Beyond the Hype: Building a Strategic Framework for AI Adoption

Alex Jonathan Brown
00:04 - 00:53
It's 11 AM in the Bay Area, 2 PM in New York. And wherever you are, grab a cup of your favorite caffeinated beverage. It's time for Coffee Break. 
I'm Alex Jonathan Brown, senior content strategist at Iron Horse.
And today, we're doing a deep dive into gen AI, the ways we've been putting the tech to work here at Iron Horse, and how you can put together an adoption plan that works for your own company. To help talk us through it, I'm joined by Iron Horse CEO, Uzair Dada, a Coffee Break regular at this point, and a rookie, brand new, bringing him in out of the bullpen.
He's our SVP of digital experiences and product strategy, Samir Mena. Hey, guys.

Uzair Dada
00:53 - 00:55
Hello. Hello.

Alex Jonathan Brown
00:55 - 02:12
So before we get started, this is a topic we're all very excited about. I think there's a high probability we're gonna go longer than a half hour.
So if you're watching and you wanna drop at the half hour point, the on demand will be here for you. Do your other stuff. Come back to it later. 
But this is a topic that, as you will see, we spend a lot of time thinking about here at Iron Horse, and it's not something that we just started thinking about. We've kind of been having these conversations internally for a long time. We've been having them externally for a long time too.
But when you title an event, “Beyond The Hype”, I think a really good place to start is talking about the hype. So we're gonna try again, as I said, this is a new platform for us. Relatively sure all the tech's gonna work, but we have, look at us, we're so fancy.
We have slides. So Samir, I don't know if you wanna, if you wanna kick things off for us, but let's, let's start the conversation at how the conversation started, really, the hype surrounding Gen AI and all of the amazing, wonderful, incredible world changing things that we were gonna be able to do with it.

Samir Mehta
02:12 - 03:12
Hey. Thanks for the, thanks for the intro, Alex.
Excited to be here with everybody. Just a quick intro about myself, I head up our customer experience and Gen AI efforts here at Iron Horse.
We've been dabbling for more than a couple of years with Gen AI. A little anecdote, I think Uzair and I, you know, it was probably 2, a little over 2 years ago, actually.
We heard about this company called OpenAI, and they had some interesting technology. And we actually, flew up to Portland, Oregon where we have some of our engineering talent, and we just holed ourselves up for a couple of days to learn about OpenAI and actually build and experiment with their API.
So this was pre-ChatGPT, and there was a lot of hype. We got some interesting learnings, but kind of the conclusion after our kind of mini hackathon session was, you know,the hype behind that it's gonna be this magic button that's gonna make everything magically work, that wasn't there. We realized kind of, you know, in that hackathon, in that sort of deep dive session that the technology was still pretty early, pretty raw.
Fast forward, you know, a few months later and you saw ChatGPT come out and it, you know, if I recall correctly, it is the fastest consumer application to hit 100,000,000 active users. So if you think about Instagram, TikTok, and other apps that had exponential growth, like, that immediately kind of brought generative AI to the masses.
Now a lot of hype because people were like, “Hey, this is gonna make everything easier in my life.” And if we put the lens of marketers on, there was, you know, on one hand, this hype, there's a little bit of fear, like, “Oh my god, is AI gonna take my job?” But there was this notion around Gen AI is gonna revolutionize marketing.
It's gonna completely turn everything that we've done, especially B2B marketing on its head. And we're gonna automate 90% of what we do so we can go do other things. And, in fact, it's gonna be so easy to adopt that within year 1 of this technology, it's gonna, you know yes. Maybe the consumer side, people are using it casually, but at the enterprise level, it was gonna infiltrate and, you know, massive ROI for organizations across the world. Some of that kinda came to fruition around yeah. And if you found the right use case, the right workflow to automate, you definitely saw impact. But largely, and Uzair can talk about this, is that there's been a lot of experimentation, a lot of learning.
I think with any new technology, it's really easy to jump in, but I'm gonna date myself here. When I think about kinda when the Internet came around and kinda the world world by web was here, I distinctly remember actually, when I was interviewing for my first job at Tech Companies, one question I got asked was, “Well, tell me how a web page like, how does it work? How does a browser? What does that like, explain to me the mechanics of that in nontechnical terms.”
And I think when it's it's it's incumbent and we'll talk about this later about really understanding the technology because when you understand how the technology works, it really allows you to understand where you can apply it. Now granted, this tech, unlike any other tech I've seen in my lifetime, is changing literally by the minute.
It is growing by the minute. So while there have been some, you know, starts and stops with Gen AI and there's a lot of success stories, but there's also stories that, you know, people have been really frustrated and haven't seen the ROI.
The expectation is true that there's gonna be an a further increase by organizations to to to bring GenAI and its benefits to their individuals, to their teams, to their organizations because the belief is that it truly is kind of a foundational moment perhaps, you know, in in our lifetimes, a the the most important technology that, that that we need to figure out and get right.

Uzair Dada
06:40 - 09:44
Yeah. And I'd say, you know, talking about impact last yesterday, when Alex and I talked to Samir, he was 10 years older. And just the magic of AI, today, a Benjamin Button happened, and sort of he's 10 years younger. So that's pretty remarkable.
And the other thing I was gonna say that Alex didn't disclose is one of the 3 of us is the fake. And so we'll let you guys figure out over time through this webinar who's real. No, kidding. It's a fascinating area.
We we are, I'd I'd say, kind of like, you know, for us creative, curious people in the marketing space, this is just amazing. I when it sort of happened 2 plus years, the and all the pundits were sort of saying it's the demise of everything, and you don't need this and you don't need that, You went through for me, it was a 2 days of, like, oh, shit moment.
And and then it was a ridiculous supercharged moment, and it hasn't stopped. And the experimentation is amazing because the more you dig into it, it's not perfect by any sense of the imagination for a lot of the use cases, but the ability to think about things and do things that you could not think about and do yesterday is just, to me, transformative.
And I think that's what's cool. Almost everyone sort of initially thought about this as a time saving thing, which is a huge component of this, and workflow automation of this, which is a huge component of it.
But to me, the fascinating part is for us to be able to do stuff that we weren't doing. That's cool. That's what's kinda neat. Right? It's sort of to be able to, perhaps, do the stuff we have to do, but, perhaps, we don't like doing. But be able to do it now because it's not hard to do anymore. It doesn't take it's not a time stop.
Like, that's transformative. So for us, the cool part is we sort of dove into it, pre-ChatGPT, and have been sort of in this lab mode since then.
I think we'll continue to be because the lab is just adding more equipment and more people and doing more things as we sort of do more. It's you know, the good good lab is always about learning and questioning and iterating, and I think that's kind of where we are, and where we see our customers are, frankly.
We are helping a lot of our customers sort of think through these things and think about the different use cases. A lot of what we're gonna talk about today is within the marketing lens, but I think it's to me, the hype is not unfounded. It's real. And but it's got a lot of, you know, classic principles of change management and adoption that we all humans need to think through and organizations need to think through, very much surrounding it just like you have for any other change. And sometimes we don't think about it. Sometimes we make big swags.
Those are additional components that I think that we're gonna talk about today that are very real as you guys are all embarking on this journey. So excited for this topic.

Alex Jonathan Brown
09:44 - 10:53
I think one of the things that you mentioned, Uzair, is that we do have this, like, little lab set up. And I think that hearing that, probably, there are a lot of people who are like, well, I know that's never gonna happen at my company. We're just not set up to do that. The important thing to remember is that, like, experimentation around AI is happening at your company.
The question is, is that an intentional effort that you are making as a company, or is it people who are excited about tech who have another browser window open and are doing it on their own. And I think the thing that that I think we all agree on is that having a framework to guide that experimentation is A, more productive for the company in general, but B, like, necessary. Because if that experimentation is happening and if it's showing up in your work, you very much wanna know about it. And I think that's maybe a good transition to start talking about what some of these frameworks can look like.
 
Samir Mehta
10:53 - 15:26
Awesome. I think that's to me. To Alex's point, with how available and accessible the consumer tools, the ChatGPTs, the Geminis, obviously, Apple Intelligence on your phone now. You know, AI has been around us for quite some time. If you think about, you know, voice assistants like Alexa, Siri, like, they've been around, and now the technology has obviously gotten way better and even more integrated into kind of your life as a consumer. But that doesn't always translate well or, like, neatly into a business organization.
And part of that is if you've got your, you know, members of your team all sort of doing different things, that's fantastic. They're exploring a variety of tools.They're attending webinars. They're trying to educate themselves.
I applaud that initiative and I think those are the type of people, kind of the trailblazers, the kind of the framework we have is sort of, you know, if your team or people haven't been ignited and I think this is probably far less common now. This was kinda common maybe a year ago where people were skeptical, dubious, fearful.
I think that for the most part, people have sort of gotten over that and have generated or have that intellectual curiosity. But if they don't, I think that's kind of like an area where you have to ignite the individual, you know, and part of that is incumbent on leaders to find those enthusiasts within your organization and kinda bring them together.
You obviously wanna, like, think through bringing, you know, a diverse team together, and we'll talk about some steps that we took to do that. And get that curiosity, that energy going.
Step 2 in the framework is now that you sort of have that energy and that ignition, some of the basic kind of literacy in place, you really wanna encourage more structured experimentation. And the keyword is structured.
You really wanna think through, both at the individual level, where hypotheses or things that I can't automate or augment, and try those things out. And, Alex, you know, I think part of the myth there as well, you know, I've given Alex these great tools. He should be able to save 20 hours of his time. That's not necessarily the case because I think to learn a new platform, invest in a new technology, it's gonna take time to figure out what works, what doesn't work.
So you need to create the space and the room and the structure for experimentation. If you can tie those experiments or those pilots to, like, true business outcomes, I think when you start off, it's more about let's get let's certainly get warm. Let's kind of try things, identify quick wins, whether they they they or quick experiments, whether you succeed or fail. It's more about getting in the motion of that lab mentality.
And then as you get more structured and focused on key use cases or key outcomes that you wanna drive at and you've been able to prove them out through your experimentation process, that's when you sort of start to hit, like, some level of ROI because you can now either significantly optimize a particular process or integrate the workflow, or you might actually create or augment capabilities that you otherwise didn't know you could do or were always resource constrained to actually be able to do. But with generative AI, it opens up new avenues and new doors.
And then kind of the 4th layer in the framework here, kind of the Nirvana, the North Star is the entire organization has been truly transformed. So it's kind of what I call, like, everybody's thinking AI first, not AI second. They're thinking, not just simply to optimize and make things efficient. They're actually looking at AI in terms of how to disrupt how they have traditionally done things, whether back office, internal business processes, but also very importantly, new capabilities, new products, new things that, you couldn't have done prior. And I think that's not just a technology change, that's very much a change management change. 
 
Uzair Dada
15:26 - 18:05
Yeah. I think I think for, you know, as you embark on this journey, the framework is super important.
It's also sort of your aperture as an organization, and how you said that aperture as an organization. For us, from the get go, it was a, we need to be in it, and we need to be leading it, and we need to because our customers come to us as advisors to think about what to do.
They are the ones. They're looking to us that ensuring that we've broken the stuff enough times that we are confident that it works or what could work. So one of the things that I did earlier as I came on is a north star, kind of this philosophy of 10 to 10. What used to take 10 weeks could take 10 days.
What used to take 10 days could take 10 minutes. May never work, but it was a cool, interesting, crazy thought process.And say, can we apply this across the board? Right? And so for us, it was the individual experimentation going on, then there was sort of a labs team that was primarily engineering. And then we sort of said, okay, this is cool. Engineering has shown some really interesting showcases, and we'll talk about some of those examples. I was like, this is rad. This is super cool. So now let's build a cross functional team and expose people.
It was a huge litany of reactions from “This sucks, this is not good.” Alex Brown was in the early days, they're like, this is interesting, you know. There was a lot of fear from, like, why is it there? So there was a lack of acceptance. And the notion there was okay, guys. Let's be real. It's not that AI is replacing you, but it's probably the person, tomorrow, who knows AI really well in that same function that replaces you.
So there was a kind of an enlightenment moment of it's not it's just like when Internet came in or the mobile phones came in or when cloud infrastructure came in. It wasn't like, oh my god. It's gonna it sort of it opened up new opportunities. It's the classic fear to acceptance to sort of success.
And that was cool. And I think that's sort of, kind of, the journey we initially went through. And from that, what led to it was really Samir and the team, sort of, leading across functional team, where each functional team, sort of, lent a few resources to just think about this as an organizational opportunity. It was just a think tank. So the lab broadened. And that was fascinating. And I'm sure we'll talk more about it.
 
Alex Jonathan Brown
18:05 - 19:11
I think one of the things that's interesting to add to this conversation too is that this kind of progression from stage 1 to stage 4 that we've got laid out here. It is rather linear, but it is not something where you have to get to level 4 in order to find success. And we just threw a poll up, in the poll tab of your little window here, asking where you think your organization is in its AI adoption journey.
But I think it's worth saying again, stage 4 is very hard, if not mostly theoretical at this stage with where the tech is. But making any progression along this, like, you will find successes at each of those stages. And it may not be, like you said, Uzair, it may not be what took 10 weeks takes 10 minutes. It might be what used to take 10 hours takes 10 minutes. Like, you can find those small progressions, moving through there. So we'll give people a couple seconds to vote in in the poll.
 
Uzair Dada
19:11 - 19:53
And and and and as you said that different we may have different tasks, different functions, different different experiments at different stages of this journey. So it's not just at organizational level. There are different maturities to different things. Right? There it may be due to technology maturity. It may be due to process maturity. It may be based on the adoption curve of certain resources. I mean, there's so many different factors that go into it. And that's totally cool.
And not everyone will move at the same pace. But it's sort of a we're going on a collective journey though, which is kind of the cool part as an organization.
 
Alex Jonathan Brown
19:53 - 19:55
Definitely.
 
Samir Mehta
19:55 - 21:31
One of the things, last comment before maybe we look at the poll poll results here, is that foundation, like, thinking about, like, you know, how AI works and really investing in the time to have individuals understand that. Because I think one of the things and this is a little back to hype in reality is, like, oh, I expect AI to work like a traditional software application, you know, 2 plus 2 equals 4.
Like, it's very, as we say, deterministic. AI is largely founded on the notion of math and probability, and it's probabilistic, and hence people talk about hallucinations. And there's lots of work and research being done to obviously mitigate a lot of that. But because of that and kind of what capabilities AI can enable and and we're only scratching the surface today. Like, there's I know a ton of companies, a ton of things out there that you're looking at. But one thing that we realized in the last couple of years is creativity. And I'm not talking about creativity in the sense of, like, design and content. I'm talking about the creativity of reenvisioning what your process or workflow or how you might think about achieving an outcome. We're naturally looking at an existing workflow we have and we're like, oh, automate, automate, automate.
Whereas as the literacy and kind of the exposure and the learning experimentation grows, you'll find a different way or they will present different ways of doing things that you hadn't thought of before.
 
Uzair Dada
21:31 - 21:50
Yeah. I would say curiosity drives the creativity. I think it's the, wow. Could this do this? Can this happen? I mean, we do that a lot. Right? And we have no idea, but it's sort of learn by doing, which is simple.
 
Alex Jonathan Brown
21:50 - 23:40
And one of the you mentioned the hallucinations, Samir. One of the things that I think is always so interesting is when you find those to try and reverse engineer how it got there.
And, like, that thinking of what it was trying to do can so often be that sense of, like, oh, I don't like where it got, but that, like, line of reasoning, for lack of a better term, can often be so productive from a creativity standpoint of, oh, I I see the gist, but now I'm gonna use my human brain and, like, get to somewhere interesting coming out of that. 
So we'll also go ahead and close the polls. We have roughly just quick math, so I'll apologize. 55% of people said they were at the AI experimentation level, which is kinda stage 2. 27% at AI scaling, AI Ignition at 18%. And thankfully for the thing I said earlier, no one feels like they've reached that AI transformation level yet.
So maybe that's a, maybe that's an aspirational goal, and also maybe it's not, and that's totally okay too. So yeah, just a little look at how our Coffee Break audience, where they're at in this, this AI adoption journey.
So we kind of laid out the steps, but I think it's also important to talk about how this happens and, like, what the driving forces are, behind that. So let's see.Y'all, Samir did so much work on figuring out what all of this should look like. I'm so proud of these slides. Incredible work to visualize these, like, really, involved concepts. So, Samir, it's your work. I'm gonna let you talk about it for a bit.
 
Samir Mehta
23:40 - 28:15
This goes back to a little bit of change management, like philosophy, but transformation you know, technology is a tool, it's an enabler, but you have to think about transformation more holistically. And what we saw early on because let's let's admit it: There were a few large companies who were caught sleeping when OpenAI launched their technology, and there was a lot of chasing, a lot of, oh my god. I don't have an AI product. Let me slap AI on my website. And then there you know, at the boardroom level, there was, hey.
We as a company need to adopt AI. We need to build AI products. Like, go. And it was a top down initiative, which in some cases can work. I'd say more often than not, it usually falters somewhere along that process or that kind of cascading effect. So with this type of technology and the fact that it's just so pervasive and accessible, I really think the theory of change or the theory of transformation really starts with the individuals.
And from a leadership standpoint, it is about encouraging experimentation. I know there are people who have early on felt guilty about using AI. You know, they didn't wanna talk to their managers about using it. And that's a cultural thing. It's like, okay. You're using it. You've established some guardrails and compliance policies. But long as you're using it within those confines, it's okay to share both your successes and your failures.
So, like, empowering individuals to experiment and having them find that little bit of, you know or in some cases, like, oh my god. This is removing a lot of, really monotonous and rote work from the things I do day to day, which allows me to focus on these other projects that I'm really interested in.
So finding those early wins at the individual level, that then creates excitement, energy to then bring it into a more structured team-oriented environment. And if you don't kind of empower, educate individuals, if you kinda just enter it, the team team, go figure this out, it it can work, but oftentimes, they're gonna have to go back and each individual needs to level up and kinda understand what are we doing and do I feel a sense of ownership and agency around sort of how we, me, we, the team, are going to use AI and how we're going to, sort of unlock its value.
I think teams need more structure and that's something if, you know, if you have a task force or a tiger team can help, you know, work with teams, identify a couple of pilots that they want to experiment with with a clear hypothesis of the problem they're trying to solve, what the impact of AI will have, what's your baseline measurements, and not everything's gonna work. They might do 2 or 3 experiments, and one might work.
I think what we've seen is people tend to kinda bite off more than they can chew. They try to solve the very complex pilot, the complex use case, and then it just fizzles out. And your excitement, and your empowerment, all the stuff you did with individuals starts to kinda dissipate. But if you can go from 1 to 2 and teams start to see benefit, they're actually being more efficient, they're augmenting their capabilities, then that really pushes into the org and enterprise. So typically, maybe one department, one team, one organization that ignites the rest of the org, to sort of say, oh my god. Look at what the marketing team did, or look at what the customer support team did.
And that becomes infectious. And I think when that happens, you start to sort of free up time and you get excitement across the organization. So then it it it actually feeds back to individuals because now they're looking for maybe not the process efficiency, outcomes, which are the probably the easier ones, the automation ones, but they're now starting to think about 10x thinking or really being more innovative and rethinking kind of the core sort of foundations of their business or how they go about doing their business.
 
Uzair Dada
28:15 - 31:24
Yeah. I think the other thing I would add to it was to sort of this flywheel to work, to me, it was not just the individual experimentation, but it was showcasing it. Let's call it. We like shiny things as marketers. So we need to create some shiny examples to get people to jump on the bandwagon and drive excitement even with an individual. So and which is, you know, we're fortunate being who we are. The majority of our organization has a very curious DNA. We all like shiny things. We all like to experiment. We all love breaking stuff.
You know, I probably have the gold badge for, you know, QA and finding bugs and breaking shit than in stuff we do. So it just innately, our organization DNA was there. But I think showcasing something simple. Something that was like, wow. That just drives more curiosity. The more you show, it deepens into curiosity. Then it goes to teams. It broadens the curiosity. Right? And then as Samir said, infectious. And I think organizationally, when you see someone else succeed, you inherently want to understand. And that was it just by just learning, by doing, and slow rolling, and encouraging. And hopefully, Samir will talk about sort of also enabling.
Enabling meaning, giving them tools. Enabling means we did a lot of interesting learning labs. We did a lot of interesting stuff like that, applied stuff. So at our own organization, in different things we were doing, we're at different stages of scaling and transforming. And even to this date, it's true. There are some things that are much further. There are some things that we are productizing that we can go to market with. There are other things that are just internal.
So that evolution has just been awesome. But I'd say to me, that when I saw it and then when I touched it, you became the believer. But when Samir and the team sort of came back, a couple of individuals, and showed us some stuff and what they had done, we do a lot of webinar stuff. And so one kind of interesting idea was, like, hey. Can we have GPT custom GPT help us structure all the content for all the different things we do as a draft, just as an experimentation, as a thought exercise? And spent a lot of time thinking about it and not then you could go and say, hey. I'm doing a webinar on x, y, z topic. To this kind of an audience that's this long, give me everything. And it produced content for end stuff.
Now a lot of it was not that great. But the notion of being able to guide and get a decent response, you know, it's sort of like having your dedicated intern give you a starting point. It was very cool. That was to me the first ignite point of something interesting beyond just search, for me personally, in our journey a couple years ago.
 
Alex Jonathan Brown
31:24 - 32:07
I think the shortcut on the content side of things, whether it's a writer or kinda strategist in general, is always what's your least favorite part of this? Yeah. And then freeing those people up to say, you're welcome to go take some time and go try to solve that with AI.
So you can do less of the whether it's social posts or email subject lines or whatever you hate doing, do less of that. Like, that is that is a quick way to, like, get someone excited about this to just say, yeah, don't do this stuff you don't like. Let's figure out how to get good results and make your brain hurt less while you're trying to work on things.
But that's my little version from my little content side of things. I know we also wanna take a look at kind of our bigger, AI adoption and transformation journey. And so shout out to Jay Jamie on our team, a human being who helped bring Samir's slides to life. Let's kinda talk about how we've established our foundation and scale and move to scale and expand.
 
Samir Mehta
32:35 - 37:11
So, if you're many of you are kind of in that experimentation phase, and so you've probably done some of this or part of this, but we just wanted to share what we've done over the last, 18 to 24 months. By no means is it this an exhaustive kind of framework, but what it is is kind of a really simple three stage model, to kind of just center yourself and where you're at. We've got lessons learned in each of these three areas. I'll touch upon a couple of them.
I'm not gonna read this slide, but, you know, I think we've talked about establishing the foundation. And that foundation, as marketers, it's like know your audience, right? Like, who you know, you can survey your team, you can survey, or have conversations and understand sort of where people's AI literacy is, their comfort with the technology, how have they used it both personally and professionally. Like, because once you understand where your team's at, you can figure out the gap and get them to, like, some of that foundational literacy.
So early on, there wasn't a lot of stuff. Myself and a few others on our team, Ezra, Gavin, we actually did workshops internally, did a bunch of research to really help explain, Alex Brown actually did some awesome internal talks as well to talk and explain how this technology works at a nontechnical level. And then we created a tiger team or a task force to explore as a unit, but also develop some key artifacts and have discussions and debates, around guidance and compliance policies, and have a Slack community, internally to share interesting things that have come about.
Now, kinda lessons learned is make sure kind of all like minded in their enthusiasm. Make sure they're like diverse because you want people that have, like, come from different functions, but have that kind of innate intellectual curiosity. So in that foundation, it was really getting people comfortable, feeling safe, and encouraging individual experimentation. We brought in tools like ChatGPT, Gemini, a few other tools for different teams to experiment with, just to get their feet wet.
Then we've transitioned and we're more or less in the 2nd stage today. We obviously wanna get to that 3rd stage, is working with different functional teams to identify kinda limited high impact and low complexity use cases. I think we tried to conquer some really complex use cases upfront and pulled our hair out kinda trying to figure out how to solve them. Part of it was some of the tools and technology just weren't there.
And one thing with AI that, you know, is good and bad is it's, like, constantly changing. You think you might solve something and there's a new release by a vendor or there's a new tool out there, and you start to scratch your head and be like, oh, like, did we, you know, did we make the right decision? And I think that the lesson learned is to expect that change to happen. Expect, you know, new technology or new innovations to trump what you've already learned. I think that comes part and parcel with where we're at and within that piloting experimentation phase. And depending on your budget, depending on how many tools you can bring in, or build, like, kind of set that as, like, you wanna experiment. You don't want so many. You don't want so few where everybody's just trying to experiment within ChatGPT. Do you wanna actually figure out what the best tool that can kind of address the pilot use case? And then you're gonna test the tool. You're gonna iterate. You're gonna optimize. You're probably gonna go back to test, and you're gonna go back to iterate before you can kinda standardize important takeaway is I don't think the experimentation ever ends. I think Uzair touched base on it.
The experimentation is kind of constantly going on and eventually things can kind of shift to the right most category, the right the right column here is that, like, that's pretty sufficient. It actually, you know, hits on the ROI we want, the outcomes we want, and it's you know, the tools are accessible either independently or integrated within some of the existing platforms we use today.
 
Uzair Dada
37:11 - 38:19
And I think one of the important thing is you'll have a lot of people on that pathway that'll start disbelieving because they didn't get what they wanted in an easy way. And they haven't invested to learn the tool or iterate, or they're expecting perfection. It's just not that. You need to have the right guidelines and mindset of what you're gonna see through the journey.
So I think that's another interesting learning to me was, sort of, how do we sort of get that to happen? I mean, these things are, as he's saying, is changing every day. We're in a constant beta mode or alpha and beta mode. And so we sort of need to think that, and that doesn't mean it's not useful, but the velocity of change is absolutely incredible. And so I think making sure that you guys have the right guidelines and ground rules laid out about we will see this, and it's learning by doing. And people who are innately more curious in learning by doing, they wanna do more. People who are expecting perfection day 1 will be very dismissive. So know that and manage for that.
 
Samir Mehta
38:19 - 39:03
And for those of you regardless of where you're at in your journey, I love geeking out about this stuff and would you know, happy to hit me up over LinkedIn or email, but, like, I love like, everybody's kind of on a path, on a trajectory, and, you know, finding a community of people, you know, outside of your organization to talk shop with, understand how they're leading a transformation or, leading this initiative within the organization. Happy to connect and chat because I think that this is, a lot of people are trying to figure it out, and, hopefully, we can not repeat each other's mistakes.
 
Alex Jonathan Brown
39:03 - 39:56
If you're looking for a shorthand right now this week, great time to ask in your Slack or Teams or whatever you have. Hey, who's played around with Deep Seek? It's been out for 5 days now. The people who have experience already are the people that you wanna pull in as fast as possible with those conversations.
One more news cycle and that'll all change. But if you act real fast, you've still got that. But this we're not just there is a part of this that's, hey, we're playing with AI for fun. But there are still desired outcomes. And if you're gonna invest in this mentally, invest in this with time, perhaps invest with this with money, you do eventually wanna see some kind of outcome. And these are kind of the 3 that you pulled together, Samir, of what we see often as those desired outcomes. And I wonder if we could dive into those a little bit.
 
Samir Mehta
39:56 - 41:02
Yeah. I kinda want this to be a little bit of a roundtable here. And I would love like, kinda you know, there are many different outcomes, but if you roughly bucket them into these 3, like, it's like, productivity efficiency. What can you use AI for to improve customer experience, both presales, postales? And what are things marketers and salespeople can do to, like, drive or impact, like revenue acceleration, revenue growth.And if you look at a map of tools out there, there are, like, a 1,000,000 other categories. But I think foundationally for marketers and and kinda, go to market folks, like, it's gonna revolve in one of these 3 categories.
Now there are tons of use cases associated to each of these that will vary by individual, by team, by department, by organization. What I wanted to maybe go around and maybe we'll start with Uzair and then Alex and myself, but talk about, like, you know, a use case that you guys are using in the tool and, you know, what's been your little mini journey in terms of, like, hitting 1 or 2 or 3 of these outcomes?
 
Uzair Dada
41:02 - 43:53
Yeah. I think for me, the few things that I am most excited about and has really transformed how I do what I do in a lot of things is one is, I'd say, just research and synthesis.
And so prepping for a topic, prepping for a meeting, prepping for anything, the ability to use things like Gemini Advanced to sort of do a research report on a topic and get smart about something. Or being able to use Notebook LM from Google to organize the content and information and thoughts for meeting prep and be able to just query or have it kinda summarize into a podcast. I can listen to the 5 minute clip when I'm working out. That to me is just rad.
Like, the ability for me to be better prepared around anything I'm doing, this meeting or a topic, has just been amazing. Like, things that I never got to that I was running to catch a bit of, I am so much more prepared because of the ability to synthesize information, condense information, and give me that snippet.
The second one that I think Samir exposed me to, a little bit ago, was the ability especially with all the conversational AI and voice stuff, it came in kind of more pervasive, is being able to literally, use and talk to a GPT and drive a conversation around any topic. So, you know, Samir is like, when I'm driving, I'm usually half talking to the GPT about anything that I'm thinking about. I do that all the time now. It's amazing.
It's sort of like a perfect colleague, virtual colleague, that you can talk to and just riff and just talk about things in kinda random times. I prefer conversing than writing. So to me, it's a really awesome modality. And another thing is I am also one that sort of reads and thinks about a lot of things.
And oftentimes, I have ideas that come up to me, like, in random times and random things, and being able to iterate that and and have it captured and be synthesized that I can communicate to Alex saying, Alex, I wanna go talk about this topic as a blog. And he's like, oh, this is really cool. What if we did it this way and that way? But rather than having a 2 minute conversation. So it's in some way, it's very good for iterating. It's really good for research and synthesis, and it's very good for helping frame things that can be used to further the conversation within the organization or whatever else you're doing. So that's it. Those are the 3 most inventive ways that it's changed how I do things as a leader in this organization. Love to hear what your guys are sort of interesting use cases.
 
Alex Jonathan Brown
43:53 - 46:12
Yeah. For me on the content side of things, it's actually been sort of transformative, but also incredibly straightforward. If you've watched a lot of these Coffee Breaks, you probably have a sense of my whole deal by now. I'm kind, I'm kind of a handful.
And so I was never thinking, oh, AI is gonna do the thing that I do. I try to bring a lot of, like, my personality to work. Sometimes that's good. Sometimes that's bad. And so it was never, oh, AI is gonna write the way I write better than I am. I've never been worried about that. The thing that it's turned out to be really helpful for is, I let AI write bad for me all the time. Like, do my first two drafts for me, AI, and then I'm gonna come back in and add the little alleged magic that I bring to things on top of that.
And that is such a time saver. Because if you're a creative person, you can spend an incredibly long amount of time watching a cursor blink at you while you're like, my brain's not working. What am I gonna do? How do I get this started? Just have AI do all that, and then come back and do the magic, the stuff that, frankly, is what you usually get paid for. Do that on top of it, and don't burn all those hours trying to get to, like, a minimum viable product.
The note on that is, I know that creatives, especially if you're thinking about reaching these conversations with creatives in your organization, don't tell them this. But it's easy to be for creatives to take the stance of, like, listen, I'm gonna be John Henry, and I'm gonna work so much better than this machine does. And the problem with that is that John Henry wins the race against the steam drill and then immediately dies. And don't do that. What if John Henry got really good at driving the steam machine? He would be employed, and all of his kids would still have their father. So that's what I bring. So, yeah, use it to do the work you don't wanna do. Get really good at that, and you will save you and your company so much time.
 
Samir Mehta
46:12 - 49:25
Given that my team focuses on customer experience, an area that I am personally just fascinated and intrigued by is how the buyer experience will change with all these tools. Naturally, people are starting they're buying journeys on tools like ChatGPT or Gemini.
If you all haven’t played with Gemini deep research, it's a tool that can write a full on report about whatever topic that you want with, publicly available citation, citations across web pages. But the way in which we research, you know, maybe we're not going from site to site to site as much anymore, and we're finding ways to kind of actually front load our research, not just aggregating and compiling it and synthesizing it, but even actually delivering it in the way that you want it delivered.
So if you're an audio person, if you play with Notebook LM, you can generate an audio podcast. If you're on the paid version, you can actually converse with the podcast host about whatever topic. So this multimodality aspect of, or capability of Genini is like, super interesting in terms of how buyers will consume information. And then if you think about when they actually want to get in touch with your company, your brand, and they arrive at your website, or have a conversation with, could be an AI agent or it could be a human.
But the notion tthat experience will like, the kind of the anti on personalization is gonna go up way more. It's gonna be a lot higher because they've already learned all about your company in, you know, 2 or 3 page report, done a future comparison, and really they're coming there to probably find very specific answers to questions that perhaps were not available using an AI agent or they're looking for a particular experience.
And so that level goes up and so kind of there have been chatbots, a lot of companies have chatbots on their site. Most of them are rule based. Something that we prototyped and now is actually launching on the Iron Horse site, thanks to Ezra, Logan, and our engineering team is our own Gen AI-powered custom chatbot. The reason we did it was we really wanted to control the information that goes into that chatbot and control the responses that come out of that chatbot.
Today, it's just text based, but we have a vision to turn it into a multimodal experience that really kinda carries that chat-based interface into the Iron Horse site and really starts to deliver on that personalized experience that I think buyers will increasingly expect. So kinda, you know, love to geek out about that as well. So, like, an area that we are passionate about investigating both on our own and with some partners.
 
Alex Jonathan Brown
49:25 - 50:01
And speaking of geeking out, we have just been throwing AI tool names around like it's going out of style. We're gonna make that problem a little bit worse and do what I think is some of our favorite things to do. Let's just name some tools, y'all. Obviously, there's different stuff you wanna use at different stages, but I think by now, everybody's kind of got their favorite little pets. But, Samir, kinda take us through some of the options of what's available and really the categories of all of these different things.
 
Samir Mehta
50:01 - 52:30
So one thing that's super interesting in the AI space today is we all sort of start in the general AI tool bucket. So largely stand alone conversational tools like ChatGPT, Claude, Gemini, Perplexity. That UX pattern is very popular, highly emulated. DeepSeek looks very much like ChatGPT.
And they're really good at kind of being the Swiss army knife for you as an individual and even for your teams. All of those platforms are starting to kind of increase the ability to build some specialized agents, but, like, kinda that's a general tool.
Copilots are things that you've probably seen. You're inside of HubSpot. You're inside of Adobe Creative Suite. You're starting to see this little, you know, a much better version of Clippy pop up and assist you, and be your guide, your partner, your co-creator in whatever task you're doing within that application. It's hyper context aware. These are some of the tools, like, that we use internally. Like, we use Fathom for meeting transcription. Breeze is actually the AI Copilot inside of HubSpot. Design team uses Adobe Firefly. There's some folks using Writer. Goldcast is, obviously, a webinar platform, but we use it to actually generate, post-event content, and then our engineering team loves GitHub Copilot. So as you can see, like, the Copilots are there.
Largely, you're gonna be incumbent tools, but there are increasingly new standalone tools coming about. Specialized agents kinda like, now are starting to move into this concept of autonomous or autonomy and meaning, yes, as a human, you need to actually spend some serious time and effort and thinking to set up and configure the agent, but once done, it really can operate with minimal human supervision.
So Webflow Optimize is a phenomenal web based, on the fly personalization engine. AISDR is an AI enabled SDR platform that we're experimenting with. Allgood is a marketing ops, agentic platform, Clay. If you haven't heard of Clay, check it out, but it's amazing at doing research and prospecting. And Common Room is another tool that we've seen. There are so many tools in every category, but the thinking here is most people start on the left. They graduate to copilots increasingly. You think about piloting and experimentation, the easiest places will be, hey, can I get a generic Swiss Army knife tool for the entire company or for my department or for my group? 
Graduate to copilots. Copilots are gonna be visible in just about, you know, whether you're using apps like Google Workspace, Microsoft, Office 365. It's just becoming more and more apparent, and they're launching on my like, they're launching so fast and rapidly, create space and time for your teams to understand how to use those agent or those copilots.
And then specialized agents, really is moving towards that autonomous realm. The last and this is very experimental. There's a lot of research being done in truly fully autonomous AI systems. It might be science fiction, but there are a few players out there beyond your usual suspects like OpenAI, Google DeepMind, and Anthropic that are building platforms, capabilities to help build autonomous agents.
Devin got a lot of press. We don't use these tools today, but Devin is kind of a software engineer that can give it a task or a product requirements document, and it kinda does everything end to end and even deploys and deploys the code after production.
AutoGPT, you should check it out. You give it a task, and you can see it reason and actually kinda do a bunch of different things.And there's custom GPTs and other things within that category. So a lot of these tools like, when you're looking at tools, it's easy to get overwhelmed. We have a framework that we kind of look at when we're evaluating tools. So we have a use case framework and a tool evaluation framework. Happy to talk about how we do that, separately. And then the bottom layer, which is super important, which like, it's easy to buy stuff.
It's harder to build stuff. But as models, Deep Seek, as Alex mentioned, as models get cheaper, etcetera, etcetera, I think you're going to see more building of tools that truly adapt to how your organization wants to achieve its outcomes. I don't think it's gonna be all buy and all or all build. It's gonna be a blend. And I think this is kind of if you've heard the term of agentic AI agents, I think you're gonna start to see more companies building things internally and enabling platforms for their teams internally to build specific  use cases on top of.
 
Alex Jonathan Brown
55:07 - 55:19
That is a great rundown. Uzair, is there any, A, anything you wanna add to that or, B, any just personal favorite, either tools or copilots or any other personal favorite? 

Uzair Dada
55:19 - 55:19
Yeah. The only thing I would say is, like, the interesting part is your favorite applications are evolving. Salesforce with their agentic framework, HubSpot with agent AI. So don't think that they won't be around. They're all doing some interesting stuff. And to me, the interesting part is today we are hyper dependent on platform UX.
Tomorrow, we will still have the platforms, but we will be more UX independent. And, the agents will sort of be our mechanism to consume and act and do what we wanna do in a way we wanna do it.
So I think that's gonna be an interesting transformation. And the only other thing I would wanna say is that vertical applications are a really interesting space, and that's kind of where a lot of growth is happening. So, you know, some of the early markets where there is huge adoption is legal, financial services. So you'll see a lot more interesting use cases come out, health care and clinical trials. There's a lot of cool stuff happening in those verticals. So beyond just the horizontal stuff that a lot of stuff we touched on, there is dramatic transformation stuff happening in the vertical application space.
 
Alex Jonathan Brown
56:27 - 57:07
It is a very exciting time to be doing any of this. Thanks, Samir and Uzair, for, I mean, this conversation is stopping because our webinar platform is telling us to stop. We could do this for the next 2 hours. I'm sure we will get this group back together and do it again soon. Those of you who've joined us, thank you so much. If you wanna learn this is the part where I usually do plugs, but since I work in content at Iron Horse, I know all of your plugs already. If you wanna learn more about what these guys think about AI, we've got a blog on the website from Samir about, you have the agentification blog, right?
 
Samir Mehta
57:07 - 57:20
Yeah. Just, you know, we talk buyer patterns changing and how agents, we will all have our own personal agents working for us.
 
Alex Jonathan Brown
57:20 - 58:11
That is exciting. And I even now, I fight the urge to be like, that sounds terrifying. Yeah. But it's because agent is a funky term for that. Uzair's got a blog post up about the current Gen AI evolution we are in and what that means for you and your company moving forward. I have a Coffee Break about how to use AI like a human.
Three different takes on where things are and where things are going. But if you enjoyed this, spend some time over at ironhorse.io. Get some more of those ideas, and then join us back here for, A, the next Coffee Break, and, B, the next time the 3 of us talk for a full hour about what we're excited about, in the AI space. Guys, thank you so much again. Any parting words in the next minute and 50 seconds?
 
Uzair Dada
58:11 - 58:14
Be curious.
 
Samir Mehta
58:14 - 58:25
Be curious. Encourage failure.
I know it's, like, controversial, but, create a safe space for your teams to experiment and fail and then succeed.
 
Alex Jonathan Brown
58:25 - 58:37
Awesome. Like I said, thanks so much for joining us, and until next time.
I mean, this month, your Coffee Break ran long, but either way, it's time to get back to work. Bye, everybody.
 
Samir Mehta
58:37 - 58:39
Thank you, Alex.
 
Uzair Dada
58:39 - 58:39
Cheers.