AI’s Impact on Nonprofit Visibility and the Changing Landscape of Foundation Support This week’s episode of the Nonprofit Newsfeed, the hosts delve into the evolving dynamics of AI’s influence on nonprofit visibility and the current state of foundation support. Main Topics: AI’s Knowledge Windows and Nonprofit Visibility: George and Nick discuss a new beta feature from OpenAI’s ChatGPT called “knowledge windows.” This feature presents information about organizations directly within the AI interface, potentially reducing website traffic as users get essential details without needing to visit the source. This development could significantly alter how nonprofits engage with potential donors and supporters, as AI becomes the gatekeeper of information. Foundation Support and Nonprofit Perception Gap: A report from the Center for Effective Philanthropy highlights a significant perception gap between foundations and nonprofits. While 93% of foundation leaders believe they understand grantee challenges, only 54% of nonprofits agree. The report also notes that foundation responses have weakened post-COVID, despite increased demand for nonprofit services. Podcast: Play in new window | Download View Episode Transcript This week on the Nonprofit News Feed, of course, brought to you by Whole Whale B Corp Digital Agency. My name is George Weiner. I’m the Chief Whaler of Whole Whale. And I have Nick Azulay, Digital Strategist at Whole Whale. And we’re here to, as always, geek out about nonprofit news, trends, things we are seeing. And Nick, why don’t we start off by something curious that I ran across on ChatGPT? Yeah, George. So our first story is an interesting one, an inside scoop, rather. We, George, have experienced among the first users to do so, apparently, a new beta feature inside of ChatGPT. What is that beta feature? It’s not particularly sexy, but for nonprofits, it’s really important. And that is the rollout of Knowledge Windows. So George, if you are subscribed to our newsletter, you’ll be able to see a picture of this. But what are Knowledge Windows, George? They are results that come up to a query or a prompt that you put into a AI chat. And essentially, what it is, is a window that, instead of linking out to a website, fills a window with knowledge, kind of ingesting that website, just as a Google AI overview might, for example. So George, we are seeing OpenAI testing Knowledge Windows inside AI chat services. What does this mean, George? That AI now has found another way to answer questions, reference your website, and in some cases, surface actual web content without ever sending anyone to your organization’s website. So George, for example, when someone asked ChatGPT, top animal welfare charities, they are now getting a slick pop-up card with your mission. They might get the foundation founding date, key facts, all pulled into the AI’s interface. No clicks, no visits to the website required. As an organization, what does this mean? People are learning about your organization without ever accessing your content at its source. George, this is a big development within AI, particularly as we’re studying user and donor behavior and how they’re using AI systems to interact with their organizations. And George, we suspect this is leading towards something very specific. But at this point in time, George, why don’t you just tell us a little bit more about these Knowledge Windows and what it means for organizations getting surfaced within AI prompts? So notice I am currently showing this. For those of you who are just listening, I am on ChatGPT 5.2. This is a paid subscription. So I have not been able to reproduce this on the free version of ChatGPT. And in this case, as I go through, I click on a link like the Humane Society of the United States, and I go and I get this little sidebar window, this knowledge source. Sometimes sources show up on the side. Sometimes they don’t. If I click on those sources, then I am sent somewhere else. Sometimes, honestly, I go to this sidebar, and on the sidebar, there are no sources whatsoever. So in this case, I am looking at the Best Friends Animal Society. I’m interested in animal welfare, and I have what seems to be a link that I could click on. I click on it. It opens the sidebar. There is nothing. And I repeat, zero things I can click on for the source of the information or to click to go to that website. So I’ll just pause on that and just be like, the logical second-order effect of this as a user interface is that you get a zero click, zero click from people that are exploring your concept with regard to your AI brand footprint, right? People that are surfacing your information inside of an AI interaction. Not great. The other sort of potential second-order effects, where I see this going is, what is the action I might want to take? Now, imagine this for shopping, right? And shopping, what happens? We try to disintermediate the endpoint. Could I shop and buy that thing right there? How do I make it easier for this person? How do I use AI agents to support my shopping? How do I use AI agents to potentially make a donation? Let’s say I wanted to donate. Oh, why can’t I just have a donate button or the actions I could take with this organization within this window right here, allowing ChatGPT to become the gatekeeper? What’s more, if I’m clicking on something like the ASPCA or AlleyCat Allies, I have key information that I could, I don’t know, say run ads against. In this case, there happened to be sources. There weren’t, and they’re not always. But I’ll tell you what you could also do. You could shove some advertising into this exploration window. What does ChatGPT need to do and begin to optimize for? Advertising real estate. Where might I find such real estate, you ask? Right here. I don’t know why we happen to get lucky and be able to see this side window. I don’t know why more people are not talking about this, but I think we should pay attention to this right meow, pun intended. Yikes. All right, George. Okay, so we have a demo of what these- Also, shameless plug here also for the fact that somebody, we need to start working around the verified giving protocol. Verified giving protocol, something that I’ve talked about before, is the concept of needing to find a way that would route not just the money, but also the data directly to the nonprofit. Because this is going to happen. They’re going to disintermediate your donation button, but it should still be under your control, under an official way of routing that information and access to your provider of choice, whoever it may be. So, all right. End rant. No, George. I think that’s so important, right? The theme here is ChatGPT, AI in general. They’re doing everything remotely possible to generate revenue for themselves while keeping you in their platform. And we need to be vigilant and see how those products develop and what that means when it comes to cultivating donor flows and recruiting volunteers and XYZ number of actions that your website is promoting, right? They are siphoning our audiences and it’s important that we are able to, in some ways, recuperate the flow. So, George, really important. I want to pivot slightly within the same theme here. We’re really studying very hard, George, what AI is doing for clients and the representation of our clients and our organization’s brands within platforms themselves, right? George, some of our clients are getting tens of thousands of real, authentic web sessions from LLMs directly, purely based on people seeing a result and clicking on a little link, taking them to the website. Tens of thousands of sessions. But, George, we’re starting to see that AI, at least for now, I don’t think this will be permanent, has a junk link problem. And we are seeing it increasingly within the social impact sector. What do I mean by that, George? If you Google, or I’m sorry, if you search, see, we’re no longer, I have to train myself not to use that terminology anymore. If I prompt an AI chat, like chat GPT, top human rights charities to donate to, the top two results are very predictable. It is Human Rights Watch and Amnesty International. And you’d say, great, that makes a ton of sense. The link next to Amnesty International probably takes me to Amnesty International’s very, very large, complex global web infrastructure that includes a global organization, national organization, and probably 100 regional or country level websites. No. What does it link to? It links to, George, a super low quality, barely usable farm link website that was written by AI that just farms content for AdSense revenue. George, I don’t think this problem is going to exist in perpetuity. But right now, major organizations like Amnesty International, that probably have larger websites than 99.9% of the internet are losing out to, excuse my language, s****y websites that are siphoning away that traffic and siphoning away those donors. So something to watch out for if you are a nonprofit or a major NGO like Amnesty International, which has billion dollar revenue. Yeah. And to be clear, it’s the top list of this for this and top list of this for this. Those types of sites are getting ingested and used as validating entities for how to answer these questions. So yeah, yeah, it is something you surfaced and something we’re exploring. And I do think it is important to note that it’s for now, not forever. These rules are changing quickly. But right now as it’s played, I thought that was a little concerning as well. A tad bit concerning. All right, George, we have some interesting tidbits I want to turn to. This comes from The Nonprofit Times. George, Whole Whale and Nonprofit Ifs, that’s nonprofit.ist, have released their 2025 consult cost and compensation survey in partnership with T