If you’re wondering what role artificial intelligence should play in your learning business—or how to prepare your people and processes to use it effectively—this episode of the Leading Learning Podcast is for you. Many learning businesses are just beginning to grapple with AI’s implications, and the gap between the potential and the current reality is big.
Amith Nagarajan, chairman of Blue Cypress and founder of Sidecar, joins co-host Jeff Cobb to talk about how learning businesses can build the capacity they need to thrive in an AI-driven future. They discuss patterns from past waves of technological change, where associations stand today with AI adoption, and the risks of standing still. Amith also shares practical steps for leaders to get hands-on with AI, the importance of reducing friction, and how AI tools are reshaping knowledge, insights, and personalization.
To tune in, listen below. To make sure you catch all future episodes, be sure to subscribe on Apple Podcasts, Spotify, or wherever you listen to podcasts.
Listen to the Show
Access the Transcript
Download a PDF transcript of this episode’s audio.
Read the Show Notes
Celisa Steele: [00:00:03] If you want to grow the reach, revenue, and impact of your learning business, you’re in the right place. I’m Celisa Steele.
Jeff Cobb: [00:00:10] I’m Jeff Cobb, and this is the Leading Learning Podcast.
Jeff Cobb: [00:00:17] Technology is always reshaping the world and therefore how learning businesses operate—from the early days of the Web, through mobile and social, and now with the rise of artificial intelligence. These shifts create both opportunities and challenges for learning providers.
Celisa Steele: [00:00:33] Our guest in this episode, number 461, is Amith Nagarajan. Amith is the chairman of Blue Cypress, a family of companies that help associations embrace AI and transform how they deliver value. He’s also behind Sidecar, which aims to educate a million association professionals in AI by the end of the decade.
Jeff Cobb: [00:00:54] Amith and I talk about patterns he’s seen across multiple waves of technological disruption, where associations stand today with AI adoption, and what leaders need to do to be prepared for the future.
Celisa Steele: [00:01:06] You and Amith also get into some of the applications his companies are building and what those tools reveal about how associations can use AI for knowledge, insight, and personalization.
Jeff Cobb: [00:01:18] And we close with a focus on the human side of all this—how learning businesses can lean into the distinctly human opportunities that AI won’t replace.
Celisa Steele: [00:01:28] If you want to better understand the implications of AI for your learning business, then this episode is for you. The conversation kicks off with Amith talking about what he’s done and is now doing with his companies.
Amith’s Background in Tech and with Associations, from Aptify to Blue Cypress and Sidecar
Amith Nagarajan: [00:01:45] I’m probably known originally, or perhaps best known, for having started an AMS company in this space back in the ‘90s called Aptify, which I sold about 10 years ago. Since then I’ve been focused on a family of companies called Blue Cypress, an AI-oriented shop. Essentially, we have a number of different brands within our family that provide a combination of AI learning, AI services, and also AI software to the association sector very, very specifically, and we’ve been doing that for about 10 years. We’ve been in AI for a long, long time (in AI speak) and are very deep in it. That is the focus for me—how do we take AI and help associations not just incrementally improve but radically transform their business models, their value delivery, and how they operate internally as well?
Jeff Cobb: [00:02:37] We’re going to talk about AI more. You’re definitely somebody to be talking to about AI. But I was struck, as you were talking, that it’s already been a decade since you sold Aptify. I know you were in the association world for years before that. I believe Aptify was your first company, if I’m correct.
Amith Nagarajan: [00:02:55] Yes. I’m from California, grew up in Silicon Valley, have been programming computers since I was in single-digit years, and started my first company at the age of 17, which was Aptify. It did not start out in the association market. I didn’t know what an association was back in ’93, when I started it. But, back in those days, I had a knack for software development, and I decided to build a toolset that would help people build database apps. That turned into a robust platform in the early and mid ’90s. And then I happened upon the association market in the mid ’90s and just fell in love with it over time and focused exclusively on it for quite a while.
Jeff Cobb: [00:03:33] So you have been doing it for quite a while. As you said, it wasn’t your original intent, but Aptify took you there, and you’ve been there all the way through Blue Cypress, which has a portfolio of companies that serve associations. What’s kept you engaged with this sector over that period of time?
Amith Nagarajan: [00:03:52] The purpose-driven nature of the people. The community is fantastic. The people in this market are focused on trying to have an impact in the world. By helping associations do what they do better, we feel like we’re making an impact. I feel like I’m having an impact. There are lots of different sectors, lots of different niche markets that you can focus on as an entrepreneur. I happened to find this one by accident. It wasn’t an intention. I just happened upon an association once upon a time back in the ‘90s and then found another one and then found another one, and I said, “Oh, well, this is actually like a whole market.” I stayed because it’s a great business opportunity. It’s a robust marketplace. Once you get to know the market, you realize it’s quite substantial, especially when you look at it globally.
Amith Nagarajan: [00:04:37] But I stayed, and I came back to it after selling Aptify because I felt a really deep connection with the space and wanted to help people in this space not just survive or stay relevant but rather thrive—drive aggressive, transformative change. That’s what I’ve always been about. It’s always been exciting to me to help people do big things, and that’s what we’re trying to do at Blue Cypress. I’ve been doing AI stuff since the early 2010s. We started getting into AI at Aptify in the years preceding the sale of it. We started getting pretty heavily into AI and then ended up spinning out those technologies into what became rasa.io, which was our first AI software company way back in the day. We still own that company and are growing it. It’s doing well. It’s been a while, but we’ve seen AI from lots of different stages of development and applicability to the market.
Amith’s Take on the Big Shifts: The Internet, E-commerce, Mobile, Social, and AI
Jeff Cobb: [00:05:27] I’d forgotten you had gotten into it that early. In some ways, our stories are similar because I too was building a piece of software that was not intended for the association world and found myself in the association world and have been there ever since. But you were there a little earlier than I was. You’ve seen some pretty big shifts from the rise of the Web. You started out in the early ‘90s; the Web was still in its infancy then. Then into the social media craze. I remember that hitting associations, and everybody thought it was going to put associations out of business. And now AI. As somebody who’s observed all of that, are there patterns you see repeating or lessons that you think leaders should draw, particularly right now, if they’re thinking about “Oh, AI is hitting. What do I do with that?”
Amith Nagarajan: [00:06:11] Yes, there are definitely patterns. Going through those major transitions—obviously the Internet and the Web, e-commerce, mobile, social, all these big shifts, and now AI, as you mentioned. Common patterns are, first of all, that people tend to overestimate the near-term impact but radically underestimate the long-term impact of technologies that are truly representative of phase changes. They’re not just an incremental innovation, where you went from the iPhone 15 to the iPhone 16, and you got some new widgets on your screen, but rather something that can fundamentally transform the economic relationship between buyers and sellers. And I mean that in the broadest possible sense. If you think about what happened in each of those transitions over the last 30+ years, you’ve had a significant shift in terms of information, information advantage or disadvantage on the demand side or the supply side of the curve. You look at distribution.
Amith Nagarajan: [00:07:09] I talk a lot about AI as being the third exponential curve following two others. One is the compute curve, which everyone talks about, often known as Moore’s Law, the exponential rise in compute power relative to cost. And then what ended up happening with the Internet was the distribution curve. We flattened out the cost of distributing nearly anything that can be digitized down to essentially zero. But now what’s happening with AI is the cost of reasoning, the core component of what we consider intelligence, is going to zero. We don’t know what to do with that. But coming back to the question of pattern recognition, as costs approach zero for something that was once scarce and is now becoming abundant, it reframes the relationship between consumers of value and producers of value—or buyers and sellers, to put it another way. People often don’t anticipate the magnitude of the impact long term, but they think about the hype cycle, and they’re like, “Oh my God, this is going to change everything.” And then they wake up a year later like, “Well, actually things haven’t changed that much.” It’s a false sense of safety in the status quo that it gives people.
Amith Nagarajan: [00:08:12] I’ll come back to my thoughts on AI on that at some point in this discussion, specifically because it is moving faster than these other curves. As crazy as the Internet and mobile and all that was, it’s nothing compared to AI in terms of speed. My main point that I’m trying to underscore in terms of the trend line across these various changes over time is that, when people get this sense of security, that it didn’t really change that much, and then the bigger tail on the change, the transformative impact comes along, that’s when they get caught off guard. Back in the ‘90s, people were saying, “Oh, associations are going to be out of business pretty soon. There are all these companies that are going to be coming out with verticalized Web sites that are going to take over associations.” Associations were still there in the early 2000s, and they’re like, “Well, that didn’t happen.” But, sure enough, the Internet has displaced the activities that associations have historically done in many ways, and AI will do that as well.
Amith Nagarajan: [00:09:08] My general observation is that people tend to not move quickly enough because they get…. Associations are generally not the first adopters of anything. We’ve talked about that before in our conversations before this pod. It’s a frustration to technologists and to people who are trying to push the boundaries of what’s possible. Some associations move a little bit faster than others, but generally they’re slow movers for the most part. Slow movers tend to have their perspectives reinforced by the initial hype cycle being greater than the initial return, and then they get totally caught off guard because the bigger wave is coming on the back end of it—that underestimation of the long-term impact that I’m referring to. We can’t afford to do that with AI is the bottom line.
Jeff Cobb: [00:09:52] What are examples of organizations in those earlier waves that either got it right—they timed it correctly; they saw what was going on and reaped huge benefits by what was happening with the Web, social, or whatever—or, conversely, maybe one or two that didn’t and got clobbered by it?
Amith Nagarajan: [00:10:14] There was a pretty big shift to the world of education delivery online back in the whole phase of Internet and mobile. You know that world far better than I do, Jeff. But the concept of people delivering traditional classroom-based, synchronous training when asynchronous was getting better and better and better, more dynamic, more capable, the learning outcomes were increasingly achieving the goal of the student. And so a lot of associations back in the ‘90s were like, “No, distance learning, remote learning,” as we used to call it back then, “is just not as effective.” Of course there are definitely, still to this day, tremendous advantages of synchronous, in-classroom, face-to-face learning, but the modalities have improved so much that many of the people who were thinking the old way—they weren’t even willing to dabble in online asynchronous learning—got crushed. You see associations where they had significant revenue declines, where they weren’t focused on that yet. You see, for example, medical specialty societies. There’s a number of them that I’m quite familiar with that have thrived in this era because they’ve invested in making it possible to do distributed learning well through the Internet, through mobile, but embracing the modality for what it is.
Amith Nagarajan: [00:11:20] You can say to me, “Oh, well, AI will never replace humans for X, just like the Internet will never replace face-to-face for X.” That’s true to some extent. Maybe eventually it’s not. People might say, “Oh, you’ll never stream video on the Internet. It’s far too slow.” That’s just a matter of time. But that’s just an example. There are certain things that are fair to say, like our biology mandates that we have connection, that we are in close proximity with others in our species. It’s not psychology; it’s biology that dictates that. We’re always going to crave elements that technology doesn’t displace. And that’s good, I think. That’s really, really good. The point would be that people forget about the new. They’re saying it won’t displace the old, but what is it going to do that’s different, that’s either complementary or, in some cases, crowds out the need for the old, even if it’s not exactly the same? It’s still a substitute good if people buy X instead of Y even if X is not an exact substitute in terms of its functionality if people stop buying the other product.
Where Associations Stand with AI Adoption—and Where They Should Be
Jeff Cobb: [00:12:27] Where do associations stand right now with AI adoption, and where do you think they should be?
Amith Nagarajan: [00:12:35] If I were to use a report card type of metaphor, I’d say it’s not looking so great today, but I’m optimistic about the future. The way I would look at the current state of affairs is that I think about education as the leading indicator for associations’ ability to adopt AI. You can go to vendors like ourselves or any other company and buy software, get help with services, et cetera. There are lots of great things you can do that way, but that doesn’t improve you. What we are always focused on—and we do some surveys around this—is how much education are people taking in, whether they’re the CEO or a first-year grad coming into the workforce? What level of investment, in terms of time and dollars, are people putting into learning? And it’s very minimal right now. It’s very, very minimal. People are doing almost nothing on average. Now, there are some associations that are putting some effort into this, and that’s fantastic. They’re going to see tons of dividends from that. The bottom line is that no vendor, no vendor at all, can solve your problem for you. You have to be knowledgeable about this stuff.
Amith Nagarajan: [00:13:38] The way I describe it is, let’s say we’re off to a board retreat, you and I, along with our board, and we’re association execs. You’re the CEO. I’m the COO of some association. We’ve got 10, 20, 50 board members coming, depending on the size of the association’s group. We’re going to figure out what the next year [or] two, three years looks like. How in the world can you and I possibly stand a chance at doing a good job in building a strategic plan if we don’t understand AI quite deeply? Maybe not down to the bits and bytes, but we need to understand its capabilities. And the way you understand its capabilities is by unpacking it, working with it, learning it, and you can’t delegate that. You can’t delegate that to a technologist. You can’t delegate that to an outside vendor. You have to make the investment in time.
Amith Nagarajan: [00:14:26] The good news is it’s not a PhD. You don’t have to go to school formally. You don’t have to spend months or years doing nothing but AI learning. That’s impractical. When I speak on the subject, Jeff, I always tell people block off 15 minutes a day every day. Listen to a podcast. Watch a video. Play with a particular tool. Read a book. Whatever makes sense. Hopefully a mix of those things. There’s nothing quite like hands-on. Use the mix of learning resources, but do it every day. Just like working out, just like anything else you practice, you get better and better. If you practice with AI at least 15 minutes a day every single day for the rest of this year, by the end of the year, I would be willing to bet you’d be amongst the top 10 percent of the knowledgeable people in AI. We’re recording this in the late summer timeframe, so you don’t have a ton of time. But, if you get started now and spend 15 minutes a day, you’ll be further along.
Amith Nagarajan: [00:15:17] There are lots of free resources. There are resources like Sidecar’s (my main focus these days), which try to help associations learn AI at scale. In fact, our mission for the next four and a half years, by the end of the decade, is to educate a million association professionals in AI. Whether it’s at the top tier of our AAiP [Association Artificial Intelligence Professional] certification or if it’s attending a free Webinar or downloading a free book, we don’t really care. Our goal is to touch a million+ association folks in that time. But that’s still a drop in the bucket compared to the number of people that work in the space, both as paid staff and volunteers. To me, the report card—going back to that—is not great today, but the good news is we can change it quite easily.
Amith Nagarajan: [00:15:57] My bottom line for the prognosis for the future is, if associations make that investment in education and put on their creative thinking hats, they have all the opportunity in the world. AI has lowered the cost of everything. It’s lowered the cost of software development, of writing, of building anything from scratch. The barriers to entry for associations to get into completely new lines of business with completely different ways of thinking in order to solve the problems that their members will have in one, two, or three years are better than ever. If you’re willing to think creatively, take a little bit of risk, do some experiments, to, most importantly, invest some time, I think the future is extremely bright for any association that wants to play in the game. But a lot of people aren’t playing in the game at all. A lot of them are sitting on the sidelines.
The Opportunity Cost of Not Embracing AI
Jeff Cobb: [00:16:42] What do you see as the cost? I’m looking at it from the standpoint of the learning business and organizations that provide educational learning, what AI is doing there in terms of providing adaptive learning experiences, personalization. If you’re not getting on board with that, somebody is going to outpace you in your market. What about more generally for associations? What’s the opportunity cost if they don’t do what you’re talking about?
Amith Nagarajan: [00:17:06] You become obsolete pretty quickly. Many associations have the most reputable brand, the best content, and the strongest community in their vertical. Let’s say your vertical is a particular profession, maybe bounded by a particular geography, where you are still the number one game in town for, let’s say, a given state and a given field, or maybe nationally, maybe internationally. And so, in your niche, whatever that is, your brand and your content are very strong. Yet you decide intentionally—even though it may not be because of any malicious intent, but it is intentional—that you’re going to make it really hard for people to get your content, to find what they need. To gain actual value from your organization is going to be tough because you’re going to make your Web site a Byzantine structure. You’re going to make it so that your Web content for learning isn’t particularly up to date. I say it in this very directed manner that it’s a choice because all things in life are a choice in terms of where you invest your energy and your dollars.
Amith Nagarajan: [00:18:10] Part of the problem is associations try to do too many things. They try to solve 15 problems. Of course, that goes back to governance, where, in these strategy retreats, everyone wants to have their little pet project on the table, and they don’t want it to die. But you have to go pick two or three things and do them well. So there are some tough choices that have to be made. My view is associations create a lot of friction, and friction is an enemy—number one enemy perhaps in some contexts—to helping your audience gain value from you. If I say, hey, Jeff, I know you guys have unbelievable expertise in education and learning. In fact, for us, that’s super relevant because we run a certification program on AI for associations. I’d love to learn from you. You make it quite easy because your content is pretty easy to access. It’s pretty well organized. You have podcasts. You have written-format material. So I can find you easily. I can find your content easily.
Amith Nagarajan: [00:19:08] Even if you had great content, but you made it really, really hard for me to get to it, I’m not going to come to you as an authority on the subject, even if I know you and respect you and think you’re the greatest source of information on it. Same thing with an association. There’s a certain tolerance people have for friction. Think about it this way—why did Google completely freak out? They lost it when ChatGPT started rising because they realized that is a displacement of their business if they don’t catch up and do something. They put some serious horsepower into it now, and I think they’re doing some good stuff. But, when you see a momentous shift like that, you have to pay attention to it. Associations have this issue that their stuff is hard to get access to. It’s hard to find things. Almost every association CEO I talk to says, “Yes, the number one complaint I get from my board or from close-in volunteers is that it’s hard to navigate our Web site.” They’re like, “Oh, let’s go spend half a million dollars on a new Web site with some vendor.” And then six months after that Web site goes live, they’re like, “Yes, it’s like glossier and shinier and smells better,” but it still basically sucks in terms of being able to access the content they need.
Amith Nagarajan: [00:20:12] You’re not going to solve it with a new version of a horse and cart. You need different approaches to solve the problem. Because the problem isn’t that anyone has anything but the best of intentions; it’s that it’s a really complicated problem. You have a lot of content, you have a lot of different people to serve, so you need AI to personalize. You need AI to curate. You need AI to make things conversational. And that’s the opportunity in front of you. Five years ago, it would have cost you a fortune to do any of this, and it would have been pretty low quality. Now it’s extraordinary quality, and it’s very accessible even to very small associations. But you have to start by understanding what it can do. You can buy things and start slapping them on your Web site, but it’s much better if you spend some time learning first.
Hopes, Fears, and Blind Spots When It Comes to AI
Jeff Cobb: [00:21:00] You run a lot of programs to help people in associations with AI. What have you learned from that around what hangs people up around AI? What are their hopes, fears, blind spots?
Amith Nagarajan: [00:21:13] With Sidecar, we’ve got an audience of about 30,000 people that consume our content on a regular basis, whether it’s our three-times-a-week newsletter, which is totally free. We also run a weekly podcast called the Sidecar Sync. A lot of listeners tune in there regularly to learn and get updates. We’ve got a free book called Ascend, which is an AI book—full-length business book about AI for associations—totally free to download, and a ton of other stuff. So we put it out there. We’ve got about 30,000 people who regularly consume this stuff, and it’s growing quite rapidly. That’s almost double in size from this time last year. Clearly, there’s a lot of demand there.
Amith Nagarajan: [00:21:47] What I’ve learned from it on my end personally is that people are hungry for this, but that it’s unbelievable how fast people become experts. We have people that we’ve brought into our ecosystem that started off going, “I don’t know which way is up when it comes to AI,” and now these are people that are writing blogs. They’re leading sessions. They’re connecting with others. Part of what we do is help folks do that, and it’s amazing. It’s so incredibly rewarding to help people in their journey. They’re doing things in their organization, punching way above their weight class, and some of these are from very small associations. There is a perception that we can’t do it because of X, whether it’s “Oh, we’re not that technical” or “We don’t have a big budget” or “We’re a very small association.” There are all these obstacles. Those are all psychological. The reality is that the tools are accessible, fairly easy to use if you’re willing to put the time in. You don’t need to be a technician to be fairly good at this stuff fairly quickly, which is the exciting part.
Amith Nagarajan: [00:22:43] One thing I would share with your listeners, Jeff, is that a big impediment is fear of the organization not wanting the staff to do stuff with AI. There are still a lot of associations who have outright banned AI. There are still people who’ve done that. Good luck with that one because your people are using AI, FYI. They’re just doing it with personal accounts, most likely free personal accounts, which have zero security, and your content is being fed to those free accounts. If you are banning AI, you’re essentially saying, “Hey, would you like to use DeepSeek and send all of your association’s content across the Pacific Ocean? Cool. Go for it.” The reality is that, if you don’t set policy, people will just do whatever they want to. There are some people who will do nothing. Don’t get me wrong—some people who are the rule followers will say, “Well, my organization said, ‘Thou shalt not do AI,’ so I won’t.” But there’s a large number of people who will. And it’s stressful too. It’s super anxiety-inducing if you want to use a tool that you know could potentially help you.
Amith Nagarajan: [00:23:45] Then there are the people out there who don’t want to change anything, but they don’t necessarily like drudgery either. They might be afraid of losing their job. There are all those kinds of issues where people are thinking, “Well, if you’re so good with AI, then I don’t need to do my manual job anymore, which is repetitive.” It might be white-collar labor, but it’s very repetitive, and there are a lot of jobs like that. It’s your job as a leader to envision that future, to find the roles where people’s experience, knowledge, and unique contributions to your culture can add value. Most associations aren’t looking to cut tons of jobs or any jobs, but people assume that that’s what they’re going to do. That’s another obstacle. There’s a big leadership gap here, both in terms of basics, like figuring out what your AI policy is going to be. We’ve got templates for that, by the way, on our Web site that are free—you can download them on Sidecar—that will help you figure out what your policy should be.
Amith Nagarajan: [00:24:33] It’s basic stuff, by the way. It doesn’t require a lawyer. It doesn’t require tons of work. But it’s, hey, is ChatGPT allowed? Is Claude allowed? Are you going to allow meeting notetakers? In this Zoom call, you have a meeting notetaker that’s helping you record the notes for the pod, and it’s going to help you put together all that kind of stuff, which is great. But do you want to allow that in general when you have a Zoom call or Teams call, especially if the notetaker is one you’ve never heard of? Because a notetaker can show up from any company that anyone happens to use, and where’s that content going? There are all these things you need to think through. Again, it’s not a gargantuan task, but the absence of such policy is an impediment to adoption, and it’s also a major stress point for your team. Fix that. You can do that in the next 30 days easily.
Jeff Cobb: [00:25:15] Your comment about how quickly some people learned and became experts just by participating, I find totally true. You mentioned later making that 15 minutes a day. I’m sure you’re familiar with Ethan Mollick, who says, “Just roll up your sleeves, and do that 10 hours.” Once you do that, you start to realize this is doable, and you can get so much out of it. I was in a mastermind group we run just the other day, and somebody was lamenting the fact that their organization won’t pay the $20 for the ChatGPT subscription each month. If you look at how much that adds up to in terms of cost annually, it’s trivial compared to most other things they’re doing with software, and the potential returns are huge. But the organization hasn’t made the kind of decisions that you’re talking about and put the structure in place.
Jeff Cobb: [00:26:05] I did want to also dive in, in addition to Sidecar, to two of the applications that are part of the Blue Cypress family. And you’re welcome to talk about others if you’d like. These two jumped out at me, partly because I was at ASAE recently and noticed them there. But also one of them we’ve referenced in our work recently, and that’s Betty, which is the chatbot-type application.
Jeff Cobb: [00:26:27] We’re now at a point where I believe that traditional, formal, structured courses are going to decline significantly. There is certainly a role for them. They will not go away. But these unstructured learning activities, where the learner is able to engage how they need to engage at that point in time to solve the problem, learn what they need to learn, a chatbot is a great vehicle for that. We recommended it as one of the solutions in an engagement recently. We’re also very big on data these days. Traditionally, the folks in our audience (the learning business audience) have not done enough with data, either to understand their market or to make learning that adapts to their market. Betty is one of your applications; Skip is another, which I know does data analysis. Can you talk a little bit about what those two applications do? And, again, if there are others you feel it’s relevant to mention to our audience, feel free to do that.
Amith Nagarajan: [00:27:27] I’ll talk about three categories, and I’ll mention our offerings in this space, which we know better than we know anything else; I can speak to them the best. The three categories I think are critical for all associations are knowledge, insights, and personalization. We have a knowledge agent in Betty, we have a data analytics agent in Skip, and we have a personalization engine in a product called rasa.io. I’ll talk about each of those briefly. Associations have, for hundreds of years in some cases, been the center of knowledge for their community. A knowledge agent—imagine a being. In this case, it’s a computer, but it’s an entity that has the deepest and best, most contextualized knowledge in your field of anyone on the planet and is available all day, every day, and can be available as an assistant, as a tutor, as a writing partner. That’s exactly what Betty is. Betty has been around since late 2022. We’ve been working, as I mentioned, with AI for a lot longer than that, but we knew this moment in time would be coming because we’d been working with both predictive AI and generative AI for years prior to the ChatGPT moment.
Amith Nagarajan: [00:28:44] But when ChatGPT launched, we said, “Okay, we need to pull the trigger on doing this knowledge agent/knowledge assistant concept now.” So we did. Betty has over 100 enterprise associations using the product. It’s growing extremely fast. Basically, what happens is this—an association says, “We want to train Betty,” and so Betty will learn your content. Your content could be all of your past journal articles. Your Web site might be all of your online courses. It could be all of that. It could be lots of different things. Betty learns that, and Betty stays up to date on all that material as it continues to change. Betty provides grounded answers to your community. “Grounded” is a very important term. You’ve probably heard the term “hallucination” in the world of AI, and that’s when the AI essentially makes something up—that’s both a bug and a feature in some respects. Not in this context. It’s not a feature to make something up when someone’s looking for answers. It can be a feature when you’re doing brainstorming and being creative about “What kind of business ideas could I come up with? How can I possibly do a better marketing campaign?” That’s where the AI’s creativity is helpful.
Amith Nagarajan: [00:29:50] Context, knowledge retrieval, and answering particularly mission-critical questions—it’s completely unacceptable, and, unfortunately, there are many AI solutions out there that still allow for this concept of hallucinations to seep through. At Betty, our number one priority has been accuracy of the knowledge because associations—especially in medical or engineering—cannot tolerate inaccuracies. Every answer that Betty provides, whether it’s through chat or through agentic AI, which is when you connect Betty to other systems—this is not just the chatbot; that’s one way to surface Betty, but you can connect Betty to the rest of your critical business infrastructure—Betty always cites her sources and is always able to refer back to the knowledge retrieved and the thinking process that went into formulating the response. It’s not some black box where you’re hoping and praying that it’s going to answer your members’ questions correctly. It’s grounded on your content, only your content. It’s not using the Internet. It’s not using random Web sites. It’s a very powerful solution, and it’s purpose-built for this sector, where fault tolerance is critical in terms of availability, where the tolerance for any risk is very low, for good reason.
Amith Nagarajan: [00:30:57] Betty’s been an incredible hit. We expect to see Betty grow for years to come because the problems she solves are ultimately different than the problems you can solve through classical means. I mentioned earlier the difficulty of accessing the information people want. Well, Betty completely solves that. Rather than having to put in the right keyword into some search tool—even if it’s a great search tool, and you get 50 articles back, that’s not necessarily helpful—you go to Betty and say, “I’m working on this problem,” and Betty gives you an unbelievable answer. You’re going to come back to that. If you’re a member and you have that experience, you’re going to come back to that. Some people say, “Well, how can the association compete with ChatGPT?” The answer is two things. Number one, your private content, and number two is your brand. People will trust your brand, for good reason, if you are grounding your answers on the basis of your proprietary and uniquely high-quality knowledge base.
Amith Nagarajan: [00:31:55] You asked about data and insights, so I will pivot to that next. We also identified what we consider one of the grand challenges of association management: insights into really anything. It can be classical database questions like, “How many members did I lose in the last month?” or “What’s my renewal rate?” or reports from events, where it’s like, “How many attendees do I have that were an attendee at some prior year, but not this year?” All those kinds of things. Those are fairly straightforward to do traditionally. Part of the problem is that people have this explosion of systems. It’s not just an AMS, LMS, or CMS; it’s those things, but it’s also probably 5, 10, or 50 other systems that are used for different specialty things. That’s not going to go away. People have been trying to fight this since the beginning of computing. There have always been specialty systems because they do the job really well. So the question is then how do you solve for a widely distributed source of truth in terms of data? How do you make it so that people can ask any question—like with Betty—and have a rapid response with a well-thought-out, reasoned answer? It’s not just for technicians; it’s specifically for business users.
Amith Nagarajan: [00:33:07] Think of Skip essentially as like Claude or ChatGPT, where you have a conversational interface, and Skip is living, though, in a secure data platform. It’s an AI data platform that we have in the market called MemberJunction, which is a totally free open-source data platform we publish and maintain. What it allows you to do is to easily ingest data from all your different data sources. If you’re using Salesforce, and let’s say you also have an AMS, and you have a financial management system, and you have 10 other SaaS products that are out there, maybe some spreadsheets and other things to throw into the mix to round it out, you can ingest all of that data very easily into MemberJunction and keep it up to date. Skip lives on top of that and has secure access to all that data and is smart enough to figure it out. The problem for us humans is that that many systems are too complex for us to know, “Oh, well, this data lives over here, and this data lives over here, and how do I connect it and make sense of it?”
Amith Nagarajan: [00:34:04] Skip’s an incredibly powerful AI and has the ability to look across any complexity of schemas and then write queries, write code, and ultimately come back to you with what we call components—reports and dashboards—that, if you like them, you can save and rerun them however often you want, or you can ask for new ones each time you use it. Skip typically takes anywhere from two to five minutes to produce an answer, sometimes a little faster, sometimes slower, depending on how much code Skip has to write. Think of Skip essentially as a software development team coupled with an MBA-level person. You take a brilliant business strategist and analyst coupled with a data scientist and a great programmer, combine them all together in a team, and make them available 24/7 and able to respond to you within a handful of minutes. It’s pretty amazing. You can start to get insights. What I love about Skip is that we’re trying to reignite curiosity. We have all been conditioned to not expect much in terms of data because it’s hard; we have to go to a vendor, or we have to go to IT, and we’ve got to wait. By the time we get an answer—because some custom query or report has to be created—even if you have the dollars to spend, typically, you wait two weeks, four weeks, eight weeks. You don’t ask the question because you’re like, “Well, I need to know now. And, if I don’t know the answer now, I’m not going to wait in my own workflow for two weeks to get the answer.” As a result of that, many of the curiosities people have about their data to gain insights and to leverage insights are stamped out over time. That’s what we’re excited about with Skip.
Amith Nagarajan: [00:35:39] Finally, one of the most important things to get right is this idea of getting the right piece of information to the right member through the right channel at the right time. That’s this idea of personalization. This is the original problem we’ve been working on solving in the AI realm, going back to the early 2010s. This is before rasa was an official company. But, back at Aptify (the AMS business I used to own), we were sitting on top of these massive, large-scale databases of information, and our belief was, “Oh, well, we could make these Web sites better and make online communities better and make the education process better if we could personalize. So let’s use all the information in the AMS to do that.” That was the original idea in 2013, 2014, 2015.
Amith Nagarajan: [00:36:23] By 2016, when I sold the company, we had spun out the technology assets that became rasa and ended up focusing in one particular narrow category at first, which was e-mail newsletters. It’s almost like an afterthought, but people forget about the fact that the newsletter is one of the most valuable things many associations provide. The lowly newsletter, which goes out maybe to 20,000, 50,000, 100,000 people a week or even daily, is your most frequent point of interaction oftentimes. How often do people come to your Web site? Even a wonderful Web site? Maybe once a week. Probably once a month would be extremely high. Most of the time people come to your Web site when they want to register for your annual conference or renew their membership or something like that. Maybe they do a lookup to try to find an article and go away after they’re frustrated. That’s happening three, four, five, six times a year. But your newsletter hits their inbox every day, perhaps, and it’s an opportunity to deliver value or an opportunity to frustrate.
Amith Nagarajan: [00:37:21] The problem is, if I have 30,000 people at Sidecar, for example, and, yes, they’re all association people, mostly association people, and mostly they’re interested in AI because they wouldn’t be part of Sidecar if they’re not, how do I then make it relevant to them? How technical are they? How far along in the journey are they? What’s their level of proficiency in different tools? What’s their area of interest? Are they a finance person, a marketing person, or a CEO? Are they entry-level? There are all these dimensions of what makes you and me individuals. This goes way beyond the idea of segmentation. This is infinite segments, essentially, and it’s dynamic. The idea was to say, “Hey, can we really target the individual?” That’s what we’ve proven with rasa for a decade now. If you send e-mails that are deeply personalized, the funny thing that happens is people ask for more of them. Associations often say, “Oh, we can only do a weekly newsletter because we send too many e-mails.” That’s true if your e-mails annoy people. If people are getting something that isn’t value-additive to them, they say, “No more.” But, if you give people something that’s improving their life, they’re going to want more of it.
Amith Nagarajan: [00:38:30] Of course, I want more free ice cream. Of course, I want more free whatever. It’s great. It’s wonderful. It’s utility. If it’s negative utility, they say, “No more.” If it’s positive, it’s very simple. Sure enough, with rasa-based e-mail newsletters—and this is over 10 years with hundreds and hundreds of association customers and non-associations as well—we’ve shown a couple things. Number one is your unsubscribe rate is cut down by an order of magnitude, literally a 10x decrease in unsubscribe rate because we’re sending relevant content every single touch. Number two, your open rates go up by about 2x. And, number three, your click rates go up between 5x and 10x. That’s an enormous improvement in engagement. Many people have quoted that they value the newsletter as the number one benefit of membership after they implement a personalized newsletter. It’s a tremendous, easy thing to do. And that’s why I like to talk about it because it’s proven. Hundreds of associations use it. Many of them make a lot of money because of it.
Amith Nagarajan: [00:39:25] If you sell ads in your newsletter and all of a sudden you have five times as many clicks and two or three times as many opens and a bigger audience because there are fewer unsubscribes, it’s really good for your advertising metrics, and you’re going to make more money. But, ultimately, you serve the member better. Most recently—last thing about rasa—we broadened the technology. We’ve had a 98-percent retention rate year over year for a decade now. The number one complaint people have had is “Well, I wish you could send more e-mails, not just the newsletter.” And that’s the thing we’ve been working on for a couple of years. The AI has gotten powerful enough where it can personalize anything. It’s not just newsletters anymore, but now you can send any e-mail through rasa. It’s a product called rasa Campaigns. There are two different things: rasa Newsletters and rasa Campaigns. Rasa Campaigns is a direct replacement for old-school e-mail blast software that doesn’t personalize or doesn’t personalize well. It’s cost-comparable to those kinds of packages, and it allows you to do any kind of campaign, sequence-based campaigns. But the key to it is personalization, deep personalization.
Amith Nagarajan: [00:40:29] A couple quick examples. In your world of learning, rather than promoting a course and saying, “Hey, Jeff, you should attend my upcoming course,” I can specifically tell you why. I can give you a subject line that’s handwritten for you by AI. I can give you three bullets about lessons in the course that are going to be super, super relevant to you. Same thing for annual meetings, same thing for volunteer opportunities, and on and on and on. I actually get excited. I love all these products. They’re all my children, essentially, so it’s hard to say which is a favorite. I don’t have a favorite. But the idea of personalization doesn’t quite have as much sizzle, but the value creation is absolutely enormous.
How AI Impacts What Humans Do and Need
Jeff Cobb: [00:41:14] Many people are worried that as AI becomes more and more powerful, the human role starts to shrink, and we’re going to have issues with that. But, of course, associations and learning businesses are supposed to be fostering human growth. That’s what they’re about. Where do you see the uniquely human opportunities that AI is never going to replace? Certainly I’ve got learning in my mind, but, more broadly, how should organizations lean into the human opportunity that is still there with AI?
Amith Nagarajan: [00:41:46] You lead the thought process with where will the customer naturally go? Do customers still go to Blockbuster, or do they stream on Netflix? It’s the latter because it’s more convenient, and the value is better. There’s more choice. It’s faster, easier, cheaper, better. No matter how great your video rental store is, doesn’t matter. The flip side of that is what would people continue to do with people? Connecting in person, synchronous learning. Even if you can automate almost all of asynchronous with AI tutors and AI-generated content, synchronous experiences with experts are deeply valuable. You mentioned a mastermind that you guys run; we do one as well. That kind of cohort-based learning is incredibly powerful, deeply human connections. There’s more opportunity to do that. At Sidecar, for example, with our audience of association AI professionals and people who are going through the certification process, we get a lot of customer service inquiries—very similar to member service, same idea. We’re working on automating all of that. We love our team, and we’re not planning on letting anyone go. In fact, we’re going to add more team members.
Amith Nagarajan: [00:42:44] But what these people are going to be doing, instead of answering rote e-mails, as we get more of this agentic AI up and running, is to make proactive outreach part of what they do. They’re going to call customers and say, “How can I help you improve engagement amongst your team? How can I help you get everyone in your association staff and your close-in volunteers up to speed on AI? How can we increase that level of service?” In the world of education—again, this is your domain far more than mine—we’ve known for a long, long time, for thousands of years, that one-to-one tutoring has been the gold standard for the ultimate way of achieving outcomes in terms of learning, knowledge transfer, skill accretion, and all that. Yet that’s not scalable. Not all of us have an expert we can go sit with and hang out with all day and learn from and have discussions with. But now we can. One of my favorite use cases of Betty, I briefly mentioned, is Betty as a tutor.
Amith Nagarajan: [00:43:39] In fact, if you come to the Sidecar Learning Hub, which is all our asynchronous content, we’ve got the 10, 11 hours of content you mentioned on AI stuff, and, in there, Betty is hanging out. If you have a question, you can talk to Betty, and Betty knows, not a ton about you, but a little bit about you, soon will know everything about you, and Betty does know where you are in the course and what you’ve studied so far and the lesson that you’re on. When you talk to Betty, Betty is also primed to not give you the answer but to take a more Socratic approach to say, “Well, Jeff, let me walk you through it. What do you know about it? How can I help you learn this?” and try to ultimately get you to the right outcome but to be much more of a tutor mindset. AIs don’t get tired, and AIs can scale to every human on earth. We can all have the most incredible tutor in any field that we want. We can’t do that with people. It takes us 25+ years to grow a human into a productive member of the workforce. We do that with AI in no time flat. We can scale it basically to any number.
Amith Nagarajan: [00:44:36] The other thing to remember is the AI we have right now is the worst AI we’re ever going to have. It’s just getting better. It’s on a six-month doubling curve, roughly, in terms of power relative to cost. Coming back to your question, I’m optimistic. If you are one of the people who’s like, “I’m going to learn this stuff. I’m going to figure out how to automate a lot of the drudgery. I’m going to focus on the connection. I’m going to focus on the relationships. I’m going to focus on new ideas.” We’re explorers, right? Built deeply into our DNA is this desire to fan out and figure out what’s around the corner and what’s on the next planet or what’s behind the next business model. And we have that opportunity now. We can go and explore at an unprecedented rate. And that means more discovery. The fixed mindset—that the earth is flat, and there’s nothing past the known universe—is very limited. Every time there’s been any era of scientific progress or philosophical discovery, we quite regularly find out that, in fact, the last generation’s thinking, as novel as it was at the time, was limited, which is, by definition, what makes the world and the universe a giant and amazing mystery. To me, I’m excited. Now, not everyone’s going to share that. People lead with fear, and a lot of times they’re saying, “I don’t know if I can learn this.” I believe that we can bring everyone along.
Recap and Wrap-Up
Celisa Steele: [00:45:54] We’re not done just yet—stick around for our recap.
Jeff Cobb: [00:45:57] Learn more about the work that Amith Nagarajan does at Blue Cypress and a link to Amith’s LinkedIn profile. He’d love to connect. He’s also open to fielding e-mails.
Celisa Steele: [00:46:17] If you’re serious about AI—and you should be—be sure to check out the Sidecar site because Sidecar has dozens of free AI resources tailored to associations.
Jeff Cobb: [00:46:32] If you found this episode valuable, we’d be grateful if you’d share it with a colleague. That helps more people find the show and supports the work we do.
Celisa Steele: [00:46:40] In this episode, Amith underscored that leaders can’t delegate understanding AI. Investing even 15 minutes a day in learning about AI and using it makes a huge difference, and learning businesses that put in the time will be positioned to thrive.
Jeff Cobb: [00:46:57] Amith also pointed out the opportunity cost of inaction. Amith’s focus is on associations, and he noted that associations have incredible assets—brand, content, community—but they risk irrelevance if they don’t apply AI to reduce friction and deliver value in new ways. So, dear listener, make the time to learn and apply a bit of AI today.
Celisa Steele: [00:47:20] Thanks again for listening—see you next time on the Leading Learning Podcast.
To make sure you catch all future episodes, please subscribe on Apple Podcasts, Spotify, or wherever you listen to podcasts. Subscribing also gives us some data on the impact of the podcast.
We’d be grateful if you’d rate us on Apple Podcasts or wherever you listen. Reviews and ratings help the podcast show up when people search for content on leading a learning business.
Finally, follow us and share the word about Leading Learning. You can find us on LinkedIn.
Why Associations Are (Still) Education’s Sleeping Giant
Leave a Reply