knowledge should be of things like, you know, what is what is the large language model, what is Chat GPT, how does Chat GPT work? You know, what is a GPT? What is a generative pre-trained transformer, right? What are what are these things? What do these things mean? And I'm not necessarily suggesting that everybody should learn to code or everybody should learn to, you know, put these things together or should learn machine learning or know how neural networks work. But what I am suggesting is, just like with websites we all know, like websites should have certain types of copy within them, you know, telling a story. We're presenting calls to action. These are all the well known pieces, parts of how websites are put together. And we know that there's content management systems like WordPress and Joomla, etc. that operate behind the scenes. Now, we don't know exactly all the functions within, but we know the general make up or architecture of of what is being put together here. Now, getting back to the future frontier, we're still putting these things together. There's still frameworks and foundational models and large language models that are being built, and there's even architecture frameworks that are being built for models that are said to surpass the GPT models, which the GPT models were created by Open AI, and that's their technology. Now there's, you know, other types of networks out there that are being leveraged by large language models that are said to be even better. So, you know, we're moving fast, I get it. But that foundational knowledge I think is going to help you. If you can only learn a little, that's going to be better than learning nothing and, you know, remaining scared or remaining fearful of the idea. I suppose not scared, but apprehensive of the fact that AI is dangerous and autonomously going off the rails or something. That's not the case. That's not where we where we are currently. There are some awesome automations that can happen, but we're not. We're nowhere near, you know, AI sort of taking over right now, maybe if we're not careful. But that's not, you know, this discussion, I suppose.
David Schwab Yeah, it's we don't yet have to submit to the AI overlords, but maybe next week we will be.
David Norris Did you say next week?
David Schwab Yeah. All right, David, this was a meaty conversation. I really enjoyed it. I learned things. I won't publicly admit the fact that I just learned what was an acronym and not just a name that someone picked out. So I'm not saying I didn't know that before, but just in case anyone else is with me, like I got you. I see you. This is great. I thank you for your time. Before we wrap, if our listeners want to engage with you further, want to learn more, can you give us some ideas of of where they can connect with you?
David Norris Yeah, connect on LinkedIn, I do a weekly newsletter called Future Frontier, and this week's just happened to be explicitly for nonprofits. There's some juicy ideas in there as to how to leverage this for things like community outreach, volunteer orientation, things of that nature. But every week that comes out, that's just, you know, a knowledge dump essentially in kind of a current status as to where we are with certain large language models. And that's Open AI things like Google's large language models as well as some open source, large language models, including Llama and things like Falcon, etc.. So I'd say LinkedIn is a good place. Also, you know, if you're interested in assistance or agents, check out BoldCrow.ai. That's our company's website. Just a little bit of knowledge there, but more so hopefully creatively gets you thinking about how to leverage, you know, agents within your organization.
David Schwab Great. Well, for those listening, I do recommend you check it out. I subscribe to the weekly newsletter you put out david and I learn a lot from there. Maybe your next newsletter, you can just drop a line and remind all the folks who subscribe what GPT is so you can do folks like me a favor. But this was great. Thank you so much for your time. I learned a lot as we were talking and I'm really excited. I mean, if anything, I take away I'm excited for the future of fundraising and the future of nonprofiteering. In the future front, I'm going to use your words, the future frontier that is nonprofits powered by AI or AI fueling the nonprofit mission. So thank you for your time today.
David Norris Thanks for having me, David. Appreciate it.
Thanks for listening to this episode of Nonstop Nonprofit! This podcast is brought to you by your friends at Funraise - Nonprofit fundraising software, built for nonprofit people by nonprofit people. If you’d like to continue the conversation, find me on LinkedIn or text me at 714-717-2474.
And don’t forget to get your next episode the second it hits the internets. Find us on your favorite podcast streaming service, hit that follow button and leave us a review to help us reach more nonprofit people like you! See you next time!