Name: AI for Philanthropy: A Leadership Conversation with Blackbaud and McKinsey
Uploaded: 2026-04-23T18:31:27.390Z
Duration: 59 min 26 s
Description: AI for Philanthropy: A Leadership Conversation with Blackbaud and McKinsey

Transcript for "AI for Philanthropy: A Leadership Conversation with Blackbaud and McKinsey": I'm Carrie Cobb, chief data and AI officer for Blackbaud, and I lead our data intelligence center of excellence. And I'm responsible for our enterprise data and AI strategy, including vertical focused data product innovation, data science and engineering, and data and AI governance and literacy. I also have the privilege of serving on the, social impact sector chairs of the advisory council on methodology for the Giving USA Foundation. I'm a member of the governmental affairs committee for the Nonprofit Alliance, the advisory council for fundraising AI, and the National Artificial Intelligence Association. So as we begin today, it's really important to recognize that we're really living in a defining moment for AI, but we're all feeling it. That pace of innovation we're seeing is unlike anything that most leaders have experienced in their careers. New capabilities are emerging faster than organizational processes or governance models and sometimes even our own comfort levels can keep up with. And for many sectors, that movement, this moment is about speed and scale. But for philanthropy and mission driven organizations, it's actually about something more. In this sector, leadership in the age of AI isn't simply about adopting the latest technology. It's about trust. It's about responsibility. And it's about ensuring that as AI becomes more and more capable, it strengthens human judgment, human relationships, and mission outcomes. And so that distinction matters deeply in the social impact sector where trust is the foundation of every donor relationship, every constituent interaction, and every outcome we seek to achieve. So today's conversation is intentionally framed as a leadership discussion. We're moving beyond the question of whether AI will impact philanthropy. Instead, we're going to explore how leaders can introduce AI with intention, how purpose built AI grounded in deep sector expertise and real world transformation experience is being embedded directly into mission critical workflows to unlock insight, free up time, and expand capacity, all while maintaining transparency and accountability. So to help ground this conversation, I am joined by leaders from both Blackbaud and McKinsey who bring both sector specific expertise and a broad vantage point across organizations navigating this transformation at scale. And together, we're gonna share practical perspectives, that's a hard one, on what responsible AI leadership looks like in practice. We'll talk about what's working, what's hard, and what leaders should be prioritizing right now. So my hope is that everyone leaves here today not just inspired by what's possible, but maybe clearer on the role that we all play as leaders in shaping how AI is introduced, how it's governed and used in service to our missions. So with that, let's begin. So my lovely panelists, if you join me up on the stage here, I thought to get us started today, I would ask each of you to introduce yourselves, of course, but for fun, share your favorite AI moment. And, Paul, I thought I'd start with you. Thanks, Carrie. So, I'm Paul Goldstein. I lead agents for good program at Blackbaud. I'm the head of AI growth. My favorite moment, it's actually been really recently. I got to play hero when I had some friends visit in San Francisco. At the end of last year, they were in town. They hadn't been in San Francisco many years, and there was a restaurant they really wanted to go to. We looked online. It was last minute, there was no reservations available, and I did a last ditch effort to call the restaurant. Usually something that stresses you out because you're gonna call somebody who's running around, has a lot going on, and likely doesn't know the answer to your question. I called, and that was the first time in my life I was greeted by an AI agent. I was like, woah. What is going on? And it asked me a question. What are you looking for? I said, hey. I'm looking for a reservation for four. Between this time and that time, anything is great. It quickly took a second and very clearly checked the back end and how the table configurations were. And the same inventory that didn't show online, suddenly, it was able to offer me a reservation for four at a specific time because it was able to see that it could move some tables around or make something work. It booked the table and sent me a confirmation email. What a wonderful experience that was for me and my guests, obviously. I was a hero. I got to take them to a restaurant they were excited about. But also the amount of stress that probably removed from the person that would usually be answering that phone call because they're that person's job is not just to pick up phone calls. That person's job is a lot of things at a restaurant. And it was just it was a great experience. We were all happy and probably ended up giving much better tips than we would have even given at first because of that experience. Yeah. I love when AI can create even more human connection. Right? You could spend more time with your guests because you are able to quickly go through that that reservation. That's fantastic. Daniel, my friend, will you please introduce yourself and talk about your favorite moment with. AI reason? Yeah. And then first, Carrie, I just wanted to thank you and Blackbaud for this opportunity. It's wonderful to be here on the virtual stage with you talking about AI and the impacts and the benefits for the social impact sector. But my background, I've been working in technology risk and resilience and cybersecurity for the last fifteen years or so. Work with organizations around the world. And lately, been very much focused on AI. How do you build AI agents and how do you manage them in a responsible way? My, favorite moment goes back to last summer in August. I was sitting in my backyard very much wanting to go to the beach and then enjoy the gorgeous weather, but I was under a deadline to submit a paper to a cybersecurity conference. And in the old days, that would take hours of methodically writing out, the responses to the questions that were posed and thinking through what I wanted to say. Here, I was able to take a couple 100 pages of material that I had, upload that to GPT along with the questions, and it instantly produced a what was a pretty good first draft. And I didn't submit that directly, but I was able to use it, edit it, massage it, make sure it reflected my voice and what I wanted to say. And then, probably about two hours, I was finished and then enjoying the rest of the afternoon at the beach. Yes. That extra beach time. I love it. Mhmm. Cecile, my friend, I'm so glad you're here with us today. Introduce yourself for this audience and then talk about your favorite AI moment. So thank you, Carrie. Also delighted to be here. It's such a pleasure. You know, as as as the people on the call know, you know, AI risk is, is really something I am focusing on right now and am quite passionate about. So, it will be it will be difficult to contain this in just one hour. But if I just briefly, introduce myself. So I'm an associate partner at McKinsey. I'm actually based out of Europe, where we like regulation a little bit more. So, focusing on that, in McKinsey kind of leading, everything AI trust, in Europe, EMEA region, working really across sectors on this topic. So working with some of the more regulated industries to also the less regulated ones on, you know, helping them think through how should they govern AI. My background is also risk management. I've been doing risk management, for for many, many years both, you know, at McKinsey and and and, you know, outside, in the industry. So if I just share my favorite AI moments. So, as as most of my favorite moments, they go back to, moments with, with my children. And, also, my favorite AI moment, it was with my children, with my nieces and nephews. I you know, outside of work, I, I try to be a little bit less serious. And typically, when I'm with them on holiday, I joke around that, you know, I'm a world champion in in various things. Right? My children are young, so they believe I'm a world champion in catching crickets and in some war, in case you're familiar with that game. They never really believed me, so I decided to instruct Chad GPT to tell a story about myself and how I'm famous. And, you know, everyone in London knows me and, you know, ask for my autograph because I'm this world champion. And that's what it did. Right? So the children read the read the text and then, you know, their eyes went wide open and and, you know, they were like, oh, oh, so you it was true. Oh, really? You're famous. So it it was a really funny moment. But, of course, it made me also realize, you know, the dangers of AI because obviously just because it was coming out of chat GPT and it was written in such a nice way, you know, you believe what was written there, which was obviously totally made up. So, that was a good moment, funny moment, but also a moment that, you know, made me think a little bit. I think I'm handing it back over to you, Carrie. Yeah. Yeah. I think every parent in the room is going to take that, as a to do with their kids is to have, have Chat GPT tell them how famous they are at something and, see see how that goes. I'm so excited to have this panel here today. I have the privilege of working with these folks on a regular basis. And I can't wait to share the experience and the expertise with you all today. So we're gonna just open it up for questions and then at the end, we'll have a Q and A with all of you. So I thought we would start by setting the stage with this group. Right? Many are calling this a once in a generation AI inflection point. From where you you sit, what makes this moment so different than any of those previous technology shifts? And, Daniel, I'd love to start with you. Yeah. I what I'd say is that just the the speed and the rapid innovation is what we're seeing. The fact that you can create something these days that before might have took you hours or days and instantaneously. And the tools that we have are Mhmm. are creating outcomes that are nondeterministic. So in the past, if I'm doing a, Google search for information or if I'm trying to, create an image, it's very much the fixed outcome. Certain steps lead to a certain outcome. But now I can create an entire report and presentation from scratch. Just even talking with Paul, before we're getting the session going, Paul was talk telling me about all the different presentations he was building or earlier today, building out some images, using AI. And it it's that speed and innovation. And I think back over time, the the last time probably you're in such an error was in the early days of the Internet boom, the .com error, even early days of cybersecurity. And this is very much the next s curve and that next evolution, but at a much faster pace with, new risks that we have to think about and effectively manage. That's wonderful. Cecile, can I ask you to answer that question as well? What makes this moment so different. from all of those other technology shifts that we've seen? So, I mean, it's been said before, and I echo what Daniel said. Right? It's the it's the pace of change, but also I think it's the breadth of change across industries. You know, I I spend a lot of time in financial services. You know, automated decision making is something we saw already probably, you know, twenty years ago. If any of you are familiar, there is something that's called algorithmic trading. This is when you get, you know, automated decision making to, you know, make money out of super small price movements. Right? This you know, back in 2010, you know, this this resulted in a huge crash, the famous flash crash and a drop in the markets and because something had gone wrong. But I think it was, you know, a specific use case, you know, in the banking industry where we were so advanced. Now the automated decision making is basically, you know, spreading across sectors, Yeah. you know, at at also at a very rapid pace. So I think for me, it's just the breadth and the pace. So kind of the you know, and across both the opportunity and then the risk side of things. And and, Mhmm. system, maybe I just add one thing, which is also the tools are available. They're out there. I can go online. I can use GPT. I can use Cloud, and it puts the power much more in the hands of the individual, to create these capabilities. So before I might have needed a lot of computational power, access to tools that I didn't have at hand. The the world has changed in that sense. Yeah. These tools are democratized. They're accessible. to everyone, which I. think, Cecile, it reinforces what you were saying here, but then even in your your fun AI story is that it's so accessible everyone can use it. That's great opportunity, but then it also presents a whole lot of new, risk as well. So, Paul, let's now think about the social impact sector. What does AI actually make possible for nonprofits or social impact organization, that wasn't possible before maybe? You know, I wanna kinda go back to the previous question you had for folks as a tee up to this one. I think Daniel and Cecile had such great insights. One of the things that's great about right now is we have so many tools. There's a democratization. of education of these tools. You can go and find videos on YouTube, podcasts with experts that are using and building these tools. They can guide you on how to use them. So I think that we're in a very special place because in the past, when the shifts like these have happened, you'd have to go and look at, where do I go? Who do I ask? A lot of this, is now a lot more available and available for free, which is which is awesome. So you can get started and learn from there. When I think about the shift in the social impact space, there's a few things that feel very important to me. One is now you can maintain personalized relationships at scale. You don't have to have a person to person relationship with everybody. You obviously want to, but you have the opportunity to scale those up with AI. Two, it's you have the ability to have knowledge survive turnover. With AI, documentation is available. You can document everything that's happened. Things don't leave your workforce. It also makes it a lot easier to as you bring people in to educate them on what's happening in your organization, what has happened in the past, what do constituents care about. And then three, you become a lot more proactive in execution, less reactive around, oh, this happened now. How do I figure this out? AI allows you to be a lot more preact proactive, especially as it relates to tools that are becoming available like the development agent that we've launched. Thinking about things as they come up, not waiting until those things are raised. That's a big, big shift. I love that. And I'm gonna I'm going to maybe create a little bit of fun here. You brought up favorite podcasts or learning opportunities. I'd love for everyone to throw in the chat. Let's learn from each other. Type in what your favorite podcast is. Type in what your favorite learning opportunity platform, where you're going to consume all of this new material, the new behaviors, thought processes that we all need. Let's while we continue to go through some of these questions, put those in the chat and we can all learn from each other throughout the the hour that we're spending together. Cecile, with all of this potential, right, when we're thinking about just the industry at large, we're thinking about the social impact sector, you know that you and I can never be in a room together without talking about responsibility. It's where we we always head when we're together. Why does responsible AI matter so much, especially for this sector? Thank you. That's an excellent question. Responsible AI or you know, we also like to refer to it as AI trust. And I think trust probably this particular sector is is something that, you know, should resonate a lot. We see it as the multiplier of the impact. Right? So when we do our AI transformations with our clients, you know, we spend a lot of time on getting the technology right. We spend a lot of time on getting the data right. It's very tricky problem. We spend a lot of time thinking through the, you know, the people side of things. What are the skills? You know, how does the organization change? But then what we realize is without the trust, without the responsible AI, without thinking through how are we going to govern this and how are we gonna stay on top of the risk, there will be no impact or that impact will not be as big. Right? So I think the importance of AI trust is it's it's an enabler of, you know, grabbing that big opportunity that we talked about. And, of course, trust is is something, you know, that is is something that when you have it, you need to protect and safeguard. Right? Because trust is takes time to build but can be easily lost. So this is kind of our core conviction, and this is why we think AI trust is, you know, a key component of, you know, any AI transformation or any AI product, that you build. And what it means in practice, I mean, I I think we'll delve into a little bit more in detail. Right? But what it means in practice is that, you know, you put in place the right safeguards, the right controls to make sure that, you know, the output that you're getting from the model is sufficiently reliable. Right? The output is, Okay. sufficiently checked. There is a clear accountability. So it's kind of focusing on all these practices, again, that allow you to kind of capture that opportunity. I love that. I I do love what you're saying there. The AI trust, it's it's once you have that, you need to maintain it and protect it. And so I'd love to dive in with this group around responsible AI, in particular and really making sure that we're defining that. So, when we say responsible AI, you talked about, you know, what does it mean in practice? Can you walk us through what what those components actually look like? Give us some some additional detail Yeah. So. happy to to to break it down a little bit. Right? And, you know, as a McKinsey person, we have our framework around this. Right? So we've tried to break it down. We've tried to, you know, use that also to understand the state of the industry. But, essentially, there is five components to it. And I'll just quickly walk you through. Right? There is the strategy part, which is essentially around the underlying principles, Typically, the starting point for, you know, any organization to build out, you know, their AI governance approach. In this bucket, we also put things like a taxonomy of risks, very crucial foundational part because it allows everyone in the organization to kind of speak the same language. What are the risks that we are worried about? But also who is then going to be in the organization is going to be responsible, right, to opine on what's an acceptable level of that risk and how do we control the risk. So that is part of the strategy pillar. If we go into risk management, you know, there's a few components in here. This includes, for example, how do we manage third party risk. Right? Because, obviously, in most organizations, we would be relying on a lot of third party solutions and APIs and data. And so kind of how do we control that risk, right, if if something if something is wrong or if there is an incident. We have the technology side and data, which means ensuring we have good data quality. Right? We're conscious about any bias that we're introducing. So making sure we have all the right protocols and tooling around that. Fourth component is around the governance. You see there's a lot of, things in this bucket. This is around, you know, how do we manage the various responsibilities in the organization around ensuring that we have that AI trust. And then lastly, we have this separate, because this is kind of a little bit new. How do we think about agent governance specifically, right, with agents now being basically, you know, AI systems that can take decisions, not just kind of create output? There's a whole additional layer of how do we control the agents. You know, how do we make sure we have an inventory, you know, where we're introducing, you know, agents as a starter. So, this is kind of when we talk about responsible AI, it's all of these things that we are thinking through and that we kind of try to build out. And I would love to go also back to you, Carrie. You know, if, you know, at Blackboard, you know, looking at this framework, how how are you thinking about it and kind of how are you how are you kind of, building out some of these components? Yeah. That's a great question. Thank you for sending it back to me. This is fantastic. You know, when we think about responsible AI here at Blackbaud, we really think about how we're building and using AI in a way that earns and maintains that trust, with those that are using the solutions, but then even extending that to the communities that each and every one of these audience members serve each and every day. And so in practice for for us, it's not about a single policy or a checklist. It's really that operating system that I think you so beautifully showed in that last slide that really spans how we think about designing and building and deploying and governing and continuously improving AI across the company. So we, you know, similar to to how you have in your slide, we prioritize the strategy, really defining our intelligence for good strategy, Our responsible AI principles, I encourage everyone to go to the Blackbaud site and take a look at both of those. And then we prioritize governance that's usable. Right? Not theoretical. Really making that distinction. And so we start with a practical, responsible AI operating layer that's aligned to NIST, AI RMF. But the key isn't always just the the framework itself. It's that, making it really practical. So AI impact assessments embedded into product life cycles, clear risk tiering, defining decision rights, all of those. It's not always easy. Right? I think whether it's AI or whether it's all of these technology shifts that have become that have come before us, the one that's proven the hardest is the change management. Right? It's not the technical piece. And so, I think making sure that you're defining principles, talking about those long standing behaviors around decision making, risk ownership, trust in technology. AI is compressing timelines and raising the stakes and and how do we think about change management and culture, when we start to bring some of these frameworks into place. Perfect. Thank you. so much. Much. Thank you. So let's think about, you know, when we think about our responsibility here, there's always something that I think about in this moment. When we all talk about agents, Daniel, Are all agents actually created equal? What really separates purpose built AI from general purpose tools when it comes to mission driven work? Can you talk us through that a little bit? Yeah. No. Definitely. And it's a great question, and there's there's so many terminologies out there, so many different tools, technologies. And when we think about AI specifically, we like to think about more first, more traditional AI. So you can imagine machine learning, deterministic, things that were around years ago. These days, we're much more into the generative AI. So you can have, for example, your ChatGPT chatbots that you use to get information. But more and more, we're shifting to an agentic world where I take nondeterministic AI and actually give it tools, give it the ability to interact and do things. And I think about the different spectrums of agents. They are not all created equal. On one side and one end of the spectrum, you could think about knowledge research. If I'm, for example, working in a nonprofit organization and I wanna do some research on a donor or do some research on trends in the industry, I'd have an agent note of prompts either leveraging information and data that I've got or leveraging information and data that's out there more broadly, to come up with a report, can provide it in a structured format or a presentation as we were talking about earlier on. On the other end of the spectrum, you've got agents that are much more focused on completing tasks even to the extent autonomously. You could think of this in the case of, of mimicking the behaviors of an employee. These days, for example, fundraising is a popular area where agentic and AI can be leveraged rather than, for example, you have a few people in a development office, that's working in a university with a 140,000 alumni just as an example, and maybe they can contact a few thousand in a in a given year. Here you've got a situation where now I complement that with an agent that can go out autonomously, reach out to individuals, raise money, custom craft messages, even have virtual personas that you can interact with and ask questions. And that's really the direction of travel that we're going towards, not just one of those, but even, groups of agents, workforces of agents. So thinking about those in terms of a spectrum, each of these have different levels of autonomy. Each of these have different risk profiles and different types of controls that need to be put in, based on that type of risk. I'm gonna have different controls and level of those for knowledge and research agent for those as more of an agent that's mimicking the behaviors of, say, of an employee within an organization. Okay. So let me double click there. So if if organizations are starting to bring in all of these varying levels and types of agents. Should an organization actually have a formal AI code of conduct? And if so, what should we all be putting in it? Yeah. And it's it's a great a really great question. It's still early, but more and more we're seeing organizations publish, codes of conduct that are specific to AI. And you can imagine including there what are the things that we would be doing with AI? Where are we using AI? Where are we making explicit decisions not to be leveraging AI? So for example, in the fundraising example I gave earlier, being able to say, hey. We're using AI in those situations, but there's gonna be a human in the loop as part of that campaign might be, in one organization's code of conduct. Also, being able to comment on transparency. If, for example, there's content in an email that's being generated with AI, that organization might choose to disclose in that email that that content was generated or or leveraged AI. And that's to an organization in terms of where they're drawing that boundary of what they're comfortable doing, versus not comfortable doing, but being transparent and open and visible in terms of that good conduct. And to me, the theme of the AI code of conduct and what you're saying goes back to, Cecile, what you were talking about, which is AI trust. That trust that we once we have it, we need to maintain it and protect it. And so, Paul, walk me through, again, this entire audience. All of our social impact sector operates on trust, whether it's from donors, from beneficiaries, from their communities that they serve. How do we ensure that AI strengthens that trust rather than You the trust? know, the core at at the core asset of social impact is trust. It's really, really important, especially with AI. We need to be clear on where AI is used, where it isn't used, and who remains accountable for it. Letting people know that you're using AI to further your mission and you're doing so in x ways or y ways and how the humans are in the loop driving this. I'll help people understand. I'll take the example of, you know, we had some very, very early adopters of development agent, last summer. And some of those customers, you know, they mentioned writing, hey. This AI agent is reaching out to you to get to know you. Caused a few customers to say, hey. Why are you doing this? Not necessarily angry, but being curious. Why are you doing it? And that started a conversation for their organization where they were able to share, here's what we're trying to do. We're trying to reach more donors. We're trying to reach more constituents and make sure that they know what's happening in our organization. And almost every conversation that I've heard, that has started that way has led to positive momentum, both for the organization and that donor, where they were excited about the innovation that was happening in an organization that they were clearly excited about and clearly had a vested interest. Love that. I love the moment of curiosity and dialogue and discussion. It's probably more important now than ever Yeah. within this moment that we're all in. So also take me into one other aspect that our social impact sector is feeling, which is we're already so stretched thin. How do we then take this technology? How do we implement AI without creating one more overwhelming technology project, one more thing to add to the list? How do we do that? Well, you know, I think the best implementation starts small and very specific. Choosing a goal, something that's a workflow or a bottleneck in your organization that you can measure the impact of with AI. Plus, it helps build trust with your staff. If you provide relief for them, not disruption, you're not taking people's jobs. You're helping them be better at their jobs. You're making some of those really annoying tasks that are the reasons you stay up later that day. You're not able to have dinner with your kids become the reason they're actually okay. I'm better at my job. I'm faith two and I'm spending a lot more time with customers, with constituents. I do think there is some work to be able to be done as, you know, especially as we get into agentic AI, like, making sure you follow the right guardrails and maybe choosing a provider that you're gonna use for those agentic, pieces so that you're not scattered. So you're not focusing on one piece or another piece and figuring out how they all touch together. Simplify it for your workforce. Simplify it for your organization of how you're gonna do that. So, Daniel, let me pull you in. So if we think about what Paul is saying, we're gonna start small, we're gonna find those pain points, give folks, you know, time to go home early and have dinner with their families. Right? How are we taking that burden off? Once that organization is up and running, we've got some of these pilots that are deploying, how do we know if it's working? Is that it like, how do we measure. success? Is it time saved? Is it dollars raised? Is it something harder to quantify that we're still working through? What do you think? No. It is a great question, and it's one that organizations are working through constantly to figure out, are we having the impact and are having the, the the results that we are looking for? Many organizations are deploying AI, but many are not necessarily seeing the benefits. Comments, for example, such as I see AI everywhere except in my bottom line are prevalent in the industry. And it's it's very much if you're using AI, it shouldn't be AI for the sake for AI, but a specific outcome. So you could think revenue generation, cost reduction, efficiency, or as Cecile, would like to focus on risk management and risk reduction as specific outcomes. And you can imagine in a not for profit organization, the social impact sector, were we able to increase our fundraising? Are we able to actually reach a larger number of constituents, than we weren't able to do before? Are we able to, develop grant applications much faster? Mhmm. And, actually, potentially, you see even a a positive result in the percent that we're winning or the the money that we're able to get, from that from those contacts. Are we able to help launch, events faster, than we could previously. And it's really a magnification of the work that they're doing and through that line seeing what the impact is. So it's almost like if you think about the outcomes that you're trying to drive, Yep. the technology and the methodology that's appropriate will follow. Exactly. Think about the outcome you're trying to drive. Think about where AI could leverage that, could be an accelerator, and think about maybe where you wouldn't wanna use it, where you're not necessarily seeing that differentiation because then that allows you to focus on the areas that you do have impact and gain greater benefits from. Yeah. So, Cecile, I'll ask you this. So, you know, Paul talked about finding those pain points, starting small, Right. scaling. Daniel talked about how do we know if it works. How do we then think about the responsibility and trust that we all carry? How do organizations adopt and scale that responsibility piece as well? Thank you, Carrie, for the question. And I think I just wanna pick up on what you said before, Ryan, that you are doing at Blackboard. It's, you know, staying practical, and to which I would also add also staying risk based. So we, you know, our our belief in kind of how we are supporting organizations on this is to make sure that their responsible AI journey, kind of scales, in alignment with the AI strategy. So in other way in other words, it's, you know, making sure that, you know, your governance, your controls, they're all calibrated in light of what are the types of AI use cases that you're really doing. Are you still an experiment, you know, in a kind of in a phase where you're experimenting when you are mainly doing, you know, use cases that are internal, they're not affecting your customer or clients, or are you actually already doing more complex stuff, right, with multi agents, you know, chained together, taking decisions, and potentially, you know, independently kind of, getting in touch with your customers or clients? These are entirely different things. So the way we approach this is to really calibrate the approach, to kind of, you know, which which are the AI use cases that you're doing and making sure we stay very practical. So we discuss the principles, for example. They are a very strong foundation, but by themselves, obviously, they aren't practical. Right? So they need to be translated. And the way you do that is, you know, you could take the principles, you translate them into what are so, therefore, the key risks that we want to avoid, and then how do we size that risk? How do we measure that risk? And, therefore, what are the practical controls without a practical checks that we want to see in place? So, again, I think in in practice, it's all about, you know, practicality, risk, you know, being risk based and making sure you are keeping the same pace in your governance as you are in your AI journey. I love that. If we then think about I love the chat. We've had the chat going of what everyone's favorite podcast and learning opportunities have been. We also need to think about our organizations and how we build for for what's next. So Paul, let me bring you in and ask what capabilities should leaders be building in their teams, not just, you know, the podcast or the, the learning platforms, but, what should we be building in our team so that we're ready versus catching up? Yeah. That's a great question, Carrie. I think first is AI literacy. We don't want everybody to be a data scientist, but we need to help our teams understand what AI is good at, where its limitations lie, and how to question the outputs. It may not get things right at first. Second is strong data foundations. Like, we need to make sure our data is clean, covered, and trusted. We as humans can't do a good job with things with bad data. We make mistakes. AI will as well. It's not reasonable to expect it to be able to circumvent bad data. And then three, it's really making sure that, we're we're reinforcing the responsible AI piece, clear ownership policies, and escalation paths. So teams that feel comfortable using AI in real world situations. We need to make sure that people are there. There's a lot of questions. We're all concerned. We're all concerned about mistakes. Being thoughtful about how do we approach it from the beginning is really important. Mhmm. I love that. I and I'm I'm seeing a thread come through this conversation of curiosity and courage and maybe some patience and grace along this this journey as well that's kinda coming out in each each, kind of question and and response there. Okay. So Someone asked a fun question to wrap us up because I know we're we're gonna have a lot of questions coming from this audience. I'm so excited. I would love to know from each and every one of you, if you look three to five years out, which I know is so hard right now, now, but if you look three to five years out, what becomes possible in the social impact sector, that simply isn't here today if we get it right? What's possible? So, Daniel, I'm gonna start with you. Terry, thank you for the question. I'm super excited about the art of the possible. And if I think the speed at which we've been moving, where three years alone will be here, what I envision and if you think today, we're very much in the single agent world. We've moved from generative and now we're focused on agents that can run a fundraising campaign for me. But I'm super excited about this multi agent world that's coming over the. horizon. And I can imagine a not for profit that's got their annual fundraiser, their annual charity auction. And rather than a team of people managing this, imagine an agent that's out there booking the event and the venue for you. Imagine one that's ordering and arranging and picking the menu and the catering. Imagine one that's, ordering the entertainment and booking that, arranging the speakers. Another that's actually running the entire auction during the event while where today it's a person and it would actually be be an agent. And then imagine, one, your your event planner coordinating and managing all of those agents, except that event planner is an agent as well. And you are at the organization is interacting with that agent as the event planner. That is unbelievable impact in the world that I hope I will see and expect to see in a couple of years. Yeah. And you're getting back to the human connection. If you have. all of that, I saw the chat. We get to go home and have dinner with our families whether. they're human or cat families. Yeah. I saw in the chat, like, that's what we get to do when we. do this right. I I love that. That, thing. Carrie, just on the point for human, I would say it's still human in the loop. It's not running. entirely autonomously. At each point where, hey. Before that venue is being booked, yeah, I'd like to see that bill and and sign off on that. Or before the entertainment's, organized, yeah, let me sure make sure I'm comfortable with that band. And that goes back to the notion of responsibility to Cecile's point of making sure those checks are being put in place, making sure you've got approvals at at each of those steps, which is what you'd want and to be a part of that. Yep. Cecile, what about you? I see a couple of, you know, I think a couple of key trends, right, that I would expect to develop over the over the over the medium term. So I think one is just about how much you can get done in a day. Right? I think we talked about it also in our personal anecdotes. Right? How much faster we can be at this. But, you know, imagine that at scale. Right? So and imagine what that means also in terms of how we can kind of serve our people and, you know, how much more time we have to kind of interact with people and kind of serve them in a in a personalized way. So I think that is one of the key trends I see. And a other trend is, you know, with the data and with the AI systems, we can be so much more proactive. So. thinking a little bit about the social sector and, again, drawing a little bit of my experience in banking where, for example, we see a movement to real time fraud detection. I would imagine social organizations will be able to be much more proactive also at anticipating risks. Right? So dropout, homelessness, health issues. These are things that with AI, you can just predict so much better. Right? And therefore, that you can, you know, intervene, much earlier. So I think these are just a couple of, you know, big big trends that I see, big opportunities that I see over the medium term. I love that. Paul, what do you think? You know, I'll piggyback on both of Cecile and Daniel's answers. I think I'm really excited for the amount of productivity we as humans can have and how this multi agent workforce or teammates can help us get more done. I really do think that we're gonna get to a point where small nonprofits, small organizations of three or four people can be able to achieve and have a level playing field with organizations 10 exercise, 100 exercise. Be if they're able to use and get access to these tools. I think it's really gonna level the playing field here and give people the chance to not just reach the ambitions they currently have, be able to set even higher ambitions for what they can do with their missions and organizations. But, Carrie, you can't escape this one. I'm gonna pass it on to you. What do you think is gonna happen? Yeah. Thank you for that. You know, I I think if we get this right, I think we can unlock more generosity. Not just, you know, in addition to that productivity that you're talking about. I'm always thinking about how do we actually impact generosity. So, you know, I always think about how AI can help nonprofits engage people in ways that feel timely and personal and respectful. So more individuals are inspired to give. More individuals participate and they support the causes that they care about so deeply. And so that's usually at the the kind of forefront of of of my mind as we build here. And so that means broader participation in giving, stronger donor and community relationships, maybe faster mobilization in moments of critical, critical need, and maybe less friction in some of the the processes that we all have in our everyday lives. And so, you know, again, for me, the the real outcome isn't, AI driving the generosity, but actually just removing all of the barriers so that human generosity can show up more often and at greater scale. I guess that's the it's usually the positive outlook I usually have on it. So I know we, we built in the time for questions from the audience. And while we do that, I'm gonna ask you all the same question I asked our our friends to put into chat. Talk about your favorite source of learning, whether it's a podcast, there's articles that you love, you have a learning platform. Where are you consuming AI information today? Cecile, I'm gonna ask you. What do you love? What's that what's the content that you go to every day? Oh, I need to pick, my favorite one. You know, there's a lot at work. I I'll just pick an example. You know, I working in the nonfinancial risk space, there's quite a lot going on. Also linked to AI. So, you know, speaking to to various organizations on how do they prepare for everything that's going on and, you know, how can they test their readiness. So one of the things that I am currently have been doing the last couple of weeks is just staying on top of everything that can go wrong and what are the latest threats and what are the kind of the latest scenarios to be prepared for. So this is just kind of my my latest favorite or kind of most used, you know, AI, AI use case. It's to kind of really stay on top of, you know, what are the risks and therefore, you know, how can I how can I kind of help, you know, the the organizations that I work with to to to think through what are some of the most important risks? But I also I mean, on a very personal note, I've also used it for menu planning for the family. So, I'll shoot them, the usual, you know, mommy is, trying to, be on diet this week. The rest of the family instead needs to needs to, have a good nutrition. So, you know, based on that, can you just suggest, you know, what should be my shopping list? And and this is actually then what I use on on on Sunday to kind of, go through the supermarket and make sure we have everything for the week. So it's it's very broad. I love it. Love it. Paul, what about you? What is your favorite, source, podcast, you name it? Yeah. I'll name two. Source, like, I love the courses on Coursera. I think there's a ton of really great stuff that's coming up. And, you know, you don't have to go and pay thousands of dollars to learn something new. You can quickly be able to implement it and get started, if you just wanna learn. You wanna just quickly dive in. And there's so many courses. I know IBM has a ton of courses on AI. It it's been a place that I've been pointing people for a while. I also think there's a few podcasts. You know, I was recently on the Fundraising AI podcast, and I've been keeping an eye on that one specifically as I get closer to the industry. Really try to understand our customers. I think they do a great job, have some great guests with very different opinions, different viewpoints, and, you know, have a great conversation. Awesome. Yeah. I encourage everyone go take a look at Fundraising AI. It has so many of community members that you know and love here that are, you know, publishing and getting on the podcast and sharing their perspectives. It's fantastic. Daniel, what about you? I I just love, AI as a way to get smarter on things. So if I'm reading an article, I can quickly drop it in and explain this term or explain the origin of something. And it's a way before I gloss over or not truly internalize and understand and use it as a way, not just, hey. Here's the explanation, but probe one or two levels deeper, to truly understand what what what's behind it, what's the meaning. And that that actually helps me learn not just one thing, but a couple of things that are new each day. Yeah. I love that. I'll share mine. I love Lenny's podcast. I encourage everyone to take a look at it. Paul, I know you you listen to that one as well. It really forces me to be curious and to think differently and to just hear so many different perspectives in this fast paced moment. I can also go on a walk and and listen to it as well, so it it just fits into our lives so so nicely. So that's a great one. And then because we're all here, I will, you know, plug the AI Coalition for Social Impact. Right? So, you know, Blackbaud knows that there are AI barriers, to adoption. And so we formed a coalition of technology, AI strategists and industry experts. McKinsey is on this coalition with us as well. We will be releasing an AI certification, product agnostic, and free to the social impact sector. Sector. Maybe our friends behind the scenes can even put the link into the, into the chat to give some more information there on how to get into the waiting list. But even that will be a great resource for all of you, to have in this moment as well. I think the more we consume across so many platforms, the the better off we will be. So, our wonderful team, again, behind the scenes has been taking in your questions, and having them prepared for us. So let's talk about the first one. So the types of agents described suggest a high replacement of people for headcount and jobs. How do you ethically ensure it's a capacity add rather than a human replacement? This question is so important. You no longer have an event planner, you have an agent manager who is determining from an event planner expertise what decisions to make or what scripts to give the agent. So Dana, I love your perspective. because you and I have had this conversation before. How do we make sure that in this. moment, the agents amplify our impact as humans rather than relate would replace? So talk a little bit about this for this this question. That's a fantastic. question. No. It's a great question. It is on a lot of folks' mind, these days and understandably. So I would say the answer is that, look, it's it's not AI that one should be fearful of taking one's job. I should be fearful of the person who can more effectively use AI than I as an individual can. And that that's the real the concern. I'd I'd go back to the the example I had earlier of the grants and the grant applications. If I. can more effectively use AI to write grant applications, to launch a fundraising campaign, to get funding for my organization. That makes me more valuable as an employee. It makes more me more valuable and effective as an individual. And that's sort of what I really focus on is how do you increase your own value proposition through the use of AI? I mean, the example of the the article, for the conference using AI as a way to be able to publish more articles and get your own voice out there. That's what I'd be really thinking through. Mhmm. Is there anything that we need to be thinking, Cecile, on this question from a risk perspective, a trust perspective? How do you think about this question? It's okay. Apologies. There you go. No. We got that. Technical issue. I still don't have an agent for that. Yes. So, I mean, I think it needs to be a consideration. Right? I going back to the risk, and the risk taxonomy as a foundational bit, I think what what's, you know, what's what's raising this question is is is something that is, I think it is a risk. Right? And I think it is something that needs to be therefore kind of considered from the start. So, I think any AI opportunity, you know, when it's developed, when it's being kind of just, you know, thought through in its initial stages, needs to be thought through also with the human implications. Right? So when we do our work, with some of the organization, their AI transformations, we always partner up also with the HR department, right, to think through, you know, what are going to be the human implications of this? How do we make sure, you know, the humans will be able to stay on top of this? They will be able to perform, in some cases, a human in the loop role in an effective way and in a way that continues to also challenge and motivate them. Right? So I think it's basically kind of, you know, in terms of our framework and how we're approaching thing, it is something that needs to be a consideration and top of mind in any of the AI work we do, and so that it can be kind of, you know, carefully managed. But I would fully agree with Daniel, right, that my experience is very much that, you know, AI is enabling it's increasing productivity. So it's enabling, you know, people to focus on other things that are more value adding. So that is where that extra capacity goes to. And also that, you know, the importance is that, that you stay on top of it. Right? And that the people are. are kind of able to, use the new AI tools effectively. But it's, it's empowering humans. It's making them stronger and more capable. Right? That's that's really what we are seeing across all sorts of organizations. Yeah. We talk a lot about here, you know, how do we amplify our impact with the use of AI? And I think that that word amplifies that that challenging piece for us to start to spark debate on how we do that. Here as well. Still, I'd love to stick with you for this question if you don't mind too. I think, we have one maybe time for this one last question. We can sneak it in. What are your thoughts on the speed of AI growth? Right? We're all feeling it in this moment. It's going so fast. How do we make sure that we're keeping up with those checks and balances as well? That's a great question, and that's my mission at the moment in all of the work I do. This is exactly what our research is suggesting as well. Right? Governance is lagging behind, and this is a problem that needs to be solved. I think, luckily, what we're realizing is that this is this is this is blocking organizations now from further pursuing the opportunity because the governance hasn't been built out, you know, very practically. You know, once once a new AI use case is going live for an organization, some some organizations haven't thought through how we're going to govern this. And this is actually them holding them back from kind of further pursuing the opportunity. So I think, you know, this is true statement, I think, is is is my a question. Second one, answer. Sorry. Second one is we're trying to resolve it, working, you know, with all of our clients. But I think the third point is because of this, you know, the opportunity is typically also slowing down. So there's a huge motivation also from organizations to to solve this, right, and make sure we're keeping up also from a governance perspective. Yeah. And sometimes I think about, you know, earlier I talked about moving governance from theoretical to practical. Right? And so we've had these trustworthy AI building blocks for decades that have served us so well before agents came along. Right? So how do we just start to make some of those pieces practical? Right? When we think about our trustworthy AI building blocks, it's governance, it's policy, it's empowerment, and it's process. So if you think about governance, do we have an AI council at our organizations that have really diverse lenses that are looking at the use cases, sparking debate, and making decisions together? Do we have a policy, an AI policy? When we look at our Blackbaud Institute research, we see that the social impact sector does not have policies across all of their organizations. If you do, make sure it's not on the shelf. Make sure that every employee sees themselves within that policy and they know how to, abide by it. Empowerment, I think has come up a lot in this conversation. How are we making sure that we are curious and courageous and learning and stretching and challenging ourselves throughout this process? And then process. A lot of what you talked about, Cecile, like how are we bringing all of that into play to make it part of the, you know, development life cycle versus a a bolt on at the end or something that's different or separate. So So I think, you know, sometimes those tried and true principles are are there for us along the way. So, thank you, my friends, as always. You are all so wonderful to me on a daily basis. Great friends here today. I'm so glad everyone got to join us, in this panel today. If there's any closing thoughts, I'd love for you to all add. Otherwise, don't forget you have everything in the document section, of that chat area to pull down and then, you know, ask questions along the way. But, any closing thoughts for this group? Alright. Then that'll be me. I'll say this is a learning moment for everyone. I hope you took a a screenshot or you pull it down around the AI Coalition for Social Impact, and how we are providing a whole host of opportunities to take down some of those barriers. Thank you again for joining. We are always here along this journey and, I bet we'll be back with a whole host of additional webinars along the way. Thank you. I hope you have a great day. Thank you. It's been a great pleasure. Thank you, Carrie. Thanks, everyone. Thank you.