Video: The Predictive Edge: AI Strategies for High-Impact Year-End Giving | Duration: 2680s | Summary: The Predictive Edge: AI Strategies for High-Impact Year-End Giving | Chapters: Welcome and Introduction (0s), Webinar Introduction (67.39s), Types of AI (212.565s), Predictive AI Strategies (454.48s), Optimizing Donor Strategies (801.235s), Context Engineering Techniques (1212.91s), Driving AI Impact (1540.285s), Addressing AI Concerns (2269.65s), Mitigating AI Concerns (2355.41s), Data Security and AI (2436.8499s), Database Cleaning Automation (2500.045s), Closing and Resources (2574.3599s)
Transcript for "The Predictive Edge: AI Strategies for High-Impact Year-End Giving":
Welcome, everyone. Thank you so much for joining us for our end of year fundraising webinar session today, The Predictive Edge, AI Strategies for High Impact Year End Giving. Before we kick it off, I just wanna go ahead and introduce my colleague, Sophie Wilson. She's our Blackbaud representative who's on the line in case you have any questions about how Blackbaud can help you transform your year end fundraising results and help revolutionize your mission's impact. You'll receive a link to the recording, and I'll include those, helpful resources as well as links. And please feel free to enter questions as you think of them throughout the session, and we'll make sure to answer them live at the end. And without further ado, I'll go ahead and pass it over to our experts, Carrie Goldberger and Stephanie Brown today to go ahead and dive on in. Great. Thank you everyone for joining us. I see some activity in the chat. It looks like maybe the docs aren't, appearing for you all currently. We'll see if we can get that resolved. I'm hoping you do also see a q and a section. That's where you can place your questions. We'll be monitoring the chat as we go along as well. So just by way of introduction, my name is Carrie, and I am a principal product manager here at Blackbaud, and I am on our data intelligence team. And then I'm joined by my colleague Stephanie as well. Stephanie? Yeah. It's great to be with you all this morning or afternoon as the case may be. I'm Stephanie Brown. I'm a senior strategic consultant in Blackbaud's analytics division. So I did see a few names that looked familiar, among the people who are here today. So some of you do know me already, but I'm one of the individuals who work with you when you do any kind of predictive modeling project with Blackbaud. Okay. Thank you for confirming. It looks like the docs are appearing now in the chat or in the the resources for you, and I'm glad to hear that because we do have a lot of resources available for this webinar today. We're going to be referencing several of them throughout. Here is an example of the first one. It's a new GivingTuesday report from Blackbaud Institute. So the Blackbaud Institute and GivingTuesday combined, our data expertise to help organizations drive long lasting success on GivingTuesday. There are a lot of great statistics and research within this report as well as recommendations for Giving Tuesday and year end giving. One call out in particular that I have here on the slide, in this report, we found that GivingTuesday donors have a 65% retention rate as compared to 52% for all donors. So even though and it calls this out as well, the total number of donors who might give on giving Tuesday may be a little bit lower. Those who do give are more likely to give again in the following year. So, again, throughout this presentation, we're going to be referencing, some of the research from this report in particular. Alright. So why is AI and why is this predictive data important? Well, according to a recent Blackbaud report, AI usage is growing in fundraising organizations. 82% of participants in this research indicated that they are using AI in their offices, mostly for drafting content virtual communications such as note taking during meetings and things like that. So this is really the future of making your fundraising more efficient. Now I wanted to dig in a little bit deeper, because when somebody says AI or artificial intelligence, that term is pretty broad, and it can mean a lot of different things that covers a lot of ground. So the three main types of artificial intelligence that we're going to be discussing today are predictive AI, generative AI, and agentic AI to some extent. So predictive AI is the the forecasting side of things, and this is not new. We've been doing, creating machine learning models and doing predictive AI for years. This is the idea of identifying patterns within a dataset to predict what is likely to happen going forward. So that could be predicting donor behavior, whether that's who is most likely to give a gift of a certain size or who is most likely to respond to a certain type of messaging. But something really important to consider is how you put that predictive AI into action. So one example might be prioritizing outreach. Are you looking at okay. If of all of our donors, these are the ones who are most likely to give a gift. Let it let's plan our strategy and our outreach based on those predictive analytics. So that's one part. The other, which a lot of people when they say AI, they are referring to generative AI. This would be the creating side. Think tools like ChatGPT or Claude from Anthropic, or so it can generate new text or images or video. And, again, this is gonna be based on a training set of data. So given some context, given a prompt, it will predict what are the words or phrases that are most likely to come next. But the way that's then applied is maybe to automate donor communications or create some personalized content. Agentic AI would be the doing side of AI. This could include things that help guide users through specific tasks or maybe representing an organization in donor interactions. Now for all of these, you'll want to have guardrails in place, but especially with that agentic AI. If you don't have a human in the loop for specific tasks, Maybe you have somebody who is setting up the guardrails that help define how it should act or how it should communicate within certain scenarios. So, again, thinking about both that research that we just shared, but in general in conversations, when people say, sure. I use AI. It is good to also dig a level deeper and say, okay. Well, how exactly and which type of of artificial intelligence are we talking about here? Alright. So we're going to start off with talking about how we can use this AI technology to help pinpoint high potential donors. And why is this important? Well, as we all know, there is limited financial and staffing resources available for all of our organizations even at the larger organizations. And especially toward the end of the year, as most organizations are doing both a Giving Tuesday and an end of year campaign and staff are taking time off during the holidays, it's more important than ever to strategically prioritize your fundraising for those individuals who have the higher likelihood to make a gift. As you can see from the charts on the slide, recent Blackbaud Institute research indicated that organizations with a higher digital maturity, those who are using AI in assisting and pinpointing high potential donors, have reported that their income has been growing at a higher level than organizations with below average digital maturity. And that's at all levels from the small annual giving level into the major gift level. So one example we have of using predictive AI, predictive analytics to pinpoint major donors comes in the form of Prospect Insights. If you're not familiar, Prospect Insights is an AI powered dedicated workspace for major gift fundraising within Raiser's Edge NXT. And, again, going back to what that predictive AI, how it works, it is helping to identify of all the constituents within your razor's edge NXT environment, which are the ones who are most likely to give a major gift to your organization. So it'll help you to identify those top prospects. It's also using predictive modeling to help you determine ask amounts, to align those ask amounts based on the predictive analytics. It can help with proactive fundraising. So not just reaching out, when somebody maybe raises their hand or there's some signal that they could be a major gift donor, but also helping to build out your pipeline of future major donors. And then it also helps to support portfolio churn. So not just highlighting who is most likely to give a gift and should be assigned to fundraiser, But maybe you have some people in portfolio who have a low likelihood to give a major gift to your organization. And in those instances, you might wanna consider removing from portfolio in order to make room for, those prospects who are more likely to give a major gift. So, again, for all of this, the the goal and the question to think through is how can we apply these predictive analytics, to our strategy. How do we put them into action? So to dive a level deeper, you can identify those prospects who are not currently assigned to a fundraiser, and they have a high target gift range and major giving likelihood value. So combining those predictive analytics with fundraiser assignment status. So if you have individuals who are not currently assigned to a fundraiser and they are coming up as high priority to give a major gift to your organization, then you can, take that action, have a conversation with the team, and see, does anyone have bandwidth for this individual? Have we tried reaching out to them in the past? Do we want to mark them as qualified and then assign to a fundraiser? Similarly, you might want to look at prospects who are considered under giving. So these would be individuals who are currently assigned to a fundraiser, and they have high target gift range and major giving likelihood values. So we don't have any recommendations there in terms of assignment status. But when you look at their previous giving to your organization, those years of giving fall below the predicted target gift range. So, again, this is an example of where you're combining the predictive analytics with the status of a constituent of a prospect in your organization to determine what action should we take next. So if these individuals are already assigned, you already have a relationship with them, and the predictive analytics suggests they could increase their levels of giving, and especially for year end purposes. Maybe these are great prospects for, to ask for a matching gift or to see can we increase their levels of giving, to our organization specifically. So with prospect insights, it'll pull all these people together for you in a prioritized list. But, again, maybe you're using, predictive analytics from a different source, and you wanna think through how do I combine these predictions with what we know about this individual to determine what is the right strategy. How do we put this into action? How do we want to move forward? Now if you don't have Raiser's Edge and don't have prospect insights available to you in NXT, there are also other predictive models that you can use through Blackbaud's prospect point predictive analytics. So these particular, predictive models work across a variety of database platforms, and you can get scores based on an individual's likelihood to make a gift at an annual giving, a major giving, and a planned giving likelihood, as well as a suggested target gift range. And, again, you can use these tools, you know, by themselves or in combination with other analytics, including with prospect insights just to get a broader depth of information, you know, to be able to fill out the picture of the story of who your donors and prospects are. Great. So now that you've used the predictive AI to help pinpoint who are the donors, who are the prospects that we want to reach out to, the next area to consider is how will we optimize our cultivation strategy for these individuals. So there are a couple of resources in the doc section that I'll point you to for this slide. First, there's a blog post related to optimizing donation forms for GivingTuesday success. So within this blog post, our colleague offers some recommendations of things to consider as you're setting up donation forms. It might be things like, using a multistep form. So each step in the process might focus on asking for a small amount of information, from the donor. But we've seen through testing that having that guided workflow leads to higher conversion rates, considering adding a goal meter. Also, it suggests adjusting your ask amounts. And so this is where that predictive AI can come back into play again, whether you're using an intelligent ask on the donation form or some of those predictive analytics that we were just referring to. You do want to tailor your ask amounts for each prospect for each individual. Also, want to think through how to tailor your messaging for each individual. Maybe you're considering their past giving patterns, communication preferences, or just interests of different donor segments. What messaging is most likely to resonate with a given audience? Again, this visual that I have here on the slide points back to that same GivingTuesday report that I referenced at the very beginning. So within that report, they have a nice breakdown of GivingTuesday donors, by each generation. So you can see that that breakdown here, the majority of GivingTuesday donors would fall into the baby boomer category then followed by millennials after that. But the other nice thing about the report is it goes into way more detail, and it also offers suggestions of where you are more likely to find these individuals, meaning their online presence. So who is most likely to see something on LinkedIn, for example. That was a call out for the section on millennials, but also considering what messaging, again, would resonate with the different donor segments that you've identified. So speaking of donor segments, Stephanie, do you have more to share on this? Yes. So if for those of you who do have, Blackbaud Analytics and have the Affluence Insights, this particular slide is going to look familiar. But this is information, again, based on some predictive characteristics on different wealth segmentations that your donors or prospects may fall into based on a lot of the information that Carrie was discussing, such as, social media preferences. For example, a philanthropist is going to be more likely to be active on a LinkedIn versus a Facebook. Also, channel preferences. Do they prefer to be contacted by direct mail or online? And, also, these wealth segments also take into account giving patterns. Does somebody tend to concentrate their giving or do they spread it out? And as well as messaging, do they prefer data driven type of messaging, like a philanthropist or an affluent enigma, or do they prefer more human centered messaging, like someone who would score as a humanitarian? Digging a little bit deeper, there's also some additional different kinds of predictive models. For example, especially for this time of year, there are those end of year and giving Tuesday likelihood scores. So an individual who is more likely to give to those particular campaigns, this might be how you would segment, some of your messaging. For example, when you're sending out that end of the year campaign, you would focus on those individuals who have that higher end of year likelihood score. Again, you can also get more data on likelihood to give to certain channels or multiple channels, which is would be especially helpful in prioritizing your direct mail, which would have that increased cost. Prioritizing that for individuals who have those higher direct mail likelihood scores. In addition, also, individuals who have a high likelihood of being a sustainer. So if you're doing a campaign around marketing for sustainer giving, you would look at those individuals who have those higher scores as well as the gift arrays and the donor personas, which, we've already discussed to some extent. Great. So I also wanted to spend a bit of time today talking about, ways to potentially craft your messaging with generative AI. I bet a lot of people are familiar with the term prompt engineering, which is the idea of really crafting the best prompt for a tool such as, whether that's Claude again or ChatGPT or whichever Microsoft Copilot, chat for Blackbaud AI, thinking through what prompt is going to get the best output, the best messaging. But more recently, I've seen this start to this concept of prompt engineering evolve into something called context engineering. And I kind of like this approach because we believe that context is everything. The context is what's going to help you get the best response from a generative AI tool. And so as I go through some of these ideas and how this applies, I'd also be curious in the chat if anyone has either heard this tool before or any best practices that you've come across in creating your own prompts. If anyone is currently using, generative AI to create your messaging or to help with campaigns, I'd be curious to to learn from you all as well. And maybe, if you don't mind, put some of those ideas into the chat. But some high level principles to consider here are just to think through what information you need to share in order to tailor the message for different segments of your audience. And as you're considering what context am I going to share and what information do I need, You may, and I would highly recommend actually, leaning on those personas, instead of using personally identifiable information, especially if you are using a public version of these AI tools. So an example prompt here that I have on the screen might be, let's create content for a social media campaign, and I want that to resonate with millennials who are working professionals. And I want to focus that messaging for a LinkedIn campaign. So that's some of the context that you can add from the start, but one tip, one best practice that I've come across that has helped a lot is I say ask me clarifying questions before creating the draft. You can even have that back and forth with the generative AI tool to help determine what context you do need to share. Sometimes I'll even ask, what context do you need from me in order to x y z, in order to create the social media campaign? And so then you'll find that it asks you questions. It'll ask you those clarifying questions. You'll respond. You'll have that back and forth. And then at the end of that, you can say, okay. Great. Summarize our conversation and create a draft document. Then you can use that document for future conversations as well. So you now have that context documented, and that's great for internal purposes too. So whether you're sharing it with other team members or you're sharing that context again back to a generative AI tool, you just have it now. A lot of times that context is, you know, in in your brain or someone on the team, they just know it, but it's not documented anywhere. And so I've found that you can use gener AI as a sort of thought partner to help you clarify your thinking, get it documented, and then use that as context going forward. Again, if you've come across any best practices or tips for using generative AI, I'd love to learn more about it if you don't mind adding to the the chat for our conversation today. Another example though of providing context would be to share examples of what good looks like. So, again, if you're creating this content for a social media campaign, maybe you provide samples of previous campaigns. What messaging have you used that, led to higher conversion rates? And if you provide that context to the generative AI, then it's going to use that with the future content that it's providing for you. So, again, here, the key messaging or the key takeaway is considering that context engineering and what context does the generative AI need in order to provide the best possible content for your use case and your scenario. So now we'll go into the section on driving that impact. You've pinpointed your potential donors using predictive AI, tailored that ask amount and messaging based on donor segmentations and based on that predictive analytics. You are crafting the messaging potentially with the help of generative AI, but always keeping privacy data privacy in mind. So then where do you go from there? So in terms of stewardship, which is a big way that you can use this AI to help you, especially at end of year and Giving Tuesday, there are a variety of different, things that you can do. And a lot of this is you will also find in the Blackbaud ultimate end of year fundraising toolkit. So for example, when you're sharing the amount that's raised and letting your donors celebrate, you can use AI tools to help you craft those social media posts and emails, again, based on those personas and other data like we were just discussing earlier. You can use it to help personalize communication to specific individuals, especially on those thank you messages. You can also craft different types of examples of, social media posts across different platforms based on those platforms, you know, specific types of messaging. And, also, you know, this will help you in directly updating all of your individuals in a way that they are most likely to be able to receive and appreciate that type of messaging. So similar to the use case of using AI to generate, a social media campaign or your initial messaging, you can also use AI to help draft the content for acknowledgment emails. Here, again, you wanna consider the context that is being shared. So, as much as possible, creating a personalized draft. This is an example of the AI acknowledgment emails within Razor's Edge NXT. And so within this context, it can, include information about maybe the constituent or the gift or whatever details you determine, you want to include within these drafts. Here you can see you also have more control over the the type of messaging. So do you want it to be fun or informal or take a more formal tone? And, again, you can use the segmentation from before that we were mentioning to help determine what sort of messaging is going to resonate, how do my donors want to be thanked, how do we, build upon this relationship that we are, creating here? And can we use some of that generative AI functionality to create that messaging for our donors? So I mentioned at the start, there's also a third type called agentic AI, and we do have a couple of links in the doc section that talk more about the agents for good from Blackbaud that, use the the agentic AI here. So this is the idea of you set up again some guardrails from how to to define how you want the agent to interact with different donors. But then from there, it can help to reach out to those donors that maybe you haven't had the bandwidth to have a personalized personal conversation with. It can help to identify new donors and then say, okay. For these individuals, here is, our cultivation plan and the messaging that we are going to send out to those who maybe have lapsed or are at risk of lapsing. Maybe you say, okay. I have a goal of setting up a campaign and set up those guardrails again for the agentic AI, and then it goes from there and it builds out more of that campaign and that plan that you can then go in and, tweak or make updates to as needed. You're always in control, so you can go in and say, I want to update this messaging before it is sent out to your donors. Agentic AI starts to take it a step further, for for your organization. Again, we do have a link in the doc section where you can learn more about our development agent and the agents for good that we've been building out here at Blackbaud. And to bookend, one of the things that we've brought up at various points during our conversation today is the importance of building an AI policy. While 82% of participants in the recent Blackbaud Institute study on fundraising in the AI area, era did identify as AI users, only 14% of individuals have an AI policy in place at their organization. And, you know, there are a variety of reasons to put this policy into place. You know, again, this is how do we use this AI responsibly? You know, this can include some of the things that we've discussed such as only, you know, not putting personally identifiable information into a public source such as CHAT GPT. And we don't want to, you know, risk the potential of that being used to train models all over the world. And, also, you know, just putting in those guardrails on how to use this particular information so that we can combat some of those concerns or mitigate them such as misinformation and data security and reputational risk. If you're interested in learning more about AI, especially, around how to draft a responsible AI policy and how to use AI responsibly at your organization, I encourage you to look into the upcoming free certification for AI for social impact. So this is coming from the AI Coalition for Social Impact, and this particular certification is especially designed for our sector. So this is a, coalition that is made up of a variety of organizations, including Blackbaud, Giving Tuesday, AFP, the Responsible Artificial Intelligence Institute. Again, a variety of organizations that will provide training on all aspects of AI skills. And this particular course is going to be beginning in January. So if you're interested, definitely encourage you to check that out to grow your skill sets around AI and how to use it responsibly. Yeah. Another, resource here that I would point you all to in addition to a lot of the resources, from Blackbaud specifically, but if you haven't already, check out Fundraising AI, and there's a link in, again, the docs. I'll point people back there. The the free certification, you can find more information at blackbaud.com/ai. I think that'll be one of the the resources in the docs. And if not, we'll put it in the chat. That's where you can find, the button to join the wait list. Again, that's blackbaud.com/ai. And then Fundraising AI, is another great resource if you're trying to put together thanks, Abby. He just dropped it in the chat. If you're trying to put together a data policy or an AI policy, you'll find a lot of great resources there, a lot from the virtual summit that just occurred. So a lot of great webinars with with more content to review at that place. I just want to double down on something that Stephanie shared earlier. We've mentioned guardrails a couple of times and the idea of responsible AI. But the nice thing about putting guardrails in place is it gives you the opportunity to really experiment and try out new things and figure out what works and what doesn't work for your organization, doing it in a safe way, a safe environment. And so whether that means you are just creating, dummy data, so fake information to sort of see, okay, if I told you, information about a donor that's not actually real, what can you do with it? What sort of messaging can you craft, based on that information? So I found that having those guardrails in place has been really valuable, for my own professional development and things that I'm experimenting with because you can sort of test out what works, what doesn't work, but you feel safe doing it. You don't have to worry about, sharing information that you shouldn't be or or any of the the downsides to consider. I don't know. Stephanie, I saw you nodding a little bit, but if you've had a similar experience, or if those guardrails have also come into play and kind of helped you, with how you're using AI. Yes. Absolutely. Using that kind of dummy data, so to speak, has been really helpful because sometimes I will have a question, especially if I'm working with a client. And I'll anonymize the data to some extent, you know, no actual names or given histories. But just saying, if I was looking at somebody who had this kind of thing going on, what would be my strategy? So using a lot of that. And one of the things that I've done as well, and this also might be applicable for those who of you who are on the call, is I've been summarizing client notes using, Copilot to summarize client notes to help highlight some different aspects for additional strategy. And you might wanna think about doing some of that with contact reports, especially for those donors where there's a long history. You know, seeing what surfaces out of that. You know, is there a particular point that, you know, you may have discussed in a meeting a few months ago that may have just kind of slipped your mind? But when you have kinda Copilot or another private source look at it, again, not for public sources, it might surface something as looking important that you might have either forgotten about or might have overlooked. I know in my case, that has happened before. And it's like, oh, I would have missed that one sentence. But this actually is a very important thing to discuss with the client or with the donor as the case might be. Yeah. And I think transparency is a big part of that as well. Right? So not just the consideration of a public versus a private tool, but it's all a part of trustworthy AI means being very transparent in in your use of AI and, like, if a call is recorded, making sure everyone is aware that that call is being recorded. I see a question has come in. Any advice for team members that are against using AI? I want to try and make my life easier with AI, but maybe getting, pushback from others. Be curious to get your take, Stephanie, as well. But for me, I think part of it is just having the conversation to understand what those specific concerns are. And I'm finding that making the conversations as concrete and tangible as possible goes a long way because I think there are a lot of really valid concerns about the use of AI, and that's where, again, you wanna think through responsible AI and ideas related to trustworthy AI. But when you're just talking in the the abstract or hypothetical, then a lot of times that doesn't lead to anywhere productive. But if you're talking more about, okay, for this specific use case, this is how I think we should use it and why. And you might find, you know what, for this use case, we're not gonna use AI because we've decided as a team that it doesn't make sense or we feel the risks are too high, and that's totally fine as well. But, again, you wanna have those really concrete examples and that that discussion, I feel like, is where most of the value comes from, having that discussion as a team about what are our concerns and why, and are there ways to mitigate the risk. Absolutely. And having those conversations with those particular individuals on their very reasons for resistance, Is it a concern with data security and something like that, or is it the concern that AI is going to take my job? So one of the ways that I frame it is finding ways that it's going to save time, doing those kind of more routine tasks that just eat up time. For example, for a gift officer, if they're spending a lot of time individually crafting some messaging that AI can help with. So that frees up more time for them to actually meet with their donors and talk with them. Mhmm. Yeah. That's a great point too. I see the comment about concern of data security, and that is hugely important because AI is only as good as the data that it's using. But, again, that's where a lot of the trust comes in. We need to be good stewards of the data and how are the data points being used. And so one reason I'm pretty excited about this certification again is, several of the organizations involved are leaders in governance, specifically in data governance and thinking through that side of things. So I'm pretty excited to see, what comes out of of these courses and that certification. Yeah. And this is, again, where that, AI policy for responsible use is so important. There's another question in the chat that, someone's asking. Can you give a tactical example of how someone might use agentic AI for year end? Yes. So one that comes to mind for me, and I suppose this is still, a developing area. So this is a little bit of me, you know, thinking out loud. But one one area that has come up and a lot of questions have come up for us is the idea of, cleaning a database. And for the year end example, I'm actually thinking about it of after the gifts are coming in and are there processes that Agentic AI could help to automate, to make sure that, again, people are stewarded in the right way. So sending out those thank you emails or thank you letters and using messaging that's really tailored for each individual automatically. But then also thinking through, are there are there any steps in the process that, where cleaning is needed, and is there something automated that can help with that process. So that's the first one that that kinda comes to mind for me in the context of of year end giving. It looks like we've hit on the questions that have come in so far either as part of the content that we shared or here at the end. If there are more questions, though, feel free to, add them here either in the, again, that q and a section or we're watching the chat. And we do have the the resource available for yeah. Abby, do you wanna close this out? Maybe you can cover this in particular. Yeah. Absolutely. Thanks so much, Carrie and Stephanie. I know this was really impactful content. So we appreciate your expertise here. Just a reminder that you can scan that QR code and download our, ultimate and view fundraising toolkit. I know Giving Tuesday is right around the corner, so there's a lot of great resources including, email templates and just everything that you need, a checklist for Giving Tuesday. So definitely check that out, and don't forget to attend our upcoming webinar, for anyone that even if, you know, all of your ducks are in a row or maybe, let's be honest, your end is is hectic and can be crazy. And so maybe you're, you know, struggling to pull things together. No problem. We have Panic to Profit, your eleventh hour playbook for Giving Tuesday and year end success. So a lot of great info there, not only for Giving Tuesday, but also what you can do after you wrap up your year end campaign. So don't miss that. You'll receive an email with the link to the recording, and I'll link all of these helpful resources that Carrie and Stephanie mentioned as well, and we appreciate your time today. So thanks so much, everyone. Hope you have a great one.