Less Data, Less Risk: Minimizing Data Collection to Protect Your Nonprofit's Mission

Hi, my name is Eli. I am a community manager at Tech Soup. Um, and I'll be your support person throughout this event. So, we know that nonprofits often hold years of personal data about children, families, and other vulnerable people in their data in their their files because frankly, sometimes we need that to deliver on some of the work we've done. But often that data may no longer be needed and increasingly it can be hard to manage your responsibility. So in this conversation, Christa Hill is going to talk a little bit about how we can reduce this risk from this data collection using some of their background around use for it as part of some of the practical use cases and conversations. We're also, of course, going to touch in on AI because it's 2026 and there's no way we're not gonna. So, who is the Hill? So, Christa is the co-founder and CEO of Cassid Edge product leadership and a Canadian AI educator focused on helping missiondriven organizations just like yours adopt AI responsibly. With a background in product leadership at organizations including Getty Images, Morgan Stanley, Benevity, and the New York Federal Reserve, she now works with nonprofits of Brown Canada to build AI literacy, reduce digital risk, and modernize operations without losing sight of their core values. And hot new news, uh, Christa is also the chief AI learning officer for learned. And so that is what we're going to do for today. So, I'm delighted to have you here. >> I want to know a little bit more about this thing that you brought to us. So, you came to me and said, "We need to have a real conversation around the masses of data that nonprofits have collected." I know I've been terribly guilty of it. I look at my CRM and it's full of thousands of fields some of which I haven't touched in years full of all kinds of [clears throat] information hopefully banking information but frankly I haven't looked that closely >> so no >> I want to you said we're carrying this data and I'm guessing what what is the problem here like what are we actually looking at around holding on to this data like are the risks that come out of that Yeah. And I want to start by saying the world has changed. We used to take a lot for granted in the nonprofit sector. We have a lot of volunteers. We have a lot of data. Let me define what data is first of all because we use that that word like as a misnomer and everybody's afraid of it now because we're like, "Oh, data data leaking. Somebody could get my data. What does that actually mean?" And so I'm going to use just by seeing the chat here and a lot of folks are coming in from all different verticals in the nonprofit sector supporting all uh different types of humans in either their family lives uh in their sporting lives or in their daily lives or their mental health. Uh there's a wide range here. So what I'd be talking about is say a spreadsheet that you've been carrying and say you create a new program and you get people to sign up on that spreadsheet. It's got their name, their address, maybe their kids' name, an emergency contact, uh you know, other relevant data that you would might have to collect for the funer, right? So, if you multiply that over 10 or 15 years of us collecting and creating just these individual spreadsheets that perhaps we passed to a volunteer and it lives now on their computer as well as someone else's and then we created a new one for the next iteration of the program. What I'm really talking about here is all the aggregation of just years and years and years of using Excel to run our operations. And nothing much has changed since we went from paper to digital. We all jumped into Excel thanks to Microsoft. But then we stayed there and there was no real threat to that data because if it ever got out, the implications might have been quite minor. But now in the age of AI, if one of those spreadsheets get out, it is it's a huge liability for us. So the cost of us keeping all that data safe has just gone up. And there is a cost to that. Whereas before, not a lot of us would have even thought of it that way. So I even even for myself, I'm a coach. I spent years in the figure skating community and I've got years and years and years of registration data on children like years. 10 years worth, their dates of birth, you know, their emergency contact information. All of those five pieces of information that I collected is enough to create a credit profile or a uh passport application from them for them in another country plus any photos I might have posted on them from Instagram with their likeness. So, we are the curators of a lot of data. And so, what I really talk about with uh nonprofits when I come into their organization is let's define data. Let's go in your practical day-to-day and what is the data that you hold? What is the data you need to keep safe? What is the data you're collecting because your funders need to collect it? And can we start to challenge the funders on how much data they really need? Because when you pass it to them, what are they doing with it? Right? So there's now a chain of events that we have to really be paying attention to because data is very valuable and it be can be used because I'm going to be very plain with you to create profiles and credit applications and identities around the world and it's a new currency and unfortunately we're ground zero for a lot of where that data can be released and fairly easily because we haven't really gotten the budgets for security practices. So, it's not our own fault that we're here, but I really wanted to come talk to you about this because I think it's really important for us to start raising awareness as a community. >> Well, so that's really interesting when you talk about being this risk. And so, I want to talk a little bit about how is this data actually leaking out of our organizations, right? >> Really simply, honestly, even if there is a board report that you're sending out, are you sending it out with a bunch of spreadsheets and a report as an attachment? Right? Okay, that is the simplest way that I would say like cuz now you're trusting a private individual with a board report for the health of your organization which which has a ton of data in it or another alternative in doing that is can you send them a link to which they would look at the report say inside of your Microsoft 365. There's just these subtle changes that you can reduce your risk without costing you anything but I would say the the biggest leak is just in our email accounts. That's it. >> So I think I've often heard around security and data which is to say like if you don't have it, it can't go out into the world. It can't be misused. But it brings to us this this hard question and you say like we have obligations from funders. We got our own practices that we built out over the years. How does an organization start looking at that data and saying how do I like how do I identify what's truly necessary versus data which is like well that would be nice to have but frankly not worth it like you know when I look at the value proposition it's not necessarily there >> so it's a change in a mindset when we have a client and whatever however we design define client whoever we're serving when they come in the door I want us now to start asking ourselves what do we really need to collect versus what we've always done, right? That that question of well, this is how it's always been done looms over our industry because nobody has had a ton of time with the resource constraints that we have to revisit how things have been done. But now we have something on the line. So, if I'm serving teens and somebody's coming in, am I taking their full name or just their first initial and maybe their last name? like let's get rid of the full first name, first initial, partial last name, even date of birth. Do I really need that or do I just need their age? Right? So, there's a lot of different ways that we can collect data to get us what we need without having the defining characteristics of someone's full information. Like I said, I want you to think about it like you're applying for a credit card or a mortgage. What are the things that they're asking for? They're always asking for your first last name, your address, your date of birth, and then all we need is two other characteristics associated with you, and we've got enough for someone's entire profile to to be given to someone else. So, above those things, what can we simplify to stop giving out everything or collecting everything? So again, it's just looking at those old spreadsheets and even I got to say this is a great use for AI is to take a look at your spreadsheets, talk about the work that you do and ask AI like verbally and I do this all the time with my nonprofit uh students, what do I really need to collect for my mission? Can you review the agreement that I've got with the fun? Can I be asking the funer to adjust the actual data sets that we're sending over? what is really necessary, how do I even go about challenging that funer in the data that we are collecting so that we can collect less because there are use cases in the nonprofit industry where we want to be collecting data and we can talk about that later for its value so we can understand the impact of our mission out in the world. But we're when we're talking about core demographic detail on individuals, it's really time for us to be getting way more precise about what we really need and what we don't need. So speaking of getting to that level of precision like obviously [clears throat] every organization has a different data set different needs but a model like what is the questions we might ask ourselves when we're trying to say is this data that's actually mission critical versus a thing we've done historically and let's stiff keep at it. like how do we get ourselves to a point where we start being able to winnow through all those options and say, "Okay, great. But these are the three things I actually need." >> I would say I want you to ask yourself with some very ancient spreadsheet that you work with all the time, what would happen if it got out in the world? What would you do? And you are the curator of that data set for that individual. You're responsible to keep their information private. how are you protecting them? And so when you look at that and you look at that 10-year-old spreadsheet, do you have a plan for what would happen if that information got out? And if the answer is no, that is quite normal, right? But the the most important part is that we're having this conversation and we're asking the question. So this isn't meant to to scare you and send off like all these five alarm bells in your organization. It's just a call to a conversation saying, "Okay, we collect all this stuff. If we pass it in between ourselves through email, are we passing it through an internal email or is it going out to say someone's Gmail account or a Hotmail account or you know a personal account, right? We're working with our boards and other volunteers. It's quite likely that this information is going out by email to other services. And if just for instance, one of those volunteers happens to be using a Gmail account, well now Google has a copy of that spreadsheet. How do we feel about that? How would that individual that whose information is in there, how would they feel about that? So I think you know in the past we didn't give this a second thought because there was no real arms race out there happening collecting proprietary data sets the way that the world has evolved to be doing right now with AI. But that is the reality. So I think really just that first conversation is let's look at what we have and let's ask ourselves are we collecting it just because that's just our habit or do we really need it and if we stopped collecting it this way and just really minimized how much we were collecting or tried to randomize and give our individuals maybe a number that only we know who their names are in a simple way that we can do in an Excel spreadsheet using AI can help us with that. uh how might we change what we're doing today and what could one thing what could be one thing that we change over the next 3 months just to reduce that risk just a little bit more. So I would say to you just starting with revising your email practices and just taking a look at them and see who's on that mass distribution list of whatever these older spreadsheets are and the reports where they're going and then also just asking yourself how much of this was really necessary and how much of it were we collecting because we could and there was just no consequence. But now that there is, what can we reduce? >> So this is really interesting because so many of the people here are small organizations. They don't have an IT department, but these conversations sound like they're not necessarily IT first conversations because you say it needs to be around what is our organization's values? And so it's so it's bringing in more people across the organization. When you actually look at these conversations, where do they start? Are they starting top down with the executive directors? Are they starting up with like the concerned person who's the program officer? Like how do you see these conversations actually starting in organizations? >> Typically it usually when it rolls in through one of my programs, we run an incredible program through the nonprofit chamber uh in Canada and then I've got another equivalent down in the states. It starts with people who are on the ground, right? And then sometimes the board, someone with some IT experience that comes in and joins a board, takes a look at everything that's been done and starts to raise a few red flags. But it can come in from any direction. But I would say most likely your ED might have gone to a conference and heard about something and it's like, holy, what are we doing? We got to get going. And then I'll get a call or somebody will ping me on LinkedIn and say, hey, Christa, where do I get started? And I can actually do a demo a little later here. I'll even just use chat GPT. I can show you what it looks like even just using something as basic as ChachiBT to help you get started. And you know, it just is asking a question and leveraging tools that are just now possible for you outside of just 365. I know through Tech Soup and other affiliates that really support the nonprofit sector, uh, Microsoft 365 tools like Copilot are around for to help you with answering some of these questions. But really it is just starting with somebody's got a question, we can go in and we can take a look uh using an AI tool to help us get some answers instead of solutions. We can ask it to help us come up with ideas on how to be safer with data because who else to help us uh think about some of these risks than the thing that we're protecting it from which is really AI and bad actors, right? So I'm not saying AI is a bad actor at all. It's not what I'm implying, but you know there is a lot of answers inside of these AI tools without having to disclose too much about your organization to get it to help you on your way. >> Well, let's touch in on some tools then. I've actually got a question here from Fu who's saying, "How should an organization whose ecosystem is really on the Google suite start thinking about how they can minimize risk?" As you say, if you go into someone's personal Gmail address, it's out there. But if you're in your own Google Suite space, it's a little bit more contained. What are the steps we can take within these ecosystems, whether it's Microsoft or Google's, to sort of hold the data in there and minimize that risk? >> Yeah. So, inside of Google, you have access to Gemini. And if everybody's on that enterprise and they're all together inside of that one account, then obviously there's that implied safety just like you would in a Microsoft 365 environment. So anytime you're working with an enterprise tools where you have seats that you've purchased for everybody, you can assume that there's a gate around your data. Now that only works though if you're keeping it just inside that ecosystem. And the second you email it out or share it outwards, then basically we become responsible for everything that happens outside of that gated community. So when you start with an enterprise license, there is that gated community that you create. Doesn't matter which one of the big ones that you go with. But again, what I'm talking about here is keeping that safe over the long term. Just because you have that gated community, I wouldn't take advantage of that that you can just keep collecting and collecting without thinking about it. I do think as nonprofits what I'm seeing is that we really need to just get a grip on what we are doing because one slip of things that we shouldn't have collected in the first place getting out is a stain we don't need when it could have been avoided by just reviewing those practices >> which actually sort of goes beautifully into Hannah's question who says like you know we're a smaller team and they're not bringing AI into the organization at this time like that's just a decision they've made >> but they still want to go through and analyze the data that they are currently collecting and say and do this these questions you've talked about which is like how do we make sure I'm collecting the right stuff and only the right stuff I'm minimizing that risk. So >> yeah, >> if we don't bring AI into the the mix, how would you look at say like what would you prioritize like as you did that analysis of data? >> I would personally prioritize around PII. So that's personally identifying information. And let me I'm just going to do a quick activity. Is it okay if I share my screen and I show you what that can look like? You betcha. Okay. So I'm just going to share >> and I'm just going to use chat GPT for the purposes of today. Okay, because I mean chat GPT is something that I know is wildly popular. It's we can talk about the different tools and and what they're used for and which ones are my favorites uh towards a little bit further into this conversation, but everybody's pretty familiar and actually if you've met one AI tool, you've met them all. And I know I'm about to use AI and you said you don't want to use AI. But what I'm going to show you something here is just you can ask questions of AI without it coming into your organization. It can be a partner that sits on the outside. So when you think about like blocking use of AI, when you've got questions around like what is appropriate, you have to remember AI is an an encyclopedia and it's more advanced than Google. Google will help you with some of these answers for sure. But if you want to ask about next steps and have a two-way conversation with that answer, then I highly recommend using AI tools in this way. So I'm going to just briefly I'm going to summarize your question here, Hannah, and I'm going to say, "Hey Ash, how you doing today? I am live on a Tech Soup uh broadcast and we're talking to incredible nonprofits about usage of data and you know if I'm brand new to say an organization and I'm looking at the data that they're collecting and they're a very small team, how might we figure out what type of data that they should be collecting and protecting when it comes to personal identifiers? Can you give top five tips on personally identifying information that we would want to protect or perhaps change our policy around collecting as a nonprofit? Okay. Now, I just used a tool called Whisper Flow there uh so that I don't have to type uh cuz I'm old and I'm lazy and I don't like to type anymore. And you see that what I did there is we talk about prompting. It's basically just line of questioning. I'm asking a question and I'm just trying to get some information as a tip. So I use chatbt as almost an advanced version of Google when I'm asking questions like this. So like I said, you don't this isn't what I would categorize as bringing AI into the into your nonprofit. This is just using a tool that is readily available just like Google. Uh she said absolutely for a nonprofit, I would make this much less about becoming a privacy expert and much more about developing good data habits. So if you cannot explain why you need a piece of personal data, you probably shouldn't be collecting it. So that's the principle really that that and I call her she sorry because I do use chat a lot and I call her Ash. So forgive me if that's weird but I spend hundreds of hours in these tools every day. So we do have names for them. Top five, do a personal data inventory. Ask what do we have? Where is it? Who can access it? And why do we have it? look beyond say your donor database. So if you collect donor information, things spreadsheets, Google forms, email attachments, volunteer lists, event registrations, someone's laptop, okay? And whether or not there's real security around these things. And let's face it, the majority of us don't. So when I teach in the nonprofit sector and the corporate sponsor that paid for the education comes in, one thing they always say was, "I thought we solved for security for the nonprofit sector." And I always laugh because nobody gave anybody budgets to solve for security around their data. So I think that there is, you know, there's some it's an internal conversation that we have to have and we don't need to be experts to do this. It's like look, I didn't share anything really about your organization or your mission and I haven't shared anything that was unusual and I'm getting tips here. Okay? So, stop collecting things like just in case data. I love this. A lot of times when I'm working within an organization, they'll say, "Well, I want to collect it just in case the funer asks for it." fair. But I want you to go back to the funer and say unless our funding can include proper security for us with our data, unfortunately we can no longer collect this. Is this okay? Can we come up with another way? So this idea of we us collecting things just in case we need it is actually if you if I'm going to be honest with you is the majority of data that I see in nonprofits, especially anything that extends past 5 years. So if I go back to if I'm going to go to a donor example, we keep donor information for a very long time because we want to go back and we want to see trends over donor behavior information and whether or not we lost a donor. Can we go back and pick them up? in a sporting organization I keep it just in case I don't know why because now we have amassed a huge amount of data on our kids that is floating around the world and after their registration why are we still keeping it right so in every vertical every industry that you're in there is always that example of just in case data and I would highly recommend that you sit down with your teams and ask them what that means to them what do they collect just in case they're afraid that they have to cover their butts and we need it. What is that? And is it if it's tied to your funding, then go back to the funer and just tune them into saying, "The world's changing. Can we change that?" Okay. Next up, number three, identify your red data, which is what you would literally classify as essentially very sensitive information. government ID numbers, financial information, passwords, credentials, health care information, information about children or vulnerable populations, precise locations where they might be, and combinations of data that could easily identify somebody that deserves much stronger protection. Okay, so this is where the riskier stuff is. So, can you label it and and actually name it as a team? And this is just an activity of you sitting down and really just asking people what is the most sensitive stuff that we need to protect on everybody's desk. What is their version of it? Everyone's going to have a slightly different version, but if you have that, how are you going to think about that a little bit differently going forward? Okay. And give personal data an expiry date. Ask how long do we actually need to keep this for? Okay. So, in every organization, this is going to be different. uh in many verticals in the nonprofit sector, we shouldn't be keeping data that long or if we're going to keep it and it's for the purposes of a study longer term. Like, you know, if you if you ran an organization that focused on mental health and you want to be able to pull data in the future to see longer term impact of your work, maybe in connection with another organization that you work with. This is really important. You want to keep that. But is that just hanging out in some Google Drive somewhere? probably shouldn't be doing that, right? It should go somewhere where there can be some security around that data and you're you're actually investing it and keeping it safe over the long term, which is different than your day-to-day data that you can just keep and get rid of. Okay? So, we just need to be start thinking in things of short-term versus long-term. What do we need to help serve someone versus what are we keeping over the long term because we want to learn from the data over the longer term. Okay, this is one of my favorites. change access from everyone can see it to who actually needs it. This is really critical. In small organizations that are quite nimble, everybody gets to see everything, right? But in larger organizations, volunteers may not necessarily get access to your donor database, right? This shouldn't happen, right? Or volunteers perhaps don't get access to a version of whatever it is for your CRM if you're serving teams. they shouldn't have access to that much personal information about any one individual, especially their healthcare data. So, this is really important. These top five things alone, I think, are incredibly important. And again, I just went into chat GBT and just asked generally, if I'm a nonprofit, what are some of the things around data that I need to be more careful about? And look, here's some tips. Then, what I would normally do with a team is I can say, here's a hot tip. Hey Ash, what could a team just like a small team do in an afternoon for free to take their first steps in reviewing maybe a little bit of a data cleanup project that they could take on over the next two quarters? What is bite-size, actionable, and non-technical? Can you give us an outline of what that could look like? Okay. Turn it into a 90minute data cleanup afternoon. Make the goal very modest. Do not clean everything today. Figure out what you have, what worries you, and what you're going to tackle first. Okay. So, very small steps here. Non-technical over 90 minutes. Here's some steps that you can take. Okay. First 30 minutes, let's go finding data. What does that even mean to us? Where do we keep it? 20 minutes, circle the risky stuff. What is something that we should really be protecting? Because it's deeply personal and very risky. 20 minutes, ask three questions. Why are we collecting it? Who actually needs it? How long do we need it? This is amazing. Okay, then 10 minutes. Pick three cleanup projects. Not 27, three. She's trying to make it like really actionable and easy because we know how busy and time constrained everybody is in this sector. And then in 10 minutes, give each one of those an owner and a date. who owns the cleanup. Simple outcome, one deadline. And do this. Try to pick these in increments and just take one bite out of the elephant at a time. There you go. As simple as that. And again, I used AI to get us this plan. And we might think, well, this sounds like really straightforward, but you would actually pay a consultant to come in and show you how to do this. But you have this tool at your fingertips. And if you did this and ex ask those exact same questions inside of Co-Pilot Claude or any of any AI tool, likely you would come back with something very similar. There's nothing really unique and special about this one except she knows me very well and I work in the nonprofit sector. So there may be a little bit of me influencing the result here in the answers. But again, all of this stuff is really simple and easy to do with any large language model you might have at your fingertips. Nice. So yeah, those are I think it's a really good we figure out how we should actually start fig like determining what data we should keep. But the other part of this is sort of we have some worries about like well what if we delete something and then we're like actually another team did really rely on that and I just didn't know about that. So Fumiko asks I'm afraid that if we delete something it might affect somewhere else in some other team I wasn't aware of. What's your thoughts around around dealing with this? Is there a way maybe some people I mean some of this work is like doing your cold storage archiving of data so it's more contained versus just a whole delete wholesale deleting of your information that you think is probably not necessary anymore. >> This is an amazing case of the just in case data. This is the fear everyone has. So usually what I do is I ask the team when has that ever happened? Have we ever needed to go back in time to get a data set versus like an individual detail on an individual per person that's maybe one or two years old? Has it actually ever happened that we needed something that was 5 years old or 10 years old? And there's always going to be someone in the room that is going to be very worried that it might come up. But I'm telling you right now, the odds that it might come up versus it might get out and hurt you. You need to weigh how much you're willing to risk to have the just in case data be an inconvenience. and maybe you can move past it and recollect it from that individual versus keeping it and suffering the consequences of it being exposed plus anyone else that might be in that data set with it. Okay, so that's a hard question to answer. However, I always say to nonprofits, there are ways that you can reach out to people to recollect things if you need them. And if they know why you're recollecting the data, it's very rare that people say no. They're usually very happy that you said, "Look, we deleted it because we're trying to protect data sets. However, we're going to need to collect it again because of XYZ reasons. If I was an organization or a person that you served, I would be very happy to hear that you were being a good constituent of my data and got rid of it and had to come get it again." Great question. So now I want to take us a little bit bigger picture. So we've talked a lot about the kinds of data that we need to be cautious about about captured. You know the the information that's personal, the information that was very program specific, but there's a whole other set of data that we produce in the nonprofit world that is data we want others to consume. I think of back when I worked at the Dam Suzuki Foundation. We were collecting data around like caribou and and their interactions with traffic and we didn't want that private. We wanted everyone to see that. And so I'm wondering if you can talk a little bit about this vision of like what if nonprofits could more easily see each other's data that they want the the sector to know about that they want funders to know about so that we can learn from each other because we all have like our own specific little view on that data like what are people doing around that around actually publishing their data and finding ways to collaborate ac with each other. So using AI and this is where AI comes into the conversation is the ability to analyze and interpret large data sets. And so this is the opportunity in the nonprofit sector that has really never before existed at least at a reasonable price point. Okay. So in the past if we wanted to report on sector like in a certain vertical across the sector there would have to be a dedicated IT team right dedicated data analysts somebody who could interpret all this data go across all the different organizations that are supporting an individual goal because I think we all have come to realize that nonprofits come in different shape and sizes and I'll I'll give an example in one industry that might help uh or assist in mental health there's versions of it that are for kids. There's versions of it for men, for women, for folks uh that are marginalized. There's a lot of different folks that are tackling that problem. What if they can contribute some knowing that comes from their interactions to a bigger pool so that we could better understand how humans struggle with mental health, right? And so we need to start thinking about our data in two ways. One is the operational opportunity, right? which is what I've talked about already and shown you the example of how you get your operational data back down to a place that it's manageable that you feel like you're in control of protecting it again. The second opportunity is what is data that could live that higher level that you could start collaborating with your peers in your sector to start really learning from each other and really reporting on the impact of your work at a larger scale instead of just as individuals. Now, there are certain people out there that are thinking about what that would look like as far as having a vault that could be a place where some of this could live. And I think this idea will be one of the biggest opportunities in 2027 for nonprofits as we think about how do we make sure we have the right funding? How are we telling the story of our impact? Can we do it at a broader scale to raise more awareness for the importance of our work? Because let's face it, the world is not getting any less complicated. The funding is not getting any easier to get and the needs of the world are increasing when especially when we have a productivity crisis. We have so many things going on in in our world that make it difficult for humans to just be supported and have their basic needs met. So that's where the nonprofit sector comes in. And I think a lot of what we can be doing with our data in a bigger sense is thinking about what could be we be collecting that actually tells the story of our impact and start working towards working with our peers to make the story even bigger than just ourselves. So there's the step one of telling your story with your data and your impact. And then there's the new new emerging opportunity which is do you know two or three peers in your space that go after a similar goal to you but maybe in a slightly different way and how could you start stringing those data stories together to show the deeper impact on either your collaboration or the fact that you all exist. Right? So there was always this goal to start to amalgamate nonprofits that were sort of working on similar things. This is something I hear all the time when I meet for with nonprofits, especially when we talk about AI literacy and trying to bring their organizations together. They're starting to see efficiency wins that they could get and then they're like, well, you know, the thing we always wanted was to maybe create a more efficient way of working between nonprofits that were are going after similar goals. How can we start doing that? We've never really been able to get there because nobody really wants to give up what they're doing as individuals. But what if we could make that first bridge our data and start telling a bigger story with our data? And I I honestly think that will be the next big opportunity for nonprofits in 2027. >> Yeah. Like and there's been some interesting work in this field. um you know the data commons work has been happening with Google who's been sort of working with groups like tech soup or giving Tuesday or groups of philanthropy organizations to say how can we basically create these do common sets of data and also standards around >> how do we measure impact so we can actually then go tell the story >> it's always been the holy grail right across the whole sector because >> you've got a thousand food pantries working in in every community, but what does the story retell and what is the need so we can actually go to funders to government? And that's been very hard to do, but as you say, the tools are starting to emerge and standards are starting to emerge to allow us to tell these overall common stories of impact. >> Yep. Love it. So, I've got a question here from Debbie who is getting into the really practical side of things who says, "Yeah, is there a simple way to anonymize data of our customers?" Um, they've got like they've got housing, they've got finance, they've got health, they've got all kinds of very personal data. And so, right now, they've got this data that they need to keep secure. And what they're doing right now is they have a spreadsheet with people anonymized >> and they delete it. And then they have like basically they create a coding system right now that seems like it's manually done which just gives them a number and maybe initials or a phone number. So they have sort of something that's personal like a personally identifiable but anonymized. Is there a more clever way to do this than this timeconuming process that poor Debbie stuck in? >> Debbie, I love this. And let's go back and I'm going to show you another trick here. We're just going to ask AI. And so I'm going to assume I'm going to assume maybe you're using co-pilot or maybe chibi and ideally if you're using an enterprise AI product like say a copilot and I'm just going to I'm going to copy and paste your question right in here Debbie so that's word for word what you said and I'm just going to add a a little bit of around it here so that my uh agent knows what I'm doing here. So, hey Ash, this is a question from Debbie and I want us in this example to say that she's might be using co-pilot. So, we know that uh you know this is one of the great AI loves the spreadsheet problem, but we want to make sure the answer is non-technical for her team. So, she could take it literally copy and paste it into say her co-pilot and help her step by step get rid of the manual. I call this the list. Am I allowed to swear on this podcast? Too late. >> Just done it. Keep going. >> So, sorry, but everybody knows what I'm talking about. This is one of those items in a nonprofit that keeps you away from your mission, right? Work like this, I think, is the best kind of work that you can give to AI. So, here we go. I'm going to put this in here. Sorry, Ash. I also gave you all my commentary around that, but can you give Debbie a stepbystep guide on how for free using Copilot this afternoon she could get out of doing that work forever? Debbie, this is the exactly the kind of repetitive work AI should help eliminate. But do not solve your privacy problem by pasting sensitive data into the wrong AI tool. I'm checking the safest co-pilot route so that we can make this genuinely usable this afternoon. Okay, she's still thinking. Okay. And mine is a Canadian instance here. So, she's going to follow Canadian privacy guidelines. So, for you and yours, uh, wherever you're located, Debbie, uh, it will pick out all of those for you. Hold on. The durables are working. Love how she's like, "This is a There you go." Okay. So, do not give co-pilot client data. Give it structure. Open your organization's approved 365 co-pilot chat, ideally using your work account, and give it column headings and fake examples only. So step by step here, basically stop using identifying information to create an anonymous identifier. Okay? And it's just going to give you the prompt on how to do this. Get Copilot helping you by building the query. And here's exactly what I would paste into Copilot. So, she even wrote you the prompt. Okay. So, this is really powerful. So, Debbie, what I want you to do if you have an AI tool um and you want to get out of a human doing this work, go to your AI tool, copy and paste that exact question in there, and then I want to give you another tip, which is ask that AI to interview you for any additional information that it might need in order to make this easier for it. Okay? So, we want to assume in the question that you know exactly what you're doing. But what we want to do is give the AI the opportunity to ask clarifying questions so that it doesn't get it wrong. So, ask your AI to just ask you any questions for any information that it needs in order to achieve this goal of getting out of this and knocking this thing off your list. Okay? And then ask it to write the prompt as well. Okay? So, that's the series. bring it the inquiry about how do I solve this problem. Be really clear on what success looks like. Success looks like we are no longer manually manipulating this data so that we are creating these identifiers that are not real to protect our people. This is amazing. And also success looks like my team gets 3 hours back every week from getting out of this work. Check. then ask the AI to ask you clarifying questions around that problem to make sure it has all the information that it needs because it needs uh it will you know we don't know it all and when I use my words when I prompt by the way when I talk I give it way more context than when I type. So just a little bit of a pro tip that if you can throw away your keyboard and just talk to either whatever AI tool it is in your realm it is so much more effective and actually frankly just so much more enjoyable just in general. But anyways, so ask it for any other data that it might be missing. Once it's gone through that interview process, then you can say, "Please write me a prompt that I can put into the AI tool of my choice to get this done." And you don't have to be a prompt engineer anymore. Okay? The technology has moved and evolved so quickly over the past 7 months. Prompt engineering is no longer a skill. I would say to the majority of people, if you can get a four-year-old dressed for a snowstorm and out the door, you have as much skill as you will ever need to be a great prompt engineer inside of AI. That is a much more difficult skill addressing a toddler than it is uh using AI now because of how progressed all of the platforms are. Okay? So, use it as a sounding board on how you might solve these problems. And to me, this doesn't count as bringing AI into your organization. What it is is actually just getting a coach on the outside helping you see options that if you're trying to solve it with technology and you're not technical, just tell it. Give it to me like I'm five. That's too technical. I don't understand that. Here's a screenshot of the error message or whatever I got when I did it. Can you help me with that? It will do it. So, the screenshotting, asking for clarifying questions, and having it interview you, and just asking it to write your prompts is the skill for 2026 and 2027 if you want to get your start in AI without the adoption going too far. >> Awesome. That's really helpful. Um, and so we're now coming towards the the latter half of this event. So, I want to get into like the tools, the fun stuff. So, first of all, what are you using right now that you love? Like you're like, what are your like at these areables? Like without this, I would feel like I'm operating with my hands tied up behind my back. >> I think I alluded to it, which is I can work with my hands tied behind my back because I don't use my keyboard anymore. So I think you know my top AI tool recommendation isn't an LLM. It is whisper flow so that I can speak to this technology in the language that it has been trained on which is natural human language not my typing. I hear so many people say that they think when they type. But last time I checked you were not born with this in your hand. >> This came much later in your life. And so I would say this is the biggest advantage to uh some of this technology and that we can really sees our workday as more of a conversation with a partner that's helping us as opposed to a technology that we're typing and talking to. Okay, so that's my top one and it's called Whisper Flow, the one that I use WPR flow. It's fantastic. I can't say enough about it. has completely changed the results I get from AI too as well because when I talk in my natural voice, I'm more likely to go into more detail and give it more context than it needs. So, it just its performance for me just went through the roof when I did that. My second one is obviously Claude is a real standout for productivity and running an operation with more efficiency, especially when it comes to just pure operations and when there's repeatable workflows. Claude Co-work. And I know Claude has offered a new $8 a month license for nonprofits. I strongly recommend people check it out if you haven't had any experience yet with Claude. It is it's an absolute game changer. I would say for a business owner and somebody that has a day that is bonkers and weeks that are bonkers. I was actually able to take a vacation for the first time in my life that was real because I had Claude running a lot of the repeatable activities for me while I was away and supporting my people that support me. You know, it doesn't replace anybody in my organization, but it sure keeps us organizes organized and and keeps me out of having to get into the weeds of a lot of things and gives them what they need to keep their work moving forward. Outside of that, I still use chat GPT a lot. It is great with imagery. So, if I'm doing any sort of social media, like if I was in doing work that I do in the nonprofit sector, which for me right now is sports because of my kiddo and just my history, uh creating social media images around, you know, our I'm going to give you an example. We were just in Finland for a tournament and we had all the the sponsors logos and I was able to create a photo of the team, you know, their backs with their big W's up in the air and on their backs and all the sponsored logos in it and saying thank you to our sponsors. Took me like 15 seconds on the bus after the game and we were able to post that on social media. So, there's a lot of content wins that I love in chat GPT uh that have been incredible. I am actually not using Google as much as I used to. I have found that they're in a little bit of a lull right now, but I'm sure they're going to release something that will blow my socks off and we'll be right back in the game. But right now, I'd say my top ones for sure is making sure that all my meetings are transcribed so I'm able to use those transcriptions in moving things forward on my desk, even when I don't have time to move them forward. Whisper Flow, so I can talk to all of these tools. No matter which one I have going on my screen, I can talk to them all using that application. and then claude for my operational work because when it comes to spreadsheet problems, it is far superior than anything I've ever experienced. And I'm spending thousands of hours in these tools, right? So, this would be my top recommendation. And because Claude just came out with that $8 promo for nonprofits, it's it's been fantastic for the nonprofits that I've worked with that have adopted and just said yes to Claude and then learned just built their comfort and their literacy and you know their fluency over time inside of Claude and are not paid by Claude by anyway. So I just want to let you know I'm not paid by any of these companies. I'm just a user and a connoisseur of the tools. >> So let's talk about what we can do next. So one of the things is you and you've already outlined like the basically the 90minute exercise which I think is >> really a lovely way to sort of say how do I tackle these problems in in a way that's not going to be so consuming but but if we're just maybe on the place where we're trying to understand where does AI fit into our organization where would I go for support if I want to start understanding AI and data and how it can fit into my organization. Two is just setting your goals around data. I would start there before we start talking about bringing AI in to do anything. I want you to sit down with your group and just have like a moment where you think about your data just as a topic, right? It's very rare that we sit down and that's actually an agenda item. What are we doing with our data? Okay, so before we jump to tools and solutions, sit down and just really think to yourself, okay, is there a bigger mission for our data, right? which is that top level opportunity that we could start learning and sharing our impact more broadly. If no, throw that away. That's not for everybody. Okay. But then once we know that's out the window, it's in or it's out. Let's go to the other opportunity which is what have we been doing to date and how do we from this point forward? I know a lot of you probably cannot unsee what I've done here today and I'm hoping that's true that you go back and you look at that the the amass of all the spreadsheets and just say okay what about this is risky and just if you have an AI tool at your disposal that you can just like I said you don't have to disclose anything about your organization but give it in broad strokes what you're collecting and what you're doing with it and how you're storing with it and just get some advice on how to create a 90minute like I a little power session with your team to take one bite out of the elephant and use plain language in in your LLM. Don't try and type in something technical that you think is going to understand. Don't use it like Google. Pretend like you have a data safety assistant and you could just ask it anything. And also judgmentfree. So just tell it the truth. Nobody's it's okay, right? It's better for you just to be able to sit there and say, "Okay, here's what we've got. Here's who we serve." Broad strokes. What would be the first three steps in us unpacking what our what our next step could be? What do we need to do? Just start small because when we think about it and we make it too big, we don't do anything. And I think we all know this is probably going to be pretty big, okay? But just do something every few weeks to just take a bite out of that elephant one at a time. And if you're an organization that that wants to bring AI into your teams and you want to build that fluency because everything that I showed you today is tool agnostic all of them do for the kind of work that we don't need a Ferrari to do everything. We in most cases just need a bicycle to do things. So we don't need to throw the Ferrari. Nobody needs any advanced skills. We just need line of questioning and just the ability to articulate our problems. Okay? So, I don't recommend using free tools for any reason because the models do not represent where the technology is at. No AI company is interested in making your life better, easier, or for you to make any more money for free. Okay? So, it does require a subscription between 19 and $29.99 a month to get a real answer like I got. Right. Also tend to give more promises around data privacy like they're not going to bring that. >> Yeah, you have more options. But I wouldn't say that that solves the problem, but it gives you more options. But I would say other than the privacy, it also actually gives you access to the real models and the real answers. Okay? So just know free tools give you free results. Don't use free tools. If you get a paid tool and you start asking it these questions, you're going to get higher quality answers and higher quality support that you paid for. Okay? So, very limited amounts of funding gets you through that door. But then, if you see results and you're like, I want to go deeper, but I want to talk about this and I want to learn in a classroom environment with other nonprofit professionals, we have an incredible program at the nonprofit chamber that supports nonprofits for it. And I can leave the link uh in this uh webinar for folks to take a look at it if you're in Canada. And if you're down in the States, I've got an incredible partnership with an organization called Learn It. And there's a Kickstarter program that I've designed that goes after in four sessions, four 90minute sessions. We get at the heart of your adoption, give everybody the fluency that they need, and we just get you going and that's it. Four 90-minute sessions. That's all you need. It really is not a huge time commitment, but it is. I know how hard it is to find 90 minutes in a day for a nonprofit. So, believe me, I know the struggle is real. So, four sessions, we get you guys up and rolling. And honestly, if you have any questions, I'm very accessible. Hit me up on LinkedIn. I'm happy to take any one of those questions. And I saw that there's a really big long one in here from I think Kieran, uh, which I'd love to address. And I understand the pain of things like deep folder trees and things like that that you're talking about. It's very complicated and unfortunately there isn't an easy answer. Uh so I wish I could give you a magic one but they've all sort of painted us into a box on how complicated our data structures are in our organizations. Like my version of that is a SharePoint nightmare, right? And so we all have these problems, but learning in community when we hear about how other people are tackling them, other things that have worked for them, this is how we learn and we learn in an accelerated way. So that's what I strongly recommend people to do. And I said, use these tools to ask these very first basic questions. There's no such thing as a dumb question with these tools. And they're happy to support you and cheer you on in your mission. And you don't need to disclose a heck of a lot of the secret sauce or anything personal in order just to get some basic advice to get you going. Amazing.