Trent Ricker: Welcome to the “Go Beyond Fundraising” podcast, everybody. It’s good to be back. I’m Trent Ricker. I’m the CEO of AGP, Allegiance Group and Pursuant, and I am honored and privileged to be joined today by Raney John, who’s our Vice President of AI Strategy and Solutions, the appropriate guy to have on a podcast like this.
Ren, say hi to everybody and tell everybody a little bit about your background and how long you’ve been with AGP and frame the role a little bit because probably last year at this time, not many people would have a title like yours.
Ren John: Yeah, exactly. So, I’m Ren, VP of AI Strategy and Solutions. It’s definitely a new title. I’ve been a technologist and software developer for over 25 years and have been at AGP for over 15 years now serving the nonprofit community. This title is new and did not exist five years ago, and I suspect a lot more AI titles will soon exist as AI permeates the workforce. Just like developer wasn’t a title, you know, or web developer or software, those weren’t titles like 15 years ago. Then, as technology and everything grew, we grew and evolved with it.
Trent Ricker: That’s great, Ren. I want to kind of frame this for everybody in the sense that first and foremost, I run an organization of 100 and some odd team members, and we’ve got lots of different functions. This kind of wave of AI, as it hits us, is for AGP probably the same that many of our nonprofit leaders are out there struggling with.
And for us, earlier on, I think through my own experiences and the experience and experimentation of our team members, like many of our listeners today, we’re trying to figure out how do we encourage innovation and adoption and do so in a responsible way. And what does it all mean? It’s changing so fast.
I think as we focus on you, the nonprofit leader audience, it’s not theoretical anymore. Adoption is not theoretical. It’s happening fast. It’s happening kind of unevenly, I think. Staff, like we experienced at AGP, I think staff are experimenting. Leaders are curious, but the training is probably thin. I’ve been working in the nonprofit partnership space for a long time. If I know anything, it’s that resources are usually scarce. So, it’s not something that is early adopted, invested in for R&D and training. Policies are often behind, and the technology itself is moving faster than most orgs can absorb.
And that’s true for us again, too. I’m in a lot of the same boat. And Ren, thank you for your partnership, because we put you in this role because of your curiosity and because of your leadership. And we’ll talk more about that throughout. I don’t think anybody listening today that is the question is – should nonprofits use AI? I think that’s kind of an obvious thing that we need to do to put a stake in the ground on. You do need to be using AI. Maybe the better question is – how do they move from a kind of casual experience, experiential type of an approach to practical, kind of responsible, mission-aligned use that actually improves how work gets done? Ren, when you and I were talking about prepping for this, well, there’s so many topics, we’re probably going to do a multi-part series on this, right?
So, for today, there’s things like fear and ethics and governance, and we’re not really going to hit on that too much today, right? Because I think it’s important. I don’t want to minimize those things, but I think today, I think most of our audience is eager to just learn more about how they can use it and benefit from it.
And so, you know, kind of before we kind of get started, I did a little bit of research, too. And I think that research has shown that nonprofits are using AI, but most of them are still using it at the surface level. And there’s an opportunity now to move forward from experimentation to impact. I think that’s probably true. And, you know, moving a little bit from can ChatGPT write an email to can it summarize a report? Can it brainstorm campaign ideas? It still matters, but now it really feels like the conversation has shifted.
Ren, when you look at the nonprofit sector right now, are we still in that kind of curiosity phase or do you think we’re crossing into something kind of a bit more operational?
Ren John: I feel like nonprofits are moving from experimentation to see how they could actually implement this to make an impact on their missions – they really want to help drive the mission forward. Some people are just realizing those goals. Some people are just, like you were saying, at the surface level – even surface level can develop and bring some efficiencies to your processes. The generation of emails, taking some mental load off of you. And as you experience more of those surface levels, then you can start going deeper and deeper.
Trent Ricker: Yeah, I think you’re right. And I think, you know, just like us at AGP, I think AI being more and more embedded in the tools that we use every day creates the kind of use case, right? Like we’re in a Microsoft environment, so within our Microsoft 365 environment, Co-Pilot has come a long way. We might have some listeners that are in a Google environment, and Gemini is kind of built into some of that stuff as well people are using it a little bit. But we’re shifting from prompts to something bigger.
I think before we kind of get into some application, I want to kind of slow it down a little bit and define some of what we mean. There’s so many phrases that get thrown around out there, right? AI can mean a lot of different things. I want you to explain, Ren, a little bit. It can be kind of intimidating, right? We hear AI a lot. We hear generative AI, automation, predictive modeling, agentic AI – lots of things, applications, agents – a lot of things that start with A, right? Agents and agentic. I think that tech guys like you understand why that is and what it is, but I think that the layperson sometimes can get lost on it. So, give us the plain English version. When nonprofits hear AI, generative AI, agents, agentic AI, what do they mean when they hear those?
Ren John: Generative AI just means kind of what it says. It can generate something for you. It can generate an image, it can generate video, it can generate text. You give it what you want to generate – write me an email to my donors that thanks them for donating, and it will generate a thank you email.
Trent Ricker: Got it.
Ren John: that you can then modify and go back and forth. Whereas agentic AI, or an AI agent – basically agentic just means it can perform tasks on your behalf. So, when we’re talking about agents, generating the content is one thing, but sending the email using Outlook, if the agent can and agentic AI can send the email for you, and it can respond to an email for you. Or coding agents are something we see a lot in the development world where an agent will write the code for you and test it and review it for you. So, an agent will perform tasks for you. And it is kind of getting blended in today because they’re building a lot of these agentic features and capabilities right into your ChatGPTs, your Geminis, your Copilots.
Trent Ricker: Yeah, that’s probably where, I mean, I know from my leadership position, there’s still kind of a scary “what if” you know? I don’t know if I want the computer acting on my behalf for certain things, right? So that is where this bridge comes. I think for a couple of years now, people might have been using ChatGPT in their personal life. And one of the stories that I heard to make it a relative – you might ask ChatGPT about a great steak recipe for your grill tonight, right? It can give you nice recipes, and that’s a task that it might perform. But to actually have AI or a robot in the future someday go cook you the steak and serve it to you would be more of the future agentic application, correct? As opposed to me learning something and then doing it. Today, we ask AI to maybe help us craft an email. We then edit that in our voice. And that’s probably the fundamental level of people getting into AI. Is that fair to say?
Ren John: Yeah, but I wouldn’t be too worried about the robots yet. I think the agent we want to take like these tedious redundant tasks off of people’s plate. One example I like to use a lot is if I’m comparing data between two worksheets, I can have AI take it and put it into its separate worksheet, build out graphs for it, and build out analysis for it. It’s doing those tasks instead of me building a table and then building a pivot table and then, all right, finding a chart in Microsoft Excel and figuring out how to get my data in there and building the labels. It does it all for me. What we want to do, especially in the nonprofit world, is we still keep the human in the loop – which is the main thing. The nonprofit world is all about human connection. We’re building things to make human society better. AI is just a tool that helps us do that.
Trent Ricker: Very important, and I think that’s really where the opportunity sits now, right? Most nonprofits are probably using AI at the task level, right? This summarize that, you know, it’s helpful, but limited. Maybe they’re using it for some brainstorming. And I think to your point, the bigger opportunity that we have to ask—and I think this is where the rubber meets the road a little bit for our session today – is where is work getting stuck? You know, where is staff spending too much time on repetitive, low-value type tasks? Where could they create capacity that does not replace human judgment? And I think that’s an important aspect, this fear that still somewhat exists, that if I explore too further on this, that this might – particularly for a staffer –could replace their job. Boy, that’s not the mindset to be in at all. Let me ask you, thinking about the nonprofits, where can they identify the right places to apply AI? Do they kind of start with the tool or do they start with the bottlenecks in the work? Where have you found that even in our own business?
Ren John: I would start with the bottlenecks in the work, the processes – where things slow down – because when you’re using AI it’s the intent to remove these bottlenecks. It’s the intent to free up space for your workers to use their human skills – their empathy, their connection – to build their relationships and create that capacity. Like you were talking about, a lot of nonprofits don’t have capacity right now. They’re very under-resourced, understaffed, and this helps bridge that gap. And that’s really what we should be focusing, especially in the nonprofit world, of using AIs to bridge that gap and build that capacity.
Trent Ricker: A good example I was thinking about as an application – let’s take volunteer training or something like that for an example. Back in the day, you might say somebody would come in and talk to the volunteers. We should still do that anyway, have that have that human element – but we might press play on a video and have them watch a video before they started to do their volunteer work. Let’s say it’s in a food bank, and they’re going to do some food sorting for that day. That’s still very applicable. But I think, kind of walking along that example, how might someone use AI for the training packet or other materials that could be more personalized? And when we need to make some edits or updates to that, how we might be able to use AI to kind of update those things in a way that would be much more efficient than the old school way of producing another video, for instance, right? Give us a little example on that.
Ren John: Well, a great way of using AI with training materials and documents like that, it’s something we actually do here at AGP. We’ll throw in its own custom GPT or agent. It uses that knowledge store, and then the food bank person who’s volunteering for the first time, instead of having to continue to go back to their supervisor, “oh, what do I do with this food?” They can go in their phone and say, “hey, what do I do with this food? What shelf does it go on?” And if it’s in the training material, the agent can answer those questions for them. A lot of times you’re watching a video, you see it once, it’s great, but you might not remember it. Or as things happen, things like you’re in the middle of things, you can start asking …
Trent Ricker: Great point.
Ren John: Questions to the agent as a first line of support versus having to always go to the supervisor there. And then the supervisor is freed up to do the normal supervisory tasks and make sure things are flowing correctly. On the other side, as far as even updating the content, you could have AI just kind of refresh, look, review the content every year and keep it green and create a process that now moves a little bit faster because AI helps you review the content. AI helps you edit and generate the content. AI can help you with the videos and training materials, and you can ask AI, “what am I missing?” Or take feedback that you’ve gotten throughout the year from your volunteers, feed it to AI to update the training manuals, which is another great use case.
Trent Ricker: I think that I want to go on a little tangent here because I think what we hit on is very important. What I found and what we found here at AGP is the source documents for any use with an LLM or whether you were going to go all the way to an agent are really critical. Source documents allow the AI to be able to refer to what might be written by somebody who has the guardrails or the vision. Maybe that’s the legal parameters related to volunteer – to keep it that example. Maybe it’s the vision specifically related to what the impact that volunteers make, to your point, what we can and can’t do with certain foods while we’re working in that volunteer. Being able to draw from that and not having to retrain is a real key element.
I love it because I think that it allows for us collectively – the knowledge of us here at AGP or in a nonprofit organization – to be in other places that we otherwise wouldn’t be able to be through our words and our thoughts that we bring to that process. In some of the other examples that we can maybe tee up, when we think about stewardship copy or something like that. Stewardship copy coming from the CDO or the CEO’s perspective – their tone, their voice training – so that the type of copy isn’t generic. It’s more based upon that particular organization’s mission, the specificity of what we’re stewarding, the tone and the voice that’s generated. And that would be considered a good example of generative AI, write a thank you letter for those who came to our gala last Saturday night, right?
Ren John: Yeah. What you’re saying is giving it the tone and voice you have in thank-you letters that you’ve written for years. Nonprofits have this information. You feed it to the AI and say, hey this is how we speak; this is how we treat our communities. This is how we show up for them. This is like the language we use. It will learn and it will adjust to that, and when it generates something, it will use your voice and tone and branding because you have trained it on that. You say, hey, here’s what we’ve done before. Now we want to move forward with that same type of voice. We’re not doing a generic generation of a thank you. It’s going to be a thank you based in your – the nonprofit’s – voice and tone. And yeah.
Trent Ricker: Right, or starting from scratch? A new copywriter that might be on a staff who’s tasked with that would take a learning curve to get to know, they could read all the letters and get to know previous tone if they’re writing on behalf of the CEO or the chief development officer. But in today’s day and age, and you hit something I think is really important too.
Sometimes people think that there’s an obstacle because, oh, Ren, I don’t have all those letters as digital files. Tell the audience a little bit how they can overcome that because we can use our phone snapshot and pull those things up. But talk about that a little bit because that’s really not an obstacle. If you have access to anything physical, AI does a wonderful job at being able to translate that.
Ren John: Yeah, AI can read text right out of pictures. I sometimes take screenshots and just feed it to AI and say, here’s what it is, and it will summarize the data it’s getting. You can ask questions about the data you’re putting into it. Like you were saying, a new employee can come in and say, oh what is the tone and voice here? And based on what you feed it, it will say, this is what we’ve gathered. And then you can even say, how do we make it friendlier – and that’s the brainstorming part. Just keep going back and forth and using it as a partner versus just as something you just ask a question to and leave it, it becomes a conversation, which is amazing.
Trent Ricker: Let’s take that a little bit. Source documents are critical. They’re all over, so don’t be limited by what you have digitally. Use your phone, take pictures, and upload them. The more we teach – think of it as an employee – the more that we teach that new team member about the role that we are asking them to play, the smarter they’ll be and the quicker they’ll be up to speed to do it well.
We are not going to expect any kind of AI to read our minds. We have to kind of lead it through the training that we otherwise might give a new team member for a particular task. You’ve started something internally that I think we do every couple of weeks called Saved by the Prompt, right? And so, prompting is critical. Talk to us about prompts. I think you’ve got a great acronym as it relates to prompting.
Ren John: I actually learned this from the nonprofit community myself. It was called CRAFT because you craft a prompt. C is for context, and then R is your role. What role do you want the AI to take on? A is for action – the action you want it to take, the tone. Our F is for format. Do you want it in an email? Do you want a short LinkedIn verb? Do you want a tweet or Instagram post? And then tone. So that C-R-A-F-T is just a great acronym for context, role, action, format, and tone, and that can build a really good prompt that’s easy to remember. You put all that data into the AI, and it can give you a better output.
Trent Ricker: That’s excellent. Well, let’s walk through an example using CRAFT and maybe an application. First, before we do that, let’s you and I kind of kick around some possible practical applications for the nonprofit, right? I think about where my role is similar as it relates to a senior leader at AGP. I’m using AI quite a bit to help me prepare for board meetings, for customer interactions, or prospects that we might be talking to better understand and do some research. Let’s bounce some other ideas around as it relates to both leadership activities that might be universal and specific things from nonprofit. What are some of the use cases that you’ve seen?
Ren John: You have to report back to your leaders how everything is going. If you feed it numbers and you say, hey, craft me an executive summary for my CEO as I – as in my tone and voice as the director of development – set the format to three key takeaways and be in the executive summary plus the summary follow-up actions that we’re going to be taking. It will build that out and it will because I gave it the context of this is of the data I’m putting in,
the key takeaways I want to give in the executive summary. That’s the format. I’m saying who it’s for and why and who it’s from and that will give out the tone and then you send that to there and you get that executive summary. It’s something you probably take a longer time to draft. It’ll get you a nice first draft, and you can edit it from there. But I use this all the time, especially as I transition to a more leadership position now. And I’m interacting more with the CEO here. I find that my detailed reports need to be shortened as they’re going through a lot of emails every day. So, I was like, oh, I’m doing too much. Let me shorten it, and then I’ve had AI help me with that.
Trent Ricker: That’s a great example. I want to think about this in two buckets, right? There’s things that we do internally, and then our nonprofit organizations are doing things that they are interacting with their constituents. Think about that as we’re using that. Let’s stay on the inside.
So, it wouldn’t be uncommon for many of our listeners that are involved in fundraising to have to do periodic reports on moves management, gift forecasting, interactions, portfolio management. There’s probably a regular drumbeat of reports that may be weekly reports to their chief development officer and maybe that CDO has a drumbeat of reports that’s a summary to the CEO who’s going to report something up to the board during their board meetings. Taking a look at that, it wouldn’t be uncommon for a gift officer to everyday log into Razor’s Edge, and they may be doing a certain number of functions. And they might, in the old-school days, print something out and organize their day with their pieces of paper. And now that moved a little bit more digitally, but still the end of week sort of performance and the insights that it could otherwise glean thinking about those processes, and I’m glad you brought that up, because I think internally at AGP, many of our reporting mechanisms have shifted. They’re smarter, they’re faster, they’re cleaner. And even myself as a leader to say can you dive in a little bit deeper, Ren, on that? You’re organizing your data in a way that can then present it in a different way, executive summary or a little bit more detail. I like that one as well.
Let’s shift a little bit then to the donors. And we talked a little bit about stewardship, volunteer management, certainly campaign briefs for internal use that could turn into letters or creative. What other sorts of uses have you seen out there? Grant report, grants are a great example, right? Grant applications – we’re seeing a lot of that. Yeah, grant writing has changed a lot.
Ren John: A lot of that AI has started to come in, and so a lot of people are using AI to help write the grants to make sure they’re hitting on all the requirements of a grant proposal and make sure it reviews the grants. You can have AI review this as a grant seeker, as a grantee or a grant funder, as a foundation. You can upload the foundation’s information and their past grantees and kind of like, okay, hey, what’s the likelihood of me getting this grant? Ask it questions. have it give you feedback on your grant writing process, which has really helped some organizations – especially those who don’t have a lot of funding or a lot of people who are trained on grant writing – they can develop better, higher quality grant applications using this.
Another great example is we talked about fundraising. A lot of CRMs now have AI starting to get built into them – Salesforce as the agent for Fundraise Up – they’re all starting to build out AI tools and it’s nice, especially if you’re calling donors and you’re tracking that data, like, oh, we talked to the donor on this day, this is the feedback we got, here’s a donation. You can look at that data over time and say, give me some more insights on this data. What’s the likelihood of then lasting? When should I contact? Again, when is too much? When is it too little? You can start asking questions based on the data, it can give you much more refined results and help you tailor the experience to the donors that you want to target.
Trent Ricker: I think that’s an excellent example. We go from whether we’re responding to a grant and putting together a grant proposal, I think the speed by which we can ingest the grant itself, the proposal then can be researched on so that we’ve got some consistency.
and increase our likelihood of receiving that grant. And again, we talk about sourcing. I don’t think I can emphasize it enough, but that’s a good example where previous applications that were successful – how you represent your organization in a consistent manner – those sorts of source documents would be very well received inside of that AI project or tool. And again, I’m hesitant to use the phrase agent for fear that it kind of can scare some of our audience away. But it is a good example in that sense that the analogy I used –a project or a specific area – that you could keep can be thought of as one of your team members, right? This area I’m going to use as my grant writer partner.
And back to your prompting, the role that you’re looking for your AI partner to play is my grant writing partner, or you are a grant writer, and we are applying for a grant from the Gates Foundation for this particular opportunity. And flipping that a little bit, talk us through how you use or what you might recommend to some folks that are kind of dabbling in different platforms. There’s differences between the strengths with ChatGPT and Claude, Copilot, Gemini, and Perplexity. Talk a little bit about the various strengths of those tools so that folks don’t otherwise get intimidated. I had a question the other day from a nonprofit colleague and they say, it’s all too much. Which one should I use the most if I only had to hang my hat on one? I say, well, there’s probably one based on your use case that would be your go-to, but that’d be like asking the question that in the Microsoft suite of products, you still have Excel spreadsheets and Word and PowerPoint that do different functions, although they do share some of the same elements at their core, but talk a little bit about the those five or more main tools that are that are pretty much used out there now.
Ren John: The tools are changing fast. Every three months, there’s a new model. It’s hard to say, okay, this is the best model, and that’s the best model. I would focus on what’s in your ecosystem. If you’re working in Google Workspace, you can lean into Gemini, and you will always have the latest features of Gemini. If you’re in Microsoft, you can lean into Copilot and get the features there. Microsoft is leaning on Claude and ChatGPT in the back end anyway, so you’re using the same models. Google Workspace is leaning on Gemini, their own AI models there.
There are strengths and weaknesses to each different platform. And as of today, like Claude is building out a lot of new skills and integrations with different platforms. If you’re working outside of Microsoft, let’s say you’re working with Atlassian, or if you’re working with Notion, or if you’re working with tools like that Figma, Claude has integrations.
Trent Ricker: Ren, this is you just flexing right now, right?
Ren John: Different nonprofits work with different tools, but the ecosystem you’re in, they’re going to be building tools. If you have Salesforce as a CRM, Agentforce is right there for you, and they’re trying to give better pricing to keep you in that ecosystem. And when you stay in that ecosystem, you have the security of the ecosystem built in.
The Google workspace has Google and the Microsoft has that. And then Salesforce, its security is built into it. Since nonprofits are the stewards of their data, that’s kind of where I would recommend. You have to be responsible for your data. And because those ecosystems have those protections around your data right now, and they are legally obliged to protect that data and they have contracts there, I would stay in the ecosystems. But that doesn’t mean you don’t experiment outside and kind of see what’s the newest and latest. I mean, that’s what I do because it’s part of my job to stay ahead of these things, but
staying in the ecosystem mainly for data security and privacy. And like I said, these tools just keep advancing so fast. So, it’s not like, oh, ChatGPT is not as good anymore. Three months later, it’s going to be as good.
Trent Ricker: Yeah, it is moved fast. I want to hit on that in a minute because we’re recording this in early May of 2026. And I would suggest that in the last three months, there was a quantum leap that happened in February of 2026. As you mentioned, there were new models that were released and it kind of became an arms war. And as their models have begun to train the next generation of model, it seems like about every 20 to 30 days, they keep outdoing one another. It is hard to keep up. And I think as it relates, I think that’s good advice – stay in the ecosystem that you’re at. You probably have one or two that you’re comfortable with. Most people are comfortable with ChatGPT from a personal level because that might have been one of the earlier tools that they downloaded on their phone and were using from a personal perspective.
I think that’s totally solid. I would say to share with the nonprofit audience that Perplexity is a really solid platform tool that draws in from multiple models, and it’s a great research tool. If you’re researching for a grant, you really want to make sure that the response for that – the grant writing – that research is extraordinarily deep, much more quick, much more thorough that you can get from a tool. Not to suggest that the other tools can’t do that as well, but I encourage the nonprofit leader to try a few different tools. You and I have had this conversation earlier on, I thought Copilot inside the Microsoft ecosystem was clumsy at best. And I think I had a little bit of a file to say, geez, it’s not working as well as I want it to, and I might have gone to another tool. You’d tap me on the shoulder a couple of weeks later and say, they’ve updated it. Go try something again because what didn’t really work all that smoothly two weeks ago is likely going to be much better today. I think that’s true of most all the tools. Is that your experience as well?
Ren John: Exactly. Like, if I was talking today, I would say I like Claude right now because of its integrations with other things and it’s pushing for different things. But like you said, for Plexi, maybe focus on research. ChatGPT just released Image Generation 2.0, which is the best image generating AI on the market right now.
Trent Ricker: I saw that.
Ren John: Copilot, Microsoft’s Copilot, just released Cowork, which uses Claude models, but gets more agentic AI tasks built into it. So, like I said, these tools just keep upgrading and adding on different things. I will say that outside of the ecosystems like Claude and Chat GP do have nonprofit discounts. And if you haven’t heard of Change Agent AI – that’s its own AI model built just for nonprofits. That’s a smaller one that’s trying to do stuff like take out bias and not train on data. So, it has data privacy built right into it and collaboration tools that others don’t, which is kind of nice.
Trent Ricker: Interesting.
Ren John: That’s specific for the non-profit space; they they’ve built that out.
Trent Ricker: I’m glad you hit that because I did want to make sure that everyone should understand that these other tools that you can work in typically extend nonprofit discounts to 501c3. I’ve seen some pretty extensive discounting. I think Claude offers for nonprofits – I think it’s up to 75 percent off for teams and enterprise versions, which is excellent.
It’s a great tool. It’s worth experimenting in. I know that Perplexity has the similar sort of discount to ChatGPT so make sure you’re taking advantage of that. And granted, I think we will see the evolution of more tools that might be plugins or applications that are nonprofit overlays, if you will. It’s a good segue though, Ren, because I think right now – without wanting to get too deep into governance – there’s some commonsense things that people should probably be aware of. And I would imagine that many of our nonprofit listeners out there don’t probably have a formal AI policy. Even ours evolves here at AGP, so there’s some common-sense elements. Why don’t you share some of that common sense? I think there’s some fear out there, but there’s also some common sense about sensitivity and how data is used. Without getting too technical on that, what would you advise our nonprofit leaders as it relates to governance?
Ren John: First, have data governance on your data before AI. You should have governance on how PI (personal information) is treated, how your donor data is treated, how you’re a steward of your data, and that everyone in your company knows that governance – so that when you introduce AI, you have to definitely turn model training off on any AI that you have, and that’s usually available on all paid accounts.
Trent Ricker: That’s a good point.
Ren John: But know that, like I said, the ecosystems like Google and Microsoft have specific language that protects your data and protects your stuff. Whereas Claude and ChatGPT, unless you go to enterprise zero data retention policies, sometimes you have to just make sure that they have the model training is turned off, and you’re not sharing your data with them. And still, I would never put PII (personally identifiable information) in there because they do have, you know, the CYA language into their terms and policies, so, which is expected.
Trent Ricker: Of course, that’s common sense, right? I think I like the way that you frame that in the sense that how would you treat your data if you were working with a partner? A new vendor? AGP is your new agency. We would always work with governance as it relates to information and data. I want to extend that a little bit because if you think about the information that you have that’s constituent data – to your point – let’s make sure we’re very careful about where we’re uploading and how we’re using it. There’s things that I would otherwise consider to be already or soon will be in the public domain. I’m less concerned about, let’s say, as a leader working with marketing material that we might be evolving because that’s intended to actually be front of house. But if I’m working on client work that needs to make sure that I’m not violating any privacy that I otherwise would be adhered to without AI, that to your point, there’s some commonsense elements to it. I think it’s also fair to assume that the aspects that might be proprietary to your organization, board communications, decks, et cetera, you want to make sure that’s more in your paid, closed environment so that that doesn’t inadvertently get out. I think that from my perspective, the advice that I might give, because again, it moves fast. As a CEO, probably my biggest concern would be a team member who, in the right spirit, does work that they think is helping benefit their workflow and efficiency. And they might have stumbled upon a new AI tool that’s outside of our ecosystem. That’s not a bad thing, per se. But to your point, if you’re uploading board decks or confidential marketing material or PII, that’s not a good thing, right? Because if it’s, first of all, it’s a free account, they can do what they wish. And if it’s outside of our ecosystem, you can at least you should default to the fact that that could, if it were sometime in the public domain, how would you feel about that responsibly or as a leader? So, we try to discourage, Ren, I think you could talk a little about this, discourage our team members from going outside of certain realms and tools and if they did want a paid account then we provide that for them with the right sort of justification for their job.
Ren John: And there’s a term for that. It’s called shadow AI. It’s when someone uses AI that’s not approved by the company because they’re trying to find efficiencies in their day-to-day. So, it’s better to get in front of shadow AI, and say these are our approved tools. This is what we’re paying for. And definitely pay for an account and make sure that the data policies are set to not train on your data. That helps protect you and protect your employees from using shadow AI because then they have an outlet to try. And then if there’s clear policies on what data can go in and out, then you’re protecting yourself, your clients, your data, and your employees from that.
Trent Ricker: When you said that, I’ve heard it and it just clicked to me to relate to what was perhaps shadow IT, right? We always, even organizations had to worry about at some point people are going to be using their personal devices to connect with Outlook or to download documents. We can’t and shouldn’t prevent that, but we need to make sure there’s governance around it so that people understand the risks that are associated with that and the protection for that. So again, common sense. As a leader, and I would encourage this to the nonprofit leaders that are listening, I don’t want to discourage experimentation and curiosity. I have to bow.
I was at a conference a couple of weeks ago in AI, and there were some executives there that were very, and that they probably were in perhaps legal roles or HR roles. And frankly, that’s their job so I value those folks on my team. But they felt the importance of governance first and locking down. The reality is that’s not really practical in the way that AI
is moving. I would encourage nonprofit organizations to foster a culture of curiosity and experimentation. And what I see is those that exemplify, like yourself, Ren, when you exemplify smart behavior related to using AI in a way that helps your job, and we share those success stories, it creates further curiosity for people to experiment.
In the last week, I’ve seen two executives build their own agents for particular rote kind of routines in Copilot in our environment, right? But one was created by someone who was just curious, talked about it, and I talked about it in a meeting. Another one followed a few days behind. That’s exciting to me as a leader because that curiosity is creating leadership that’s saying, hey, what can this do to help that when this thing that I do every week or that my team does every week is over and over again and it’s taking two hours to compile that information and create the report. What would it look like to turn that into a function where I could feed the information or tell it where to find the information and give me that first cut look? And we’re seeing that here inside of AGP.
Ren John: I think we’ve done a good job of providing a safe playground for people. Like I said, when you provide those paid accounts and when you provide guidance and the governance and the policies that we have, it’s given people that freedom to explore and experiment. And for nonprofits that don’t know how to start with policy or governance, I would highly recommend going to NTEN’s website. They have a lot of resources for starting with the AI policy and governance that you can get for free.
Trent Ricker: Great suggestion. That’s a great piece of advice. The trade organizations that many of you are probably already a part of are ahead of you, and they’re using their part of your membership to help guide you through these uncharted waters.
I want to shift a little bit as it relates to flipping the script. Now, as we as consumers use AI,
I think it’s an increasingly likelihood that we’re already seeing it in search results, right? They’re dominated by AI results. I think there’s an increasing likelihood that as AI assistants or the constituents themselves are researching causes or comparing your organization against somewhere else or maybe making recommendations as to where to give, that has an implication for your website related to transparency, credibility, and how you’re explaining your mission and impact.
I’m of the belief that more and more of those paid SEM and the organic SEO searches where people who may be mission recipients or who may be constituent supporters that may support you through grants, personal donations, volunteers, events, et cetera. So, talk a little bit about that, Ren, because part of AGP, we do web and technology development, and we’ve had to integrate that quite a bit lately. Share that with the audience because they might need to think about that.
Ren John Especially with websites right now, web traffic is down across the board mainly because people will ask questions to AI, and they will get the answers right there. We want to make sure we’re producing good content. The same rules of SEO apply, but they’re more important now. And now we want to make sure we’re producing schema documents for AI to read and be able to ingest your content. And the quality of the content matters. There’s the EEAT model (Experience, Expertise, Authoritativeness, and Trustworthiness) from Google that has been how they rank your content is going to be becoming even more important. And we do a lot of AEO (answer engine optimization) or GEO (generative experience optimization) audits of content now. And then we’re seeing what pages that do get referred from AI. Like if you get cited and someone clicks through and we see what pages are, those are the ones that are bringing your content up and we’re like, hey, we need to review this page and make sure it is up to date, it has the information that is relevant, it has all the keywords, and it’s got the authority. It’s got the author, the date – everything on there that makes it show up for AI overviews. So that’s a service we provide is just
kind of reviewing your content. And it’s just something you can, if you need to do it on a regular basis now, because your content becomes a lot more important.
Trent Ricker: That’s right. I think it’s fair to say that we’re heading towards a world where, you know, nonprofit organizations’ websites need to be understandable, not just to humans, but to AI agents. They’re the ones that are going to also be searching on behalf, donor advised funds, et cetera, as it relates to making the recommendation. I want to be careful. It’s not about chasing the shiny object or rebuilding your websites for robots, right? It’s about clarity. It’s about structure. It’s about discoverability. It’s the same best practices we might otherwise recommend when we’ve been building websites for the past years as well. But as it relates to your technology and just making sure that your organization’s story, and the impact that you make and the donation pathways are easy to understand in an AI mediated world because at some point that element – it’s already possible today. Somebody may make a donation through their AI tool of choice as opposed to ever even touching the site. That’s possible today. Not a lot of folks are using it that way, but I could see that happening where we have to have the right information on behalf of the website, on behalf of the nonprofit organization and be able to still allow them to register for an event, volunteer, make a donation, or as a mission recipient, where can they be taken care of for their family members?
Let’s kind of close a little bit on thinking about what are some practical ways that our nonprofit friends could use AI if they aren’t already. Let’s think of a couple, two or three key takeaways for leaders to do in the, let’s say the next 30 days.
Ren John: The next 30 days – I would say every nonprofit deals with spreadsheets, and that is something AI does so good. It can analyze the data for you. You can ask it for insights on your data. You can ask it to organize, clean up, and reproduce that data, like building key takeaways for your executive leadership from that data. Take that data for different audiences, like, what do we need for our program managers? What do we need for our executives? What do we need for, you know, our volunteers to help them? Can it help me really learn more about the communities we serve? Because if we can get more insights into our communities, we can serve them better.
Trent Ricker: That’s great. I’ll offer a couple up to our leaders specifically, from leader to leader. First, I would say I had to change my thinking, and this was probably about six months ago, and it was a slog until I think February when some of the tools were really doing. But I had to think every day, where can I try to use AI within my regular workflow? Where might I ask a question of AI or explore research with it because it wasn’t front of mind for me. I am a creature of habit. I have been in a senior leadership role for a long time. And so, as I was preparing for a board meeting or preparing for a client visit or reading through our quarterly business reviews for our clients and working with our team for a leadership meeting, I was doing it the way that I’d always done. And I had to pause and say, how might I use some of these tools to make that better or more polished or inform me? And it was transformational, but it takes some time. So. persevere.
I think the other thing I do is find your Ren. There needs to be an internal AI champion. I think Ren, your partnership here at AGP is critical because people know that they can go to you and there’s no question that’s a dumb question. And then a flip side of that, you’re championing when people are finding a really cool use case to be able to amplify that throughout our organization. Whether that’s through lunch and learns or the email blast periodically or showcasing where people are using AI in a way that might then trigger curiosity. So, stay curious, figure it out in your own workflow because I think we as leaders are going to set the pace. If you are in an organization where your leadership is talking about, you know, don’t do it, or I don’t understand it, or those are actually places where leaders are going to have to pivot or they’re going to start being replaced. There’s just, there’s proof of that in the Fortune 500 world right now. Lots of CEOs have been replaced lately because of the pace by which those organizations need to go. So, you as a leader with your senior leadership and your board need to have a commitment that this is not something, as we said earlier, that is optional anymore. It’s something that we need to be learning as new tools.
Any other thoughts from you, Ren? I’ve got a few to close, but I think those are some pretty good things to give some practical applications.
Ren John: I agree with your sentiment. I think about how can I use AI for this? Like, that’s now my first thought. Can I use AI to do this or do I need to do this? And that’s really helped me start using AI more for different tasks. I didn’t even think about using AI for like a LinkedIn post – it reviews mine for now. And I was like, oh, that made it better English for me. That’s good. And another thing is you’re talking about leadership and buying into AI. When you share your story about how you use AI when it comes from the top, it really sets the tone for the entire organization that, hey, okay, AI is something that we are using. It’s not like a mandate. It’s more like this is something that we share. When you share those stories, it really makes impact. Then people can start getting creative and thinking about how can I use AI to make my life better? I mean, our boss is using it, maybe I can use it. And then once those stories get shared – like you were talking about the show and tells we do here – people in different job functions are sharing different stories, but that are impacting other people to spark ideas of how they can use it in their work.
Trent Ricker: I love that. I believe as we close here, I think it’s the most exciting time in business that since websites were introduced and the internet was part of how we would work. And that took quite a while actually to adopt overtime. This is happening at a much more rapid pace. It’s an extraordinarily exciting time.
So, AI adoption in nonprofits is already happening, and it’s happening in companies everywhere. The opportunity now is to make it more intentional. And the next phase is not really learning how to prompt better, but it’s learning how to work better. It’s a practical, responsible partner as part of your mission. Asking better questions is the appropriate way to go about this tool. There was a point by which when we all were not fully embracing email, for those of you that are old enough to remember that, and it was a new, novel thing. Now we can’t imagine working without it. It’s going to be embedded.
Ren, I thank you for your leadership at AGP. And if any of you have any questions or would like to banter some ideas around in the spirit of what we do here at AGP, we love to banter around and have conversations and try to help you out. There’s certainly some things that we are doing to help our clients out, but we are also a company that’s applying these principles and this technology so look us up. Ren or I or anyone on our team would love to have a conversation with you. Ren, thanks for your time today and we’ll look forward to having another episode soon.
Ren John: Thank you.
