Mallory Erickson  00:11

Welcome, everyone, I am so excited to be here today with Cherian Koshy. Cherian Welcome to What the Fundraising.

Cherian Koshy  01:51

Thanks so much such an honor to be here.

Mallory Erickson  01:54

So I’m excited to dive into AI with you today. And I feel like when you first started to talk to me about AI, it was only like six months ago now. But I was in a very different place with AI, where I had felt like I had never touched it looked at it, anything like that. And here we are in the world of chat GPT and all of this AI awakening. So can you just start us off with explaining from your perspective? Like give us some background on what AI is what got you involved in it in the first place? And we’ll go from there.

Cherian Koshy  02:30

Yeah, absolutely. Way back in the day, I remember Garry Kasparov losing to a computer. And that was maybe like 1997, or something like that. And I was like, Well, this is fascinating. Like, this is like a grandmaster chess champion, losing to a computer. What does this mean? And then sometime in the early 2000 limb, maybe a little later than that I don’t remember exactly when. But the big deal for me was one of my heroes, Ken Jennings lost to IBM Watson on Jeopardy. And this was like in the summertime, like 1015 years ago. And I was like, Oh, my goodness, this is unbelievable. And I kept thinking, how overwhelming it is to think that a computer could be, quote, unquote, smarter than a human with all these things that were happening. And then, as you mentioned, like shaft GPT came out. And I was like, Whoa, what is this? This is fascinating. Like everybody else, I started playing with it, I probably started playing with it a little bit earlier than most people because I was working as a fundraiser in an arts organization. And we didn’t have programming. So I had a little extra time on my hands. I was like, What is this? And how can I use this to figure out my own stuff to make my own stuff a little bit easier, a little bit faster. And then I got to a point where I was like, Oh, this could be super helpful. And of course, like, there are lots of other companies that are doing cool and interesting things in the space around. But I’ve always been this kind of early adopter in tech and playing around with stuff. And so this was just the next natural evolution of all of that.

Mallory Erickson  04:05

Okay, and so tell everybody about like, okay, so actually, let me back up for a second when you started to use this. What were for you like, the biggest kind of like, Aha was in terms of, oh my gosh, this is really going to make X easier for me.

Cherian Koshy  04:21

Yeah. Back when I started playing around with us, it was during the pandemic, as I mentioned, I had seen everything on Netflix, and I had too many open bottles flying. I was like, I got to do something productive with my time. I’m gonna learn about this open AI project, which was very new. So I was back on like, DaVinci to maybe when I got on there, and I started seeing platforms like Jasper AI and copy AI, being able to write sales copy, and I was like, Well, if it could write sales copy for the pizza shop, or my hairstylist couldn’t write copy for nonprofits. And I started learning about that and I was like, wait, you You mean to tell me that it could write the fundraising email, it could write the impact report, it could write the annual report, he could write the thank you letter, oh, my gosh, this is incredible. And I kid you not what actually would happen at work, even during the pandemic is a 40,000 square foot facility, my colleague would come knocking on my door, and she would say, Hey, we got to send out this appeal, when’s it going to be ready, and I’m looking at this blank screen like dear donor. And then like, I would get the appeal done. And she’d be like, Hey, we’re gonna send this out, we need to have the thank you note for this, when you’re going to have that. And I was like, tomorrow, like tomorrow, and part of the problem, honestly, was, I’m not super good at writing. People think that I’m good at writing. But it takes forever for me to like edit something to where I don’t hate what I’ve written, then the other part of it is, I would always use some sort of like template that was online, Tom Ahern template, a gale Perry template, if somebody has a template somewhere for a fundraising email or direct mail piece, or thank you note or whatever. And I would do that I would use that template, I would reformat it to whatever organization I was working in. But then you can’t use that the second time, right? You got to do something different the second time, then, if you’ve been in fundraising for any portion of more than a few years, you’ve run through all of the templates, and you’re done. Now, you’re like, oh, geez, what do I do now? So this is a story of my life.

Mallory Erickson  06:30

Was there anything when you first started to play with it, or even now, where you’re like, Okay, these types of tools can help us up to this point. And here’s where they get kind of dangerous, or here’s the thing we have to watch out for?

Cherian Koshy  06:44

Totally. The example that I will share with you is that a lot of the stuff a lot of the content that’s being produced out of AI, whether it’s for profit stuff, or nonprofit stuff, whatever it might be, is not perfect. It’s not like the old Ron co set it and forget it, whatever AI spits out, you just automatically email to your donors or post on social media or whatever. The way that I think about it is, does it get me 80% of the way there? Does it get me to a point where I can now edit it, which I’m not as terrible at and get to a point where I’m like, oh, okay, now we just ended up with our organizational language, like how my voice sounds and put it together in a way that even like a team of people can review it. And I actually was having this conversation with some other nonprofit fundraising copywriter, folks, and they’re like, it’ll never perform. And I’m like, that’s not the barometer of success. I think the barometer of success is how quickly can Mallory close her laptop and go spend time with her family? Or go hang out with her friends? How much more? Can we engage with donors on a one on one level? How can we facilitate those real authentic connections that are our authentic selves? I don’t think there’s anybody who loves like transcribing data, or moving things from one column to another, or the ticky tacky junk that we have to do in our days. That’s what I think causes burnout and frustration. But like if I could sit in front of a donor and have those conversations with them, that’s what I love to do. That’s what I really enjoy. Can I get to a point where that’s really the best use of my time? 

Mallory Erickson  08:16

Yeah. And you know, it’s interesting, like, I was just writing my email to my list tomorrow, and I was talking about how totally different subjects I was talking about the way that the chatter that happens inside our head says fundraisers and the way we sort of gaslight, our own emotions, or over rationalize things and try to just tell ourselves, like you should be over that already, or whatever it is, and how that chatter actually takes a lot of our time. And I think what you’re talking about too, right, like we hear so often in this sector, understandably, I don’t have time for x, right. And it’s like, everybody has too much on their lists like that is just definitely. But also, there are different ways to use our time and how to best optimize our time. And I think what you’re talking about here is recognizing, okay, like if staring at the blank page is just like weighing you down, depleting your energy, feeling super overwhelming, there are tools to help you get over that action line. And like what I love, and I want to talk a little bit about nonprofit operating system, because one of the things I really like about it is like that I feel like you have addressed the barriers to action at a few different points. So you have a tool for both inspiring like topic inspiration for blogs and content. It seems like you have the ability to sort of like draft some content from that as well, like that first draft, but then you also have like an editor tool and a finalization tool, which I think is and you can tell me if I miss categorizing everything, but to me that’s like, you know, different people have different barriers when it comes to their content writing and like for me, it’s the finalization piece like both of them blank page. But also it’s like, how do I make sure that? Okay, I’ve been reviewing sentences so many times that I’m not even sure they’re in English anymore. This like pass through to like, clean up the grammar or make sure this makes sense. So can you talk to you about sort of how you thought about that? And just tell everyone a little bit about nonprofit operating system?

Cherian Koshy  10:20

Yeah, absolutely. So it was exactly that same mentality of what are the pain points that I have, personally? And then what are the pain points that I hear from nonprofits all the time throughout that entire journey of whether it’s content creation, or whatever? And so there’s a part of me that’s like, maybe I Frankenstein the system so much, because I’ve built out all these things that everybody has said, like, Oh, here’s the pain point. So I’m just like, well, how can we solve that, but to your point, one of the things that we talked about is like, what are questions that our donors or stakeholders might have? I might not even know what those questions are. So there’s a prompt that allows me to figure that out. And I don’t even have to know what I don’t know about chap. GPT. And how to engineer a prompt. It just says fundraising ideas or frequently asked questions, and you type in your nonprofit or even the website of your nonprofit. And it says, Here are questions that donors or stakeholders or volunteers might ask, and then you have those questions. You could go crowdsource it from your department heads or from donors or something like that. Or you could take that and get a starting point through the tool to say, how would you answer this question? And now you have a first draft of that same thing with like, social media. I’m not awesome at social media. But like, if I could come up with a just have an idea and put it into the tool, and it creates something for me. I’m like, Okay, well, this has to be more concise. So I could put it into a tool to make it a little bit more concise. But even like work this week, you know, we were talking about revising something. And my boss said, like, what’s another way that we can reword this? And I’m like, Oh, no. So I honestly I opened up the tool. And I was like how to reword this and came up with a different idea, a different way of phrasing it. And I put in the Google comments, I was like, chat, GPT says that this is the better answer than the other wording. And he laughed at that, you know, I think that’s part of it. I think the other thing that I struggle with is like with grant applications, and these absurd word limits, how do I take something that’s been created for 1000 words and shrink it down to 100? Or I have something that’s 100 words, how do I expand that to 1000. And the AI will do that for you, which is fantastic, right? It saves you a tremendous amount of time and word count, right? Like trying to figure out, what’s the word count, and all of that, being able to post directly to social media or share something via direct mail, those are all things that I’ve seen. But we’ve been talking so much about content, I think it’s important to remember that this revolution also has a tremendous impact on some of the things that were excluded from particularly small and mid sized nonprofits. In the past, even large scale nonprofits, I am privileged to work in a pretty sizable nonprofit. And we have people are, you know, data analysts and running data from Salesforce into Tableau. And I was sharing with them even like, Hey, here’s a tool for natural language query of a Salesforce database, or a big query, or things like that, so that they’re not spending time building SQL reports. For every person that wants it in our organization. Everybody can just say, I want to know how many donors started giving in 2016, and only came in on social media. If that information is in the database, it will pull that information out, create a pretty little graph, and then share with whoever on the team like it’s that is stuff that we didn’t have access to before, or because I’m not trained in data analytics. But if you have machine learning that can say, Hey, these are the trends or opportunities in your database like that would cost a ton of money or a lot of time or skilled volunteer to help you with that you can easily and inexpensively do nowadays, which is really fascinate. I mean, it’s super exciting. Coding websites, coding apps, like cheap and easy.

Mallory Erickson  14:13

That is like mind blowing level for me. I really want to double click on what you said about the grant applications because I hadn’t even thought about that in terms of these tools. But that brought me to my knees and I for a lot of neurodivergent folks like me that piece the how do you turn this into 100? Where I mean the amount of time I wasted word counting. It broke my brain.

Cherian Koshy  14:41

Oh words of my life that I will never get back taking something from 100 to 1000 or 1000 to 100.

Mallory Erickson  14:48

Yeah, you know, you said something earlier that piqued my interest around how when you put in that suggestion into your Word document you said you know, chat GPT suggests blank and I have Seeing this disgust, although very well might be because I’m not on the end of any of this, but like, what are sort of the ethics? Or how are people thinking about when you should? Or shouldn’t note that something came in formed by AI? Or does it ever mattered to say like I watch people on LinkedIn, sometimes write a post, and at the bottom, they’ll say, like, written by chat GPT. And I’m just curious, like, for folks who are feeling maybe a little bit more uncomfortable with the use of AI? How do you think about when it should be attributed or not?

Cherian Koshy  15:36

So I’m going to take a very unpopular stance on this one, I don’t really think it does need to be disclosed or disclaimed, primarily because I would do some pretty heavy editing on whatever chat GPT or any AI tool produced, I would never let it go, just straight out of the box. And I’ve written a few blog posts that have that, and certainly written like, webinar titles and webinar descriptions. And I’ll usually include that it was written by AI just as the joke, right, because I’m talking about AI. But I feel like with content that we produce for our organizations, it doesn’t really matter if four consultants worked on it, or we got help from our neighbor, or our friend or whatever. So there isn’t the same kind of threshold or conception of plagiarism as there would be under an academic discipline, right? So if you’re writing a paper for a grade, the point is that you show your work, right, like my kids get frustrated, because they know the answer, but they don’t show their work on math and whatnot. And well, you have to demonstrate that you understand how to do it. That’s the point of that activity. But when it comes to putting out a blog post or a social media post, you could have just copied it from somebody’s good at social media, or retweeted it. Right. So I don’t really know that it makes a difference in that regard. I guess what I would say is, the warning label on that statement is that you shouldn’t be pumping out content that isn’t relevant to your organization that you haven’t edited, that you have not adopted your organizational voice. And you have some sense of what’s needed, right? Like, why are you putting it out there. So that’s how I think of that. I think the other thing to remember, especially when it comes to just generic, generative AI content production, is that accuracy could be a challenge. I think GPT four, which I got to play with over the weekend is way better, and way more impressive than what people have access to with chaps. GPT, for example, but it makes up stats, like it makes up sources and things like that. So you’ve got to be really careful, especially if you’re going to include something around like a fundraising appeal that says X number of people live under the poverty line in Arizona or whatever, that’s potentially not going to be true. So you got to do the fact checking in that regard. Anyway. So those are all concerns that I would have around the core generative AI, the bigger question is around how AI affects personal information. So one of the things that has been true already, is that financial institutions are using AI to determine who qualifies for a home loan, who qualifies for credit cards, who qualifies for life insurance, that scares the bejesus out of me. Because the machine learning the code doesn’t take into account necessarily systemic bias and systemic oppression that has occurred. And so there are all kinds of things that can result from that in the financial sector, and all kinds of other scenarios like jobs, and whatnot that we need to be really careful about. And then when it comes to nonprofits, what does that mean for what I call like, philanthropic exclusion? Do we look at our donor data set and say, These are the people who are most likely to give and those people and I’m doing air quotes, which people can’t see in the podcast. Those people look like just the donors you’ve already had. And if you’re predominantly white institution in the United States that constitutes philanthropic exclusion from my vantage point.

Mallory Erickson  19:09

Yeah, I think that’s really important. And you were talking about before how you put into the tool, basically asking the tool to help you shift your perspective, right. I don’t know how to say this in a different way. And you are asking the tool to help you shift your perspective. And it’s interesting, I think that is what’s really powerful. And my fear is the same as you what happens when it reinforces biases, or are we asking questions in a way that feed it to produce content or to hold some of the same biases that we do? Yeah, I think noting that is really important. One of the things I’ve wondered Is do you think folks are using it to do research on individuals or philanthropic institutions particularly has that something you’ve seen yet?

Cherian Koshy  19:57

Yes, absolutely. So there are tools out out there that are using. And I don’t know exactly what’s underneath the hood in terms of the code base. But there are definitely entities out there that are using AI to identify people who might have the propensity to give might have the capacity to give, especially along information that’s publicly available out there, like wealth screening, right? So well, screening gives you a certain layer of third party information. And AI allows you to understand the connections between that data and other pieces very, very quickly. So I think it rapidly allows you to prioritize, but again, coming back to what we said before, we’ve got to be really careful about how we’re using that. And are we using that to the exclusion of other outreach strategies or engagement strategies, but I think it might be helpful to see like, Okay, if we take the information in our database, plus wealth screening, plus Experian marketing mosaic data, and we’re able to say, these types of people are most likely to respond to a social media ad versus these people are most likely to respond to a direct TV ad versus email or like a one on one or a one to many conversation, like, Let’s invite them to a gala, or whatever, that’s really useful information, if the information is actually credible. Like if we can use that to test and form a baseline around. I think the other thing, which I’ve been playing with quite a bit longer is using AI to help us understand the motivations of donors beyond just sort of their core, like what we think they would do, or what we expect them to do. So one of the things that we can do in this day and age is like actual testing of messages with existing donors or with prospective donors, essentially, like mirror donors. If we said like, there was a Children’s Zoo in Indiana, somewhere, we have that group of donors, and we can see like, do they actually click a button to say I would donate to this? And what prompts them to click that button? Is it this language? Is it this color and test those variables iteratively. And very quickly, or we can take a group of donors in Indiana, for example, and see what drives them to make a gift or whatever it might be, and then use that as a starting point to be able to change our messaging and grow that way. I just think it’s really exciting to see what those opportunities are to be able to improve what we’re doing at scale, ultimately. So it’s not just like sharings opinion on this wording is a good idea. This worked for me before two jobs ago. So let’s try this pro forma on this other organization that doesn’t do exactly what that does.

Mallory Erickson  22:37

Yeah. Okay, can I tell you a use case of how I use chat GPT. The other day, you can tell me if I’m crazy or not. So I was on a hot seat call. So I had a power partners member, they have 30 minutes with me, we’re coming up on time, and they were like, what corporate sponsors would be aligned with our organization. So first thing I did was I put into chat GPT, one of organizations in this region are the most similar to and then gave information about their organization. So the thing came up with like, I think I said by 10 organizations, then I said, Can you look at the corporate sponsors of all 10 of those organization, and identify patterns, or like the organization that then would be the most likely to give to this person’s organization. And he was like, pretty quick came up with a very strong prospect list.

Cherian Koshy  23:29

Oh, yeah, you should try it again, when chat GPT has GPT for loaded, because they will actually read the websites. Because right now what it’s doing is saying like, here’s what we know, from it’s 3.5, as at the time of this recording, so I think it’s 20 or 2021 data, but you can actually have it look at more current information. And then it will build out the outline for the sponsorship kit, it will build out the marketing plan, just to ask those questions. And like, oh, my

Mallory Erickson  23:58

gosh, yeah, I’m gonna go back because I just got the GPT four, I’m going to try that. Yeah, I mean, I think that piece that you said around, like with any content, whether it consultant wrote it, or you hired someone on Fiverr, to write it, you would never just post it, you would always edit it, you make sure it’s in your voice, you would confirm the quality of it, you have to stand by this content, even if the AI wrote it first. And so I think that’s like a really big takeaway for folks is to figure out how does AI help you get over the action line, and ensure that you have the checks and balances in place to make sure that you’re using it effectively and responsibly?

Cherian Koshy  24:37

Yeah. And what you mentioned made me think of how many consultants have to say no to working with an organization because they don’t have the budget for it or a consultant will like when I was consulting, I would say, Okay, go back and fill out this information. And then we’ll come together and weeks would go by and they’re like, you know, like came up and we have a board meeting and blah, blah, blah, and we just and get to it. But how much more effective can consultants be in this space? Because a nonprofit can go into the tool and say, draft a marketing plan, draft a development plan, and update it to their specific use case. Or like I was talking to a social media person in nonprofits and them saying, like, we spent so much time crafting both the strategy but the actual implementation of all of those pieces, what if they were just editing it? What if they what if it created an entire month’s worth of social media content, and then the nonprofit went in, modified it and the consultant came in and said, Okay, now let’s level up the strategy. Let’s coordinate this with email and direct mail and create a full on Omni channels approach that really works and be able to now test it and iterate on it. And the opportunities are endless. Now that we have what I call, you’ve heard of software as service, I call it technology as a sidekick, where we really have this superhero sidekick standing behind us all the time, that’s able to do some of the work that we just don’t love doing. And for each organization, they have to figure out what is the work that you don’t love doing somebody like Juliette Campbell, I’m sure loves doing social media stuff. I do not. So we’re gonna use AI very differently. But there are things that I think like everybody dislikes like, well, I shouldn’t say that. But I don’t want to be a shithead. But I feel like things like taking notes in meetings is a terrible waste of someone’s time, especially like a board member who you want to contribute to the meeting. So I have an AI notetaker that joins every meeting. 

Mallory Erickson  26:36

Yeah, otter. 

Cherian Koshy  26:38

Oh, I use fireflies.ai. But I used to use otter

Mallory Erickson  26:41

two options for y’all.

Cherian Koshy  26:44

No, no, like affiliate anything there. Oh, no, no, just like solutions. Yeah. And so fireflies transcribes everything it gives me actually, there’s a summary. It goes immediately to our notion database, like, it’s super cool. I mean, imagine having that historical record for your board meetings.

Mallory Erickson  27:02

Oh, my God, it is amazing. It has changed my life. I love this. I love thinking about. I love the way Chad GPT if nothing else, I think has opened up people’s eyes to like what is possible and made people a little bit more comfortable. And then using tools that are tailored to the thing that you need it for. And so we’ll include a link below as well. So folks can go check out your work and try it out the nonprofit operating system. Anything else you want to make sure before we hop off,

Cherian Koshy  27:34

I would encourage everyone to not feel overwhelmed by the technology. I feel like there’s a lot of technological advances that are happening right now amongst all the other things that are happening in our world. So please, if you take anything away from both this podcast episode, this should free you up to be able to spend time with friends and family to look after your mental health to be able to do things like listen to a podcast at 1x speed and actually take in what Mallory and her guests share and go through this professional development and enjoy the company of other nonprofit professionals and learn from them as well. So I just hope that this is rather than feeling overwhelming. I hope this is encouraging to your listeners in a way that helps them feel empowered and excited about doing their best work.

Mallory Erickson  28:23

I love that. Thank you so much for joining me for this conversation today.

Cherian Koshy  28:28

My pleasure. Thanks again for having me.

Mallory Erickson  28:36

Okay, I so enjoyed this conversation. And here are a few of my very top takeaways. Number one, explore and experiment with AI tools like chat GPT for content creation and editing. Number two, use AI powered tools to help with grant application word count adjustments, expanding or condensing content as needed. Holy moly, did I need that? Number three, consider using AI generated suggestions for rewording and improving written content in various nonprofit communications. Number four, try out AI generated social media content ideas for engaging donors and stakeholders, test some things and see what happens. You can also share and collaborate with your team members using AI tools to streamline content creation and editing processes. And lastly, use AI generated frequently asked questions to inform blog topics and other content for your nonprofit. I use it a lot for some back and forth asking of questions and getting deeper insights into some of the things that are top of mind for me. Okay, for additional takeaways and tips inside this episode, head on over to Mallory erickson.com backslash podcast to grab the full show notes and resources now. You’ll also find more information there about cheryan and nonprofit OH S Thank you for spending this time with us today. If you enjoyed this episode, we would love it if you would give it a rating and review and share it with a friend. I’m so grateful for all of my listeners and the good hard work you’re doing to make our world a better place. And if you miss me between episodes, stop by and say hello on Instagram under what that fundraising underscore. Have a great day and I’ll see you tomorrow for the next episode in this mini series.