Supersonic Moments: The NFL’s Flight Problem & 100 Trillion Tokens | [Sidecar Sync Episode 114]

Summary:

In this special holiday edition of Sidecar Sync, Amith Nagarajan and Mallory Mejias unwrap two stories that feel wildly different—but point to the same big idea: technology breaks barriers we once assumed were permanent. First, they dive into the NFL’s quiet fascination with supersonic flight and how faster-than-sound travel could reshape sports, conferences, and global strategy. Then they shift gears to AI, unpacking a massive new study analyzing 100 trillion tokens of real-world usage to reveal how people actually use AI—from role-play and storytelling to agentic workflows and open-source models. Together, these stories challenge associations to rethink what’s “unsolvable,” adopt an abundance mindset, and ask the most important question of all: what’s your supersonic moment?

Timestamps:

👥Provide comprehensive AI education for your team

https://learn.sidecar.ai/teams

📅 Find out more digitalNow 2026:

https://digitalnow.sidecar.ai/digitalnow2026

🤖 Join the AI Mastermind:

https://sidecar.ai/association-ai-mas...

🎀 Use code AIPOD50 for $50 off your Association AI Professional (AAiP) certification

https://learn.sidecar.ai/

📕 Download ‘Ascend 3rd Edition: Unlocking the Power of AI for Associations’ for FREE

https://sidecar.ai/ai

🛠 AI Tools and Resources Mentioned in This Episode:

NFL & Supersonic Flight ➔ https://shorturl.at/hnedE

OpenRouter’s State of AI ➔ https://openrouter.ai/state-of-ai

OpenRouter ➔ https://openrouter.ai

DeepSeek ➔ https://www.deepseek.com

Qwen ➔ https://qwen.ai

Claude (Anthropic) ➔ https://www.anthropic.com

👍 Please Like & Subscribe!

https://www.linkedin.com/company/sidecar-global

https://twitter.com/sidecarglobal

https://www.youtube.com/@SidecarSync

Follow Sidecar on LinkedIn

⚙️ Other Resources from Sidecar: 

• Sidecar Blog
• Sidecar Community
• digitalNow Conference
• Upcoming Webinars and Events
• Association AI Mastermind Group

More about Your Hosts:

Amith Nagarajan is the Chairman of Blue Cypress 🔗 https://BlueCypress.io, a family of purpose-driven companies and proud practitioners of Conscious Capitalism. The Blue Cypress companies focus on helping associations, non-profits, and other purpose-driven organizations achieve long-term success. Amith is also an active early-stage investor in B2B SaaS companies. He’s had the good fortune of nearly three decades of success as an entrepreneur and enjoys helping others in their journey.

📣 Follow Amith on LinkedIn:
https://linkedin.com/amithnagarajan

Mallory Mejias is passionate about creating opportunities for association professionals to learn, grow, and better serve their members using artificial intelligence. She enjoys blending creativity and innovation to produce fresh, meaningful content for the association space.

📣 Follow Mallory on Linkedin:
https://linkedin.com/mallorymejias

Read the Transcript

🤖 Please note this transcript was generated using (you guessed it) AI, so please excuse any errors 🤖

[00:00:00:14 - 00:00:09:17]
Amith
 Welcome to the Sidecar Sync Podcast, your home for all things innovation, artificial intelligence and associations.

[00:00:09:17 - 00:00:27:13]
Amith
My name is Amith Nagarajan.

[00:00:27:13 - 00:00:29:13]
Mallory
 And my name is Mallory Mejias.

[00:00:29:13 - 00:00:42:13]
Amith
 And we are your hosts. Mallory, we are recording this slightly ahead of Christmas Day when this episode will drop, so this special edition of the Sidecar Sync is not being recorded live on Christmas Day.

[00:00:43:20 - 00:00:48:00]
Amith
 But what are you excited about with Christmas right around the corner?

[00:00:48:00 - 00:01:00:10]
Mallory
 You know, so many things. I try to make Christmas as magical as possible as an adult. I mean, nothing really ever beats when you were a kid, Amith. But we have the tree up, the ornaments, the lights.

[00:01:01:12 - 00:01:13:04]
Mallory
 I'm excited to see my family. By the time you all are listening to this, I will be back home in Louisiana soaking up time with my nieces and nephews. So I'm excited for the holidays. What about you, Amith?

[00:01:13:04 - 00:01:44:20]
Amith
 Really similar. You know, my kids are late teens now. I've got one that's leaving the house next year off to college. The other one is midway through high school and just spending time with them hanging out. You know, it's a super fun time of the year. And the tradition and all the other elements are wonderful and amazing. But the quality time is really what boils down to you. And especially in the world of AI and the speed at which all the craziness is moving, it's good to have time to reflect back on what really matters. So it's a great time of the year.

[00:01:44:20 - 00:01:50:07]
Mallory
 And I thought maybe you were going to do like a ho ho ho welcome to the podcast. You don't have to, but that was just my idea.

[00:01:50:07 - 00:01:59:00]
Amith
 Yeah, I don't know if I could pull that off. But you know, people are kind of used to my boring intro. But you know, we'll see if I can work up to that by next year.

[00:01:59:00 - 00:02:00:04]
Mallory
 100 percent.

[00:02:01:05 - 00:03:50:04]
Mallory
 Well everybody, whether you're unwrapping presents, recovering from a big meal last night or just enjoying some of your well-earned time off, we wanted to bring you something a little bit different, a little bit fun today. Two stories about technology creating possibilities that seemed impossible until suddenly they weren't. So first, believe it or not, we're talking about the NFL's secret obsession with supersonic flight and what it could mean for the future of professional sports. And then of course, we're going to be talking about some AI. We're talking about a huge new study that analyzed 100 trillion tokens of AI usage to find out what people are actually doing with these tools. And answers might surprise you. Both stories have us thinking about what barriers associations have accepted as permanent that might just be waiting for the right tech to break through. Let's talk about supersonic flight though. First, it means traveling faster than the speed of sound that is roughly 767 miles per hour. We actually had supersonic flight from 1976 to 2003 with the Concorde, a luxury British French jet that flew New York to London in three and a half hours. It retired after a fatal crash, rising costs and regulatory challenges. Since 2003, no commercial supersonic flight has existed. But in January 2025, a company called Boom Supersonic broke the sound barrier with a test plane, the first civilian aircraft to do so in over 20 years. And NFL executives have been quietly monitoring this. So why does the NFL care? Well, the league has expanded internationally. London is a mainstay and 2025 saw debuts in Dublin, Berlin and Madrid. But permanent European teams face one barrier and that is travel time.

[00:03:51:11 - 00:04:39:19]
Mallory
 Supersonic flight changes the math. Boom's commercial jet aims to carry passengers by 2029, New York to London in under four hours. The company already has 130 pre-orders from major airlines. But Trump signed an executive order in June directing the FAA to repeal the 50-year ban on supersonic flight over US land. And there's some historical precedent here. As a fun fact, the Dodgers and Giants only moved from New York to California after transcontinental flights became available. Rail times made West Coast teams just impossible. So this is a story about a technological barrier that's held back major strategic goals for years. While we didn't explicitly mention artificial intelligence, Amith, do you think it's playing a role in the era of supersonic flight?

[00:04:39:19 - 00:05:29:16]
Amith
 Well, Boom specifically has been at it for a number of years. I actually have been following the company, I mean, for probably 10 years, eight years, and been on their mailing list just because I find this fascinating. It just sounds really cool. And I hope one day and during my lifetime, certainly, that there will be an opportunity to fly supersonic and as a civilian. And I think that it's just an amazing aspect of the world we live in. Certainly, I have to imagine that AI has played some role in some aspects of the development of the aircraft. But I don't know firsthand. I'd be really interested to dig into that and see how this company has been using AI during the development of their plane and certainly now. I would imagine the original designs probably were, you know, considerably pre-AI, certainly generative AI, but other types of predictive AI might have been used. But I would imagine they're heavily using it now.

[00:05:30:18 - 00:05:35:00]
Amith
 So four years, you mentioned 2029, that's right around the corner.

[00:05:36:04 - 00:06:26:07]
Amith
 In aviation timescales, actually, it's minutes because it takes a long time to get through all the necessary processes, in my opinion, appropriate processes to ensure the safety of these new technologies coming into use. So I find it really exciting. I think that this podcast, even though it's about the intersection of, as we always say, communications and AI, really what it's about is the intersections of associations and transformative change. So it happens to be that AI is the most significant transformative force in our world at the moment, it probably will be for many years to come, but it's not the only transformative force. And certainly if supersonic flight becomes a reality in the near future, it could reshape a lot of things, not just the NFL's decisions on where to place games, but many aspects of what many people do in their lives and certainly professionally.

[00:06:27:12 - 00:06:46:00]
Mallory
 Well, you said you would like to ride supersonic flights. I mean, I'm iffy on it. Maybe if you did it a few times, I would say, "Well, I'll give it a shot." You know, me safely traveled in under four hours, New York to London. But I don't know, something about supersonic flight, it doesn't seem to me, kind of like space travel. I'm like, "Let someone else do that for a bit in my opinion."

[00:06:46:00 - 00:07:50:06]
Amith
 Yeah, I would do that. I'd probably do the same thing. I'm definitely not John Glenn or Chuck Yeager or anything. I would let other people kind of be the test pilots. I think once there was a bit of track record, I'd love to jump on one of these planes. I think also, you know, there's other technologies that are on the horizon. There's a company called Auto Aerospace that has a plane coming out in the next few years called the Phantom 3500, which is not supersonic. It's a business jet. It's subsonic, but it's essentially right below the speed of sound. It's like 720 miles an hour or thereabouts as it's cruising speed. And there's some reasons for that that have to do with efficiency, where once you go past the speed of sound, you're talking about a pretty significant profile in terms of energy consumption. Obviously, there's the issue of the noise. But ultimately, that's a big jump over the average airliner is flying around 520, 550 miles an hour. So if there's planes on the horizon that are even 100, 200 miles an hour faster, that can add up. And if you spend a lot of time traveling, you know, even cutting off one hour instead of cutting off three hours from a flight like New York to London or New York to LA would be quite nice.

[00:07:50:06 - 00:08:15:08]
Mallory
 Yep. It would get you all home a lot quicker for the holidays. It's something to pay attention to. Amith, we have the NFL paying attention to aerospace innovation, even though they're a sports league. As a fun question, are there any industries outside of the association world that you think leaders should be watching besides perhaps the ones we've already discussed on the pod? Like anything just a little bit different that you would be keeping an eye on as an association leader?

[00:08:15:08 - 00:08:46:23]
Amith
 That's a great question, Mallory. I think casting a wider net or having a broader lens when you look at the environment around you and being able to detect patterns that could affect your sector is a really smart thing to do. And it's actually very difficult to do. You wouldn't necessarily affect a technology as different as AI to have historically affected associations as deeply as it is, but it is. And so, you know, supersonic travel and the NFL might seem like it's several degrees apart, but in fact, they're very closely related.

[00:08:48:04 - 00:10:13:24]
Amith
 The general answer to your question is, I think there will be many transformative so-called exponentials or technologies that are rapidly accelerating, right? That we talk about from time to time, AI being the one we focus on the most, that association execs and company leaders should be paying attention to. The way I would probably frame this and the best resource I can probably think about guiding people towards is really a frame of thinking around having an abundance mindset versus a scarcity mindset. So it's actually a book by the same name called Abundance by Peter Diamandis that's now, I think about 15 or 17 years old, where the author talked about the essence of what's happening right now in our lives, which is this convergence of multiple accelerating curves in terms of essentially the ratio of the cost of something relative to its availability and the growth and the power and capabilities of the same resource. Computing was one curve that he talks about, but really what we're saying is when we go towards an abundance economy or an abundance system, you have so much of something that you can do more of it and it opens up the doors to doing different things, things that you wouldn't have expected before. Think for example about the inputs to doing anything that we do in our world. Everyone needs energy. We all need to produce electricity or other forms of energy to propel ourselves around and to do what we do, to eat and to live, but certainly to do business.

[00:10:15:00 - 00:12:19:05]
Amith
 If energy is something we can solve, and what I mean by solve is if we have essentially environmentally friendly abundant energy at either low cost or no cost, what would that solve? It would solve for a bunch of things downstream from that. Another common constraint that you think about classically is water or water rights. There's a lot of water in the world, but a lot of it's dirty and a lot of it's got a lot of salt in it. There's ways to solve for that, but it requires a lot of energy. If you keep going down that path, you end up seeing that a lot of things that seem unrelated ultimately actually end up being very interrelated. Your industry, your profession, your sector ultimately can get impacted. I think the general way to look at it is be curious and keep a wide aperture so that you understand the world around you. Certainly focus in on your domain as we talk about all the time in this podcast. Domain specific knowledge is a driver to value creation, but you can't be so myopic as to not realize that things around you are changing and some of your core assumptions therefore are requiring a revisit. Coming back to the supersonic travel thing, I'll say one last thing that's really probably going to hit home for associations. If supersonic travel was here tomorrow and it was cost effective, would that change your strategy? For example, with respect to your conferences, you might market your conferences in a very regional way. You might say, "You know what? We're on the West Coast of the United States or we're in Europe or we're in Australia. We're just going to market our conference to people in this one region." Education like asynchronous online education that's not really time zone bound and doesn't have too much localized cultural stickiness to it has already spanned borders quite fluidly in the world we already live in. What about in-person education? What about the opportunity to connect with colleagues? Well, if you can get from New York to LA in two hours instead of six, would that change the equation? Especially if it was super cost effective. I don't think that's going to be the case in 2029 or 2030, but it might be in 2050.

[00:12:20:10 - 00:12:34:14]
Amith
 I think there's a lot of interesting things to look at. My general perspective is just try to be broadly read, try to pay attention to things that seem outside of your lane. Be curious. I think you'll find that there are unexpected connections where you can connect the dots essentially.

[00:12:35:15 - 00:13:04:18]
Mallory
 It's a really simple framework you share to me from Peter Diamandis, because I often struggle with this. How do you look ahead? I think of you, Amitha, and I've told you this. You have a really solid sense of vision and looking forward, but it's hard to dilute that into something practical. How do you think that way, but really just daydreaming with an abundance? An abundance of intelligence or an abundance of creativity or of energy, what does that look like? You can apply it across the board. I feel like that helps you think in the right direction.

[00:13:05:18 - 00:13:06:03]
Amith
 Yeah, totally.

[00:13:07:16 - 00:13:56:14]
Amith
 For those of you that are on our YouTube channel, my background right now is the Blue Cypress flywheel. On the very top of the flywheel, we assert that our first components of the flywheel is to seek and destroy unsolvable association problems, not unsolvable. We specifically chose that particular word because what we're looking for is the need for a shift, a shift in technology, a shift in abundance availability in some resource to make it now possible to solve that which was previously unsolvable. Our mindset is we only want to play in that game. The products and services we launch and will launch will only be focused on cracking the code on things that were and still are by most people thought to be unsolvable.

[00:13:57:19 - 00:15:40:20]
Amith
 If you think about something where you say, "Well, couldn't every association on the planet have world-class customer service?" Well, think about that statement. We all like to say that. Your mission statement might even codify that, that you do world-class this and world-class that, but you're in association with 50 staff, 100 staff, maybe a lot less. You don't have the resources of the Ritz Carlton or the Four Seasons or one of these world-class customer service brands or Apple or Nike or whoever you hold in high regard with respect to customer experience, customer service. How do you do that? How do you pull that off? If there is a bar separating you based on classical resource constraints, for me able to deliver that, you're sitting still. You're not making any progress, not material progress. You might have a little bit of an improvement in your member services process here and there that makes you slightly more efficient or whatever, but it's not a step change. Now with AI, we have the ability to put in your hands an AI agent that provides the best in the world member service nearly instantly to any number of members and in fact to flip the script because most customer service professionals, if they are being honest with you, will tell you that a doubling or a tripling or a 10X increase in the volume in their inbox would not be a welcome sight. If you're running member services for an association and you go from 100 emails a day to 1,000 emails a day, it would crush you. It would destroy you. It would basically be the worst day ever. But in fact, if the reason these people were contacting you was because you're providing such an enormous amount of value that they can't get enough, it's a different conversation. It's not that they're writing you that many times because they're mad at you. They're writing you that many times because you're creating value for them each time they connect with you.

[00:15:42:05 - 00:15:45:00]
Amith
 That is an inconceivable world. Going back to unsolvable,

[00:15:46:04 - 00:16:44:09]
Amith
 up until basically now, that was very much an unsolvable idea because there was a classic resource constraint or choke point which is high-quality human labor that was necessary to do that and the constraint was both the people being available and then of course money. So now if you can get that for a very affordable cost or very close to zero essentially in the grand scheme of things, why wouldn't everyone have that kind of caliber of quality member service and the good news is you can do that, right? But my point is that that's the kind of mindset you have to have. You have to look for those problems. Otherwise, you're just going to get incremental gains. You're going to say, "Hey, how do we make that spreadsheet we currently have? We want to be able to create that spreadsheet 20% faster." That's awesome. I totally applaud those efforts. That helps you get the wheel spinning. But what I'm looking for personally are those unsolvable problems to basically rack my brain and work with colleagues and work with the industry to say, "What are some of your pain points that you do not think are possible to solve?" And let's go focus on those things.

[00:16:44:09 - 00:16:57:13]
Mallory
 That would be a fun game or idea to happen at your board retreat even to just think about the unsolvable problems in your association or in your profession and then seek and destroy as Amit said.

[00:16:59:16 - 00:17:03:05]
Mallory
 Amit, do they play football games on Christmas Day? You would know this.

[00:17:04:11 - 00:17:07:01]
Amith
 I'm pretty sure there's Christmas Day football unless I'm hallucinating something.

[00:17:07:01 - 00:17:16:21]
Mallory
 I know there's Thanksgiving. Well, all of you, if you're watching an NFL game today, I want you to think about unsolvable problems like supersonic flight and think about this podcast.

[00:17:18:08 - 00:18:14:18]
Mallory
 Moving to the state of AI, we are of course talking about a new research paper from A16Z, an open router that analyzed over 100 trillion tokens. You heard that right. 100 trillion tokens of real world AI usage across the past year. Quick explainer, a token is roughly a word or word piece and it's how AI models measure input and output. So 100 trillion tokens means billions of conversations and prompts. This is one of the largest empirical studies of how people actually use AI and we've distilled it into some key takeaways, which I'm excited to share with you all. The first, and this was a surprise I would say, role play dominates. Over 50% of open source model usage is for role play and creative storytelling, not productivity and not coding. This is surprising to me. People are using AI for interactive fiction, character conversations and entertainment and programming actually comes in second.

[00:18:16:02 - 00:18:27:19]
Mallory
 This next one will not be a surprise if you're an avid listener of this podcast. Open source is surging. Open source models now handle about 30% of all token volume up from essentially nothing a year ago.

[00:18:28:20 - 00:18:59:11]
Mallory
 These models like DeepSeek and Quinn have been major drivers. Chinese OSS went from 1.2% to nearly 30% of usage in some weeks. Next takeaway is the rise of agentic inference. Usage is shifting from single questions to multi-step reasoning workflows where the AI plans, uses tools and iterates. Reasoning models now represent half of all usage. Usage prompt length has quadrupled from about 1500 to over 6000 tokens.

[00:19:00:14 - 00:19:23:06]
Mallory
 This one I feel like I can relate to. We're going to call it the Cinderella glass slipper effect. When a new model launches, there is a narrow window where it can capture users whose specific workflow finally fits. Once that fit happens, those users become deeply loyal and stick around through everything. Early adopters who find the right fit have dramatically higher retention than later cohorts.

[00:19:24:14 - 00:19:54:12]
Mallory
 The next insight, again this will not be a huge surprise if you listen to the pod, programming is anthropics territory. Claude handles over 60% of programming related AI usage more than any other provider. This is the most concentrated category by provider. And then the last takeaway is that geography is shifting. North America is now less than half of global AI usage. Asia has grown from 13% to 31% with significant growth in Singapore, Germany, China and South Korea.

[00:19:55:15 - 00:20:02:24]
Mallory
 So, Amith, I'm curious on your take for the most surprising insight and then the most unsurprising insight, if you have to pick one of each.

[00:20:02:24 - 00:21:27:06]
Amith
 Yeah, for me, the first bullet you covered on role play and creative storytelling, I would have guessed that that was meaningful, but over 50% of open source usage for that is tremendous. It's exciting too, because that sounds like people are enjoying themselves, which is pretty awesome. It's good to hear. So I love that. That's really cool. So what's totally unsurprising to me is the rise of so-called agentic inference, which just means systems that are using AI automatically behind the scenes. I'll give you an example. I kind of alluded to this earlier in the last topic, but we're just about to launch a new AI agent called Izzy. We've talked about it here and there very quietly with just a handful of folks, but it is a member services agent. It's kind of a companion agent that's good friends with our other agents like Betty and Skip and others. The idea behind it is to essentially take all the inbound asynchronous communications, which would be emails, text messages, Facebook Messenger, WhatsApp, any asynchronous channels you may have, and to automatically answer the questions. The questions could be inquiries about your upcoming annual event. It could be things about your certifications. It could be about in-domain knowledge, where Izzy can talk to Betty or other knowledge agents. The reason I point to this is that a typical inquiry to Izzy, where you might have an email with, say, a couple sentences, will often utilize somewhere between 300,000 and 700,000 tokens for a single inquiry.

[00:21:28:21 - 00:22:18:05]
Amith
 What's happening in this is that Izzy might actually be running anywhere from a dozen to 50 different prompts of varying lengths. Some are tens of thousands of tokens. What it's doing essentially is breaking apart the problem into lots of different pieces. If you were to just take an email and drop it into chat GPT and say, "Create an answer for me," you may or may not get something good, but that might take up 1,000 tokens or something. It's very, very light. What we're trying to do is say, "Listen, we want to break the problem down into many different sub-steps, and each of those steps, we want to validate them, check them against FACT bases, which are FAQs or knowledge bases, make sure that our knowledge is grounded, and then we want to go through an iterative process of ensuring the quality of the response is extremely high. We use a variety of different models, many of which are the models you're talking about that are open source.

[00:22:19:06 - 00:23:04:11]
Amith
 Think about the problem we're trying to solve. We're trying to solve for creating a magical beast that is capable of somehow being the best customer service member service agent on the planet. That is going to be not just doing the simplest possible thing and the fastest possible answer, but it's the best possible answer. To do that, you have to go really, really deep. There's a lot of people that are building similarly minded agents that are also very heavy consumers of tokens. Point is, is that if a single email with two or three sentences results in half a million tokens of consumption by a tool like IZZY, and there's a lot of other people doing similar things again, that's massive, massive growth. I think that agentic inference was going to be by far the lion's share of inference consumption in the coming years.

[00:23:06:02 - 00:23:26:23]
Mallory
 I was curious because I knew Enthropic leading the way in programming. I was like, "There's a couple of this open source of surging, but that makes sense to me. I am not quite as up to speed on IZZY." I'm actually very curious about this, but I know there's been some internal testing at Sidecar and maybe across the Blue Cypress family. What has that looked like? What have you learned? Has there been anything surprising in that process?

[00:23:26:23 - 00:23:49:20]
Amith
 There have been a lot of surprises in the sense of what do people actually look for? What are people conditioned to expect? When people email customer support, what are they conditioned to expect? Think about your own experience as you listen to me say this on the pod. You probably are used to sending a message to a company or to an association and hopefully getting a response.

[00:23:50:20 - 00:24:15:24]
Amith
 Hopefully within, say, a day would be considered pretty good, if within an hour or two would be considered quite outstanding. Usually, labor constraints prevent that, but a day, a day and a half, something like that. Then a lot of times there's round trips where the person has to clarify their question. You ask, "Well, I'm not quite sure how to do this thing." Then the rep on the other end writes you back and asks you one question and then you have to answer it. Then it goes back and forth and back and forth and back and forth.

[00:24:16:24 - 00:24:21:02]
Amith
 That can be very frustrating to the customer, to the member, because they haven't gotten their problem solved yet.

[00:24:22:11 - 00:25:00:12]
Amith
 What ends up happening is, let's say, for example, I send an email to your association. It's coming from my email address and I say, "Hey, what time does the meeting start?" It's kind of a ridiculous email for me to send on the one hand in classical thinking, because what am I talking about? What time does the meeting start? If I'm a member services rep on the other side of that email browser, rather than I'm looking at the email that's coming in, I'm Mallory and I'm going to do my very best to solve every problem. I'm also not time constrained. If I need to spend all day long figuring out how to answer this guy, me, I can do that.

[00:25:02:03 - 00:25:07:06]
Amith
 What's Mallory going to do? Mallory, what would you do? You'd probably research me a little bit, right, if you got an email that vague.

[00:25:07:06 - 00:25:18:18]
Mallory
 I'd look at you, see what events you were signed up for, see what events we have going on, and then I'd probably make my best guess and say, "Is this the one you're talking about?"

[00:25:18:18 - 00:25:43:06]
Amith
 Totally. Totally. You might also do a search in your Outlook or your Gmail saying, "Hey, let me look for prior emails from this guy who's really horrible at communicating, but let me see if I can find some other messages to get some context clues from past interactions." You walk into a restaurant where you go regularly and you say, "I'll just have the usual," or they just say, "Hey, you want your usual." It's like, what do they mean? It's like, well, they have context on you. They have memory of you.

[00:25:44:08 - 00:26:55:19]
Amith
 What if the agent did exactly that? That's precisely what Izzy does. Izzy has access to your email history, your messaging history, your CRM or your AMS, has access to your knowledge base, and has no constraints really in terms of how much time. In human time, it's hours of work or whatever, but of course in AI time, it's compressed down to typically between 90 and 180 seconds. Izzy's able to do all the stuff you're talking about to access your AMS to say, "Okay, well, which meetings has this guy I'm actually registered for? Let me get that back. Oh, and let me get the last five or 10 messages that I've had with him." Let me also just do a quick search on meetings and see if I can get the most relevant recent messages with him, not just the most recent three or four, but the most relevant messages. Get all that, pull it all together, put together a little bit of research on this guy I'm eating, and say, "Okay. I think what he's probably talking about is the upcoming symposium on blah, blah, blah." That's probably what he's talking about because that's coming up around the corner. He is registered for it. Let me go ahead and write him back and say, "Thanks so much for reaching out. I believe you're referring to this particular meeting which he registered for. That meeting is located at this location. By the way, I'm going to give you a little links that you can click on it and go to Google Maps or Apple Maps.

[00:26:57:05 - 00:27:18:03]
Amith
 The meeting starts at 9 a.m. I'd recommend that you arrive at 8.30 to enjoy a continental breakfast and to connect with your peers, blah, blah, blah." A really nice message that's contextually relevant and some helpful things that goes a little bit beyond what was asked, but also is taking into account this research that's been done. How often do you actually get a response like that from a company that you message?

[00:27:19:07 - 00:27:19:21]
Amith
 Fairly rare, right?

[00:27:19:21 - 00:27:20:06]
Mallory
 Yeah.

[00:27:20:06 - 00:27:39:13]
Amith
 You're thinking. It's like, can you think of a time you've gotten that kind of customer service response? The only time I can say I've gotten anything even close to that is maybe at a really fine hotel that has concierge. If I stayed there for several days and I've asked them for advice on which restaurants to go to and then I go ask them for another recommendation, maybe they remember me.

[00:27:40:14 - 00:28:48:24]
Amith
 But really, even that is not so consistent. That's the kind of thing that results in a shift in mindset in terms of what's possible, which of course leads to, "Wait a second. That was really good. That was an awesome experience in 90 seconds or 10 minutes even. I got this amazing response. Wow. I'm going to be conditioned to want to go back to that resource more." Now instead of emailing you only when I really, really have to, maybe I start emailing you more regularly. I realize, "Wait a second. Mallory is able to answer a whole bunch of questions about knowledge in my domain too. Maybe I can go to Mallory for that." Mallory's like, "The librarian who goes and researches everything for me and gives me this great set of resources back that meet exactly what I'm looking for." The list goes on and on. If you have this kind of service mentality, you can flip the script on going from something that you actually try to push away. You will not ever hear a customer service professional say they're trying to minimize the number of tickets they log, but they're talking about being efficient with resolving them. They generally are not talking about how to radically increase the volume of tickets or messages they're getting. But here, actually, I think that could be a very much desired effect for both sides.

[00:28:48:24 - 00:28:54:11]
Mallory
 Wow. I'm fascinated. Is Izzy up and running at Sidecar right now?

[00:28:54:11 - 00:30:03:15]
Amith
 Yep, sure is. Izzy is handling customer support at Sidecar, not exclusively. So we very much believe in the idea of human in the loop, which is a fancy AI speak way of saying someone reviews Izzy's work. So there's a separate AI that's part of the Izzy system, but a different AI that checks Izzy's responses and assigns what we call a confidence score. How likely is it that the answer is spot on, meets the needs, and if it's over a certain very high bar, then it automatically goes out? If not, then a Sidecar team member reviews the message, can edit it, and then send it out with changes. And every time a team member makes a tweak to the message, either in some cases changing significant parts or maybe just adjusting tone or language, that goes back into a training loop which makes Izzy smarter over time. Because the more you adjust it, Izzy knows, oh, this is the response they created, and this is the response that actually was sent out. That's actually a perfect way of doing reinforcement learning with a system like this where it can become smarter and smarter over time. So that's what Izzy does. Izzy is something we're super excited about. It's massive tangible value to the members of the association, and of course internally as well.

[00:30:04:20 - 00:30:13:19]
Mallory
 Wow. What a great Christmas gift to meet Izzy. I mean, I've heard, I've heard rumblings of Izzy for, when did it come about? Maybe like a year and a half now?

[00:30:13:19 - 00:31:01:00]
Amith
 Well, a lot of the projects we take on, you know, this idea of trying to find or seek and then ultimately destroy unsolvable problems, they're things we expect to take a long time to solve them, right? And sometimes we release products that we think have utility but are early versions, we tell our clients that. And then, you know, in some cases we'll just develop something for 18, 24 months before we release anything. And that's the case with Izzy is we've been working on it for a year and a half, and we've kept iterating and iterating. We've made our architecture better, but we've also been waiting for the models to become smart enough where we feel that the fairly high bar that associations have and should have in terms of their quality standards, we're now there. And so that's an exciting time. And yeah, we'll be launching Izzy into what we're calling an early adopter preview. By the time you listen to this, we're starting it in January. So it's right around the corner if you listen to this over the holidays.

[00:31:02:11 - 00:31:07:15]
Amith
 And yeah, we're working actively with a handful of early adopters right now to get deployments done in Q1.

[00:31:07:15 - 00:31:20:19]
Mallory
 This is huge. I think when I started at the Blue Cypress family of companies in 2022, maybe you could have imagined it in me. But if you had told me what we're talking about now, I would have said, "Heck no. No way. Could there be such a thing?" That's insane.

[00:31:21:23 - 00:31:40:02]
Mallory
 Moving to the glass slipper effect of me, because this is one that resonates for me for sure. The fact that there's a narrow window when new technology launches to capture users who've been waiting for that exact capability, do you feel like there's anything to learn from this for associations about the first mover advantage that might be underestimated?

[00:31:40:02 - 00:33:14:11]
Amith
 I think to some extent, the people who are model developers, I hold them both in high regard and I'm empathetic towards them because their ability to produce a world-changing technology and have it be the hottest thing that lasts literally days, that's a tough spot to be. And it's a very, very difficult, challenging race. But at the same time, obviously, everyone else in the world benefits from that. And I'm quite enthusiastic about it. I think the context you're framing around this is like, what can associations learn from this? And I think the world the associations live in, they're used to fairly slow pace of change. They're the center of the universe and their sector in a lot of ways. And so one might think that associations have a good bit of time to figure this out. I guess the main takeaway I have is that the models are becoming so, so good that even if you might be the best in the world, if there's something else that's out there that's almost as good and it's free or cheap and you are hard to deal with and you are expensive, both whether it's in dollars or in effort level to engage with you, if there's something that's lower friction and more available and nearly as good as you, that's a problem. And so the adoption rate of AI is what it is, not because it's a buzzword, not because it's cool, but because there's economic utility. It's that simple. People are flocking to this thing because it is saving them time, making them money, helping them in their lives, ultimately making everything they do easier and better. And that is not something that comes along all too often. So I guess the main point I'm trying to make is this.

[00:33:15:12 - 00:34:29:03]
Amith
 In the context of what you described, it's when a new model launches here, I think it's in the context of when an association is thinking about some shift happening in their industry and who's the first to help people in their sector figure it out. That's the association's golden opportunity today. It's not just about your internal adoption of AI, although that is important. What I think is the bigger issue, the bigger challenge and the biggest opportunity of all is being at the center of the conversation in your field to guide and to advise and to help your audience, help your members grow and adopt and transform with AI. Essentially what Sidecar is trying to do for you as association leaders, you in turn should pass along and do for your members. You should be their Sherpa in the world of AI. And that opportunity is not one that's going to come along again, probably ever or certainly not in our lifetime. So the opportunity is there today. And I think the first mover advantage really has to do with who these people in your sector choose to build a relationship with that will help them in that transformative period of change. If it's not you, then they might not come back. That's a possibility because the good enough solution, which you might say, hey, I have the best knowledge in this field. Well, but if you're not thinking this way, you might not be the most helpful to them.

[00:34:31:20 - 00:34:59:24]
Mallory
 Well, that's a good place to wrap this episode up with me. I feel like both of our topics today are ultimately about the same thing. That technology doesn't just improve incrementally. It creates windows of possibility that unlock what was previously unthinkable or unsolvable. The NFL couldn't have European teams without supersonic flight. And the AI landscape a year from now will look nothing like today based on how quickly things are shifting. For associations, the question for you is what is your supersonic moment? What strategic

[00:34:59:24 - 00:35:05:09]
 (Music Playing)

[00:35:16:02 - 00:35:33:01]
Amith
 Thanks for tuning into the Sidecar Sync podcast. If you want to dive deeper into anything mentioned in this episode, please check out the links in our show notes. And if you're looking for more in-depth AI education for you, your entire team, or your members, head to sidecar.ai.

[00:35:33:01 - 00:35:36:07]
 (Music Playing)