Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
Andy Sack (00:23.202)
This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation in business. I'm Andy Sack along my cohost, Adam Brotman. In each episode, we bring you candid conversations with business leaders who are transforming their businesses with AI. No fluff, just real talk and actionable use cases and insights for you.
Andy Sack (00:56.792)
Today we have the first of what will be a multi-episode journey with Dan Todd from Influence Mobile. And we thought it would be really good. Dan is someone that's working with Forum 3. We thought it would be great to highlight his onset of implementing the AI First playbook and working with Forum 3 to transform Influence Mobile. Welcome Dan and Adam.
Thanks Andy, glad to be here.
Dan, let's kick off right away and just like give the audience just a brief introduction to Influence Mobile. Why don't you tell us about the company and what it does?
Great. Influence Mobile has built an app called Rewarded Play. It's in the Android App Store, and we invite tens of thousands of people today to download Rewarded Play, the app, and then discover new games. It's best to think about it in the context of like an airline rewards program. So as people download and play these games, we give them points towards real-world rewards like Amazon gift cards. They can earn very large rewards, $100, $500 playing these games.
in certain time frame so game developers pay us anywhere from five to twenty five dollars every single time a person downloads one of these games and then we use a portion of that money to fund those rewards and so we want to encourage people to quickly get into the game experience of games enjoy the games but then what we're really looking for is the people that spend money in those games they they drive all the value for game developers we put a lot of focus on finding and keeping those people we have
Dan Todd (02:32.142)
retention programs and all kinds of rewards for those people that spend money. And we're generally giving them 10 to 20 % cash back, which keeps them around and engaged, which makes other game developers want to continue to buy traffic from us. And then we just kind of continue to repeat that cycle.
And Dan, can you tell us a little bit about the sort of size and scope of the business in terms of employees and revenue?
Yep, so we have about 65 employees. This year we'll do somewhere between 40 and 50 million dollars in revenue.
Great. And that revenue comes, your core customers are whom?
the primarily the largest game developers in the world. So companies like Scopely that build Monopoly Go, that game generated $2 billion in its first 10 months. They pay us to get people to play Monopoly Go, whether it's Candy Crush or Solitaire Games or whatever the most popular games in the world, they're primarily looking for people who want to spend money in those games and we help them find them and keep them.
Adam Brotman (03:33.912)
You have sort of a Mr. Beast model. in the same sense, gets paid for generating audience and he gives that money back to fund his business model. And you have a game publisher, game player loyalty version of that. So that's a cool business, Dan. And on the AI front, we've been working with you now for a year and you clearly as CEO and founder of the company, you are
a believer in that this is a game changing technology that business leaders and CEOs and teams need to be taking advantage of it. But you're also self admittedly a company that is still going through its transformation journey on AI first. And that's why we wanted to do this mini series with you because we thought it might be a fun thing for other people to sort of see in real life, like what happens and kind of where other similarly situated companies and CEOs are.
So can you start out by telling us a little bit about like set the scene for us? Like, you you believe in AI, you've gone through one of our boot camps personally with your team, but you know, if you were to sort of describe, you know, how often you're using either ChatGBT or Gemini or Cloud, how often your team is, kind of set the scene for us in that kind of where you're at, like the baseline.
Sure, yeah, I think when we first started talking with you guys, there was a lot of concern about using JAT-TBT because there were stories of people putting personal company information into the system and it getting distributed everywhere. you know, it was the, you know, if today is the Wild West, I don't know what it meant a year ago, but we were both anxious to get started, but also hesitant to not do something wrong. I would say we've largely,
Secured ourselves against doing things wrong. You know, we have corporate accounts for Gemini that is very secure. We have an AI policy where we are clear with our employees what they should and shouldn't be putting into AI. But beyond that, you know, so it's been intentional, but I would say it's still very early. We started making people aware of it. We started encouraging users or employees to be using it regularly.
Dan Todd (05:54.668)
but we haven't necessarily followed up on it consistently. I'd say usage varies. use, I use chat TBT so much, my wife stops responding to my questions because she thinks I'm talking to chat TBT. So that's how, that's how much I've talked to chat TBT all day long on every topic from personal to business to everything. just becomes a source of creativity and bouncing off ideas.
I would say we have other executives who maybe use it that much. Some people are using it less frequently. Our dev team uses it. I would say each developer has certain elements of their day-to-day flow where they use AI. But we're kind of just getting into this phase with you guys where we want to start to systematize it, make sure that we know what are the best use cases for each employee, for each department, make sure that we're...
giving them all the tools because those are constantly changing and we're excited to work with forum three to see what the best tools are today besides just chat tpt. I'd say there's moderate usage. There's people who are embracing it just like any other company and they're doing fun stuff on their own kind of solo. was just talking to a guy today who's setting up his own servers on render using N8n to create AI agents. like there's a lot of independent thought that's happening and we're encouraging that. But we want to kind of bring it all under one roof.
and make sure we're all using it to the best of our abilities and also not wasting our time doing tools that aren't going to be effective.
This whole podcast and this idea behind a tracking the influence mobile AI first journey came as a result of about, I don't know, six weeks ago, you and I were having coffee and at coffee were like, I want to push and become AI first at influence mobile. And you were picking my brain about that. Can you tell me what, what prompted you to say we need to go AI first?
Dan Todd (07:57.272)
Well, I just from those early wins, like, I mean, we have, we have an incredible amount of data, for example. So like there's the way I think about it, you can apply AI and a lot of different functions, right? So for content creation, it's incredible, right? So it's very good at. I can just give it a stream of thought and it can give me a condensed version of an email to investors or to whatever in like a millisecond. And then I can tweet that. But when I start thinking about like what I could do.
with AI, if I could access my own data, connected AI would be incredible. I have to spend hours sourcing data to get an answer that lets me make a decision in one minute, right? So like the data collection process is very onerous. And so if we can get to the point where AI just has access to all of our data, there's so much opportunity there alone, apart from anything else that I've been talking to my BI team. I'm like, we have to get
or AI integrated into BI. Then when you think about developers, right? If content creation, when I'm creating emails is enhanced, how much better will code creation, code testing, code documentation? Like we spend a lot of money every single month hiring and paying for developers. And if we could make them 20%, 30 % more effective, that just scales the growth of the entire company because we're a product-driven company getting products out.
making sure they're tested, making sure people know what they're supposed to be doing and that they're doing them. So just everywhere I look, it just seems like at a minimum AI could bring 20 to 30 % lift. And then that directly translates into company growth. And so then when I hear you telling me stories about other companies who are doing innovative things, I'm like, we got to get on this, right?
If you were to describe, you have a, you mentioned you have an AI use policy. Like, would you say that they're using different platforms? Would you say that they're very, really varied in sort of how much they're using it how much they're getting out of it? Like describe a little bit about like where, you know, how you would rate or describe kind of where, your company's at.
Dan Todd (10:07.852)
Right. Like I said, I would, I would, I don't follow up with everybody and ask them about, their AI usage regularly. we were doing in our bike by week or by yeah, biweekly company meetings, we were having like an AI segment and kind of showing people how to use Gemini. So there are people who kind of use Gemini primarily. And then if you, if someone comes to us and has a good use case to switch over to chat, TPT, we let them use chat, TPT, for example, chat, TPT just.
really improved its image creation quality. And my wife who actually works for our company does all kinds of creative marketing stuff. And so I just told her like, you should switch over to ChatGBT because I was testing those same image creations on Gemini and it was inferior by far. So depending on the use case, some people prefer ChatGBT. I'd say majority of employees like the developers have it just because we're a Google shop and we're integrated into all the Google sheets and
those types of things. To be honest with you, I can't tell you other than kind of anecdotally, you can tell when people have been using it for their emails. And so like you can tell who's using it for emails. We certainly encourage it, but that's kind of where I'm hoping to get to the point where, you know, if AI can improve my productivity 20 % or 30%, I want to make sure that that baseline is being applied consistently across the board.
What outcomes are you looking for? What does success look like?
I mean, I would say success looks like it can be as little, right? As like you guys telling us one or two nuggets that some other company has done. And all of a sudden we change one process that adds a million dollars in revenue, right? Like that, even though that wouldn't be what I would be aiming for, that could, that would be a success. I think across the board, making sure that employees aren't wasting their time. I can, there's a lot of experimentation. Like I said,
Dan Todd (12:07.692)
Some of the experiment experimentation is awesome because you've got employees showing initiative, but there needs to probably be some cohesive strategy there. And then just make sure that we are using the best tools in the most effective way for improving our people, you know, the people effectiveness, the processes, like are there things like the data collection and AI? I mean, we could be wasting tens of thousands of dollars a month in people time, you know, pulling reports. We're not.
We're not actually spending that time because it's so onerous. don't do it. So then we're losing out on the insights of all that kind of stuff. you know, improvements in people process. And then of course our product needs to have AI from fraud protection to app selection. Like there's all kinds of things we could be doing. And I don't know the first step to do it because I don't want to necessarily cram it down my employees throats that they have to do AI and everything they do. But I want to show them.
the upside by showing success stories that you guys have seen elsewhere and encourage people to do it out of excitement.
So Dan, let's use that as a good segue. You actually just said something that I want to build on. You said, I don't want to cram it down my employees' throat. And I think that's actually a and a common sentiment that we hear a lot. Like you actually mentioned two things. Like we hear a lot of people being scared about their data being leaked or available in the public domain by accident through this process. And then secondly, I feel like there's
There's not quite yet a cohesion between CEOs that are getting it like yourself and the rest of their team. seems to be some interesting impedance there and I want to kind of dig at that for a second. do you, I'm going to, you you're coming on this podcast, you're being a little vulnerable. It's like a reality television show a little bit. you, so I'm going to ask you, like, are you feeling that
Adam Brotman (14:11.692)
Your employees across the board, a hundred percent, your leaders and your team, a hundred percent are like, we want to lean into AI and we are leaning into AI and don't worry, Dan, we got this. We're excited about it. Or do you feel like it's, there might be a past of aggressiveness there that might be going on that you're not aware of or somewhere in between.
I would say somewhere in between it. I don't think there's passive aggressive. think, you know, we, we, we ask and encourage people to give us positive feedback. I don't think there's really any reason for people to, dislike the concept of AI. It's just, they don't know where to start. Right. And so you can tell people go use AI or you can show somebody a single use case that will blow their mind. And like, all of a sudden they're excited to use AI. Right. And so
but we have a lot of different employees in lot of different roles. so like for sure, the situations where again, like in our marketing team, right? Like we create thousands of ad creatives a month. so in video ads and all kinds of different things. And so the amount that AI can help with the creative process is just enormous there, right? And so there can be hesitancy in the creative world because artists really like
to have ownership over the creative process. So I'd say there's occasionally hesitancy there. But beyond that, I think it's just a matter of like helping people get into the habit of these use cases that save them time. And then just building up a functional library of use cases where they're like, when I do this thing, I use this tool, right? And then eventually, just like I use it now, it becomes...
something that you enjoy using so much because you see the benefits of it that they start incorporating it in other areas.
Adam Brotman (16:03.374)
Do you actually yourself, I'm just curious, like this is both a question and a bit of a best practice, I think. Are you, when you're having discussions with your team, do you have it sort of on your mind to ask the team almost every time? Like, hey, what have you done to try to further this or improve productivity or improve quality using AI? Like, is that a question you're asking every day of your team?
No, I wouldn't say no. It's definitely not. don't ask every day. It comes up. It comes up. There are people who are more interested in more excited about it. Now part of that's because I just know we're ramping this whole thing up right now. So, you know, I just, my last call, we spent a significant amount of time talking about AI tools with my product. One of my product managers who's very excited about stuff. He's being proactive. He's setting up servers. He's like, I'm not taking up any dev time. I'm creating my own server.
I'm setting up this innate instance that allows AI agents to do this, that, and the other thing. It's freaking amazing. I love it. Right. And so, he's in a different category though, in terms of his excitement and his ability to use that and see kind of like, he's kind of on the blue ocean of unlimited opportunities in the product side. Right. Whereas like maybe my customer support team, once they've got a bunch of scripts for emails, they don't need to like sit there on AI creating 50 more.
So again, I also think there's an opportunity to use AI more in certain areas. And so in those areas, I do bring it up more, but you're highlighting a potential weakness here, which is, yeah, I'm not beating the drum nonstop every day to make sure you're using AI. I tell stories about it and hope to encourage people, but we do want to figure out like what is the right cadence to tell people, you know, remember to use your AI tools.
I want to ask Dan a couple of questions. For our audience, this is totally unscripted. first of all, Dan, have you been playing with the deep research model of OpenAI?
Dan Todd (18:05.512)
I I used them all up. You gotta be very careful and uncheck that box so you don't use them all up too quickly.
I've hit that limit too. Yeah, it's pretty amazing, right? And is the rest of your team, like is that just you that has access to that? Cause that's actually, you know, at the pro level, I think they've actually now rolled it out to the plus level, but there's even less, you use up those things even more at the plus level. do you kind of, who has access to deep research and have you talked about deep research with your team at large?
I, I would know. I haven't talked about it at large. did mention it a few different times. Like I said, it came on, Andy told me to like pay for the, you know, the $400 thing. But then like right when I went to go do it, it was available at my level. And so I started using it and I just went crazy. I didn't really realize I could use it all up in like two days. So I just kept doing it for things that were not useful. Like I just kept the stupid button on. And I would say I only got two probably like really good
forms of the deep research done. I'm now like a few days away from my next month of allocations, but it was very powerful and I could see using it as I get more personal use cases under my belt. I'll probably start talking about it more with employees, but I don't recall having a conversation with anybody about it just yet.
So Dan, two topics that haven't come up that I'd like you to just touch on lightly. One is Azure increasing the promotion of AI within amongst and with your staff. How do you deal with concerns of the employees around specifically being laid off or job replacement?
Dan Todd (19:48.226)
I've been pretty clear with everybody that I actually think the biggest risk of people losing their jobs is that they get replaced by people who are better at using AI. actually don't see like, we're not.
Do you hit that straight on? that like a, you hit that straight on with the employees?
Yeah, I mean, what I would tell people is we're not a 1900 person or 19,000 person company where there's massive, you know, such massive potential for layoffs that you could see that happening. What I want to encourage people is like, we want to get into the growth mode so that when we start growing again, we just need less people. Right. So that's really the focus is like we want
to be more effective with the group that we've had. I've even toyed with this idea that if a group of people could become 50 % more productive, right? So in 40 hours, they could do 60 hours of work, and maybe I would give them one day off, right? So like we split the value. So I don't want them to think that I'm...
I specifically go to a four day work week is what you've sort of promised.
Dan Todd (21:06.326)
Right. Well, I didn't promise anything, but I'm saying like, I don't want them to view it as like, I'm using AI as a tool to squish out all of the value and they get nothing out of it. Right. I want to figure out how do I make it a win-win? So my employees are more productive. The company is growing faster and maybe it's not a whole day. Maybe it's half days on Fridays or something like that. But I want to take the tact where people are excited because it's not just going to make them more productive.
and ultimately more valuable down the road, but that it actually might make them do more work in less time and then they get some time off for their own families and free time.
Let's transition. Adam, what jumped out at you from, mean, this outlets, that'll conclude sort of our interview of Dan. I'm just sort of, want to have you summarize what jumped out at you for the audience from this first of the mini series.
I, well, I wrote down a couple thoughts that are Dan that you mentioned that I thought was interesting. You you've mentioned a couple different times this notion of a productivity lift of 20 to 30 % and, that you're you've seen and you believe without even having us prove anything that if somebody's leaning into AI tools, whether it be for strategic thinking, decision making or
Creative, you know, working, marketing, coding, BI, you've mentioned these different functions and you're like, I believe that if people lean in, they ought to be able to get to 20 to 30 % lift. And if enough people are doing that over time, that's going to translate to the bottom line of your company, both in terms of growth and in terms of efficiencies and making better decisions. So that was number one that you kind of mentioned that you also mentioned something else that I was really interesting, especially for Andy and I, when what we've been talking a lot.
Adam Brotman (23:04.93)
amongst ourselves lately about, is you mentioned, you're gonna get more out of these AI tools, the more easily you can get to the source of data that you know is gonna be relevant to the problem you're trying to solve or whatever. you're like, God, you spend a lot of time getting the data together and then you get these great answers. so I wanna sort of...
you know, put an asterisk or a highlight next to that, because I thought that was the second thing you mentioned that was really interesting, which is like, can we help you or can you help yourself to like solve that a little bit? Cause that can be a big unlock to get to your 20 to 30 % productivity lift. If you get after those sources in a more time efficient manner. The third thing that you mentioned, which I thought was this interesting is you, you gave some examples of, know, how your, your team.
is using it in different ways, some more advanced, some less advanced, and that there wasn't so much a general resistance to it, but there was an unevenness to it. And you were looking for ways that you could sort of make that a little bit more fully integrated and present in your team. And that was a challenge that you're kind of going to take on in the next 90 days after having coffee with Andy and hearing, how are, and wondering how other people are
So those are the three things I took away from this call so far and kind of sets us up for the next 90 days.
Yeah. Well, I would say if we can solve the data thing, it's a lot bigger than 20 to 30%. I think unlocking the ability to quickly access data is like exponential, like more than a hundred percent improvement because I don't do things because it would take too long to crunch all the data for an unknown upside. Right. So like we launched this new product that's been incredibly successful called Boost. It's all new data.
Dan Todd (25:04.63)
So, you know, in my 25 years, there's done a lot. I've done a lot of data work. So I kind of have a spidey sense for what certain things should look like, but this product is not that. so, but I just don't have the time because like literally like each of these campaigns are running and optimizing against different numbers. And I have certain reports, but I would ask myself like, Hey, I wonder if I can figure out this whole thing. And it's either going to take two weeks, literally two weeks of my BI teams time to go.
set up a report to pull all that data, or I might spend 30 hours trying to pull that data and it could just be I picked the wrong thing, like it wasn't a meaningful data, but there's for sure massive insights, bigger than 20 to 30 % improvements if I could just get to the data and ask questions of it without having to like, you know, build it into reports.
Yeah, that's awesome. Dan, a couple of things that I want to highlight from what you said. First, I think you're, you know, as a CEO, I mean, the audience really for this podcast is C-suite and CEOs. I think you did a great job early in this episode highlighting sort of just leading by example, your own usage and process over the last year.
of starting to adopt AI yourself and getting comfortable with it and getting over the, the concerns and to the point where you said six weeks ago, I want to become AI first. That's sort of best practice. Like the, think the CEO needs to lead by example. And, and I thought your example of using it verbally and talking to the, to AI to get your use up. I thought it was just a great anecdotal story. So that's one thing I wanted to.
The other is, just you're looking the way in which you're approaching it with your employees being upfront and transparent about the productivity gain and dealing with the job replacement, which is, think, looking for win-wins with your employees. And I think there's a variety of different ways to approach that, but your example of possibly splitting that. it's a win for the employees as well. I it's also really.
Andy Sack (27:26.164)
exemplary and other people can learn from. I'd also highlight the, and I think it'll be really interesting to track the real functional use cases. Is it customer support? Is it coding? Is it BI? And I also echo Adam's like just from a personal standpoint, we're going to have lots of questions about accessing data, getting access to more enterprise data more easily so that you can make those kinds of
analyses that you were just referencing. Those are the things that jumped out at me.
Andy Sack (28:04.172)
Welcome back Dan Todd, CEO of Influence Mobile. Glad to have you on the pod again.
Yeah, happy to see you both. Thanks, Andy.
So this is the second installment of AI First following the AI First transformation of Influence Mobile. And we plan on doing at least one more installment. It would spend about five or six weeks since we first recorded and started the digital transformation process in Influence Mobile. And just to set the stage for the listeners or remind the listeners, I've been working
closely with Dan and his AI counsel on a weekly basis to support them as they go through their digital transformation. Adam has not, so this is really Adam's first check-in. And we're going to hear Dan's reflections on the process as he's seen it for his company, as well as for himself and the positives, the challenges.
the negatives, if there are any, and what his key takeaways are. So with that context, let's dive right into it. Cause we have limited time. Dan could like tell us this, tell the, tell the listeners, the story of the last six weeks and what you've seen from your team, lay that landscape out for us.
Dan Todd (29:27.778)
Yeah. Yeah. I'd say there's two different, there's two different paths, right? There's my usage, which is, which is influential because I do a lot of things that drives like the product implementation. And then there's me pulling people into, we have an AI council meeting with Andy. So, you know, it's a, it's a pretty big deal for these guys to be on a call with me. And we're, you know, constantly talking about AI and the first week we were on there, several, several of them were like,
I don't use AI too much. And it became pretty clear that that better not be their answer two weeks later. Right. And so we've seen everybody's usage step up. And so each, each week, Andy would sort of bring new things, best practices that he saw elsewhere. And we're like, okay, well, we'll apply that. So the first one was just start measuring it and start talking about it. And so we started trying to figure out how to measure it manually. I would say that was more or less a failure, but it did push us to.
get the enterprise license for ChatTPT, which has automated tracking. And then we have Gemini already rolled out and that has not as good, but reasonable enough tracking. we, but we were asking everybody to self-report. So it definitely made, it made people more aware that they needed to be using it more. We did see usage increases. And then on top of sort of just tracking weekly usage and reporting it, we launched a Slack channel that awarded winners.
on different topics. So best personal use, best multi-model use, best prompt, and then just overall usage. And so then people would submit their ideas on what, how they'd use it. And we started promoting those. And I think those were probably the biggest factors. Cause when people started seeing how creative others were being with AI, I think it was blowing their minds. People were taking pictures of their backyard and it was helping it figure out how to do gardening plans and people were creating.
all kinds of crazy things for their home or, uh, you know, eventually I got to this, the synthetic, synthetic personas where I had a fraud character and an anti-fraud character and experts in our business model and synthetic customers all talking about like what we should do relative to some product advancements. And again, it's a little bit of a step of faith on how accurate it is, but it was 100 % in mind expanding because it let me
Dan Todd (31:53.832)
ask a question and immediately take into account more viewpoints than I would have without that. And so we're already moving forward with product evolutions that came from those conversations, primarily moving more towards choice where players will allow players because there were so many different, there's four different personas that came out of like 250 survey responses. And they became clear that like,
They all weren't motivated by the same thing in our rewards model. And then as I shared that people were like, it was blowing their minds. And so then you started seeing people who maybe were skeptics using it. And so I heard, just had a call last week with a guy who had previously wasn't going to invest in AI. He thought it was a fad. He used it and he said it did about a week's work worth of refactoring in less than 30 minutes. And so you're talking about 40 hours of work.
done in 30 minutes, like that was mind blown. then as that's, you know, as those stories continue to get perpetuated, you just get everybody excited about what they can be doing. And so I think that's where we're at. We're continuing to see greater usage. We're applying the things we're learning, whether it's synthetic personas or multi-models, you know, I'm using more than just chat CPT myself. And each of those things is building up. And that was one of the things you guys mentioned.
or at least Andy mentioned four or five weeks ago was that some of the companies that are doing the best work, it's not one big thing. It's thousands of small integrations across all of your employees that raise the bar and raise performance.
Dan, what has been your biggest surprise? Like what's the thing that you were not expecting that's either good or bad or otherwise?
Dan Todd (33:43.694)
I would say that there are simple things that you guys have told us that have profound positive impacts on your AI usage, right? So like you would think, for example, just last week, Andy started telling me the difference between 4.0 and 0.3.
Those are two models for the listeners.
Yeah. And, know, I had been using grok. I'd been using other things. thought I was a sophisticated user, but I felt like I wasted my whole life because I had done all this work with four and as soon as I started using three, was like, Holy smokes. This is a whole different ball game. Right. And so, and there's just different like stages of learning that you guys were able to, Andy would share each week. Like,
You guys should think about this or that other thing. And each one of those things have been very substantial in its magnification of the successful use of AI.
Mm-hmm.
Andy Sack (34:47.118)
Dan, would you add a little bit of color as to why that was such a big surprise for you? Like, what did you done with 04 and then what did you do with 03 that had you go, oh my gosh, somebody turned on the lights.
Yeah. So there was two, two major amount. I I spent days, full-time days doing one of the other things I learned from you guys, which was building a bunch of synthetic personas. And so out of, think our second week call, Andy's like, you should look into this and build these personas and have these personas speak into the product, evolution, even your marketing messages. And so I ended up serving.
hundreds of our elite players and getting back a lot of information. And I went through this whole process of creating knowledge files and, custom GPTs with each one of these players. And then I wasn't satisfied with that. So I actually created like a meta version of that, where I invited seven different personas all to the same custom GPT. And, one of them was this, version called Randy core, which I spent.
probably five hours and it just kept getting so many things wrong. And so it just was actually raising my concern. like, how come it could just, you can't do the simplest math. Like, you know, it could get 85 % of the stuff, right. But then it, you know, it's telling me stuff that I can just see on the screen is not accurate, let alone when it's giving me advanced reasoning that I can't even comprehend, like, you know, and so as soon as I used O3 though, it started.
clicking in me that these were two very different things. Not only could you see all that stuff, Andy was like, you should go back and recreate all those personas, which I haven't had time to do yet, because this was just last Thursday, but I was doing a personal project for my wife, which was this round the state scavenger hunt. And 4.0 was horrible at it. It couldn't figure out the simplest of clues. And then all of a sudden I started using 03 and it
Dan Todd (36:52.662)
like blew my mind a level of sophistication and thinking through and like, Nope, that's not right. This is, know, and it now ended up being wrong on a lot of things anyways, cause it was like trying to predict the posters on the back of walls in hotels and things like that. Maybe things that it was accurate at some point in time, but became changing. But to your point, I don't know what a gentic means Adam, but I can't, you, when you can see the process of it's working itself through.
I'm like, felt like I was ruining. mean, I just wasted my time with four. Obviously when it's writing just a simple email, it's fine, but I spent a lot of time to a data analysis that I now feels completely wasted because I should have been doing it in no three.
Well, that's, that, that's a great, so that's a great call out because I feel the same way about a version of O3 called deep research, which you may have heard of, which is it's actually the same O3 model, but they've actually created a special thing called deep research, which is fine tuned for analysis and research and report writing. And I found that when I put a spreadsheet into it, it would one shot, meaning I didn't have to like work with it and explain things to it. It would one shot.
Nail data analysis, right? In a way I've never seen before from even for all with code interpreter. So like to your point with these models, there's a lot of trot. You got to like learn like, when would I, when would I use a reasoning model versus a chat model? When would I use Gemini or Claude versus a chat GBT? When would I like, when would I use perplexity versus search in chat GBT or Grok? Like they're all different, but they're, they become sort of a key.
part of the learning curve of being an AI first leader. And so I think what you're describing is very typical in that you've got to go through it and do some trial and error and kind of stub your toe and learn what works well and, you know, keep up with the stuff because it's changing every day. But it's, but you're right. Otherwise you could end up thinking that it can't do something that it can. And what I mean by the way, what I mean by agentic, which is worth mentioning for a minute, which is this notion of
Adam Brotman (39:02.082)
The promise of AI, believe it or not, when you really get into the mindset of the people who are building it, like the Sam Altmans and the Demis Esabazes of the world, who are like the leaders at OpenAI and Google, they will talk about that the future of AI is all about being these agents that are like these almost employees of yours. so you give them a task and they just go off and get it done.
check their work and then come back like an employee would. like when you start to look at a reasoning model, it's sitting there thinking to itself, well, Dan asked me to do this, but I really know the context that he's trying to accomplish this in his business. So I'm going to double check this other thing. Nope, that's not right. And so it's basically like checking its own work. It's thinking things through with context that you've given just like an employee would. And that's agentic, agentic, meaning like it's, it's, on its own going to go
do things, think things through, use whatever code writing or search or analysis tools it has, and then come back and help you. And then you realize like, wow, that's even, if I thought ChatGBT was good back a year ago, this is even more interesting because it's more like what a knowledge worker would do for
I mean, we, one thing that we didn't talk about in this call is we're sort of building towards a list of AI first initiatives and ultimately being a choosing one initiative. That's, that's more than just employee productivity. And so I expected that, but when you look forward, where are you excited to take to lead the company in terms of strategically with the help of AI as it relates to.
the next like let's just say the next four to six weeks.
Dan Todd (40:50.51)
Sir, well, I think the next four to six weeks is a little shorter timeline. So I think within the next four to six weeks, we'll continue to see more individual evolutions. We are working towards the first product integration towards AI. do have, Andy likes to call him a young whippersnapper who's actually all in that agent stuff. And he held the training session on explaining how he's got a bunch of agents.
working on things all the time. And so we're working on some data collection around device models that we think are going to be very informative in a variety of different areas. And so I'm going to work with him on putting some agents into analysis of that kind of stuff. that's actually probably going to be, it's not going to be a fully integrated live product, but it's going to be live data. Then we're going to start using AI and agents to analyze the
very important information for us to help predict fraud and also predict high value spenders. But then I also think just the next four to six weeks, we're raising the bar in our accuracy of what product evolution we're doing because we're challenging ourselves by not just assuming we know the right answer, we're putting this, you know, each thought of like, well, well, should we spend time building this or that and getting a more complete, you know, package of ideas before we determine.
Yeah, we're going to build this product over this one. So I'd like to think that will manifest itself with more productive developer work.
One thing that's worth that you just talked about that I want to highlight for the audience, cause you talked about it, which was what you, this whole notion of synthetic personas, what you did effectively was create a, you, you surveyed your, a subset of your top users, got real data from that and then use that real data to create synthetic personas, which you then invited to a
Andy Sack (42:54.132)
Meta chat bot. you're now talking to a subset of say five personas who are representative of your user base that's informed by real data. And then you also created another synthetic persona that has your business KPIs and business results so that you can chat with both the users and the bit and someone who chimes in all the time about your.
business results. And as a result of that, you then wanted to work on your onboarding. And I, you know, what you reported to me was you came up to a totally different conclusion than you otherwise would have by talking to that group of five synthetic personas. And I re-tell that story because you teed it up just a moment ago to make sure that people understood exactly what was going on. Because I, I thought that was a fantastic use case.
One that I had the idea of synthetic persona, but Dan, it was really your idea to create another synthetic persona with, who could chime in with this is what these are the KPIs, the day to day KPIs that we focus on and have that come in. it's like having a, a CFO business analyst always sitting over your shoulder as you're making product decisions going, I wouldn't do that. Cause that's not going to emphasize the right, the right, revenue results or the right profit results.
I would do this. And so you're basically getting a, I'm reminded of Abe Lincoln's, what was Abe Lincoln when he had the, the, group of, would, he would have, he would have, he would have a group of both, what's the term that I'm looking for? The, council of council of something of like different, he would have the Democrats and the Republicans, the council of elders and they would all chime in and then he would make a decision.
And I, and you're kind of created your own council of synthetic elders for influence mobile to assess your onboarding. And then it led you to a different strategic insight, a totally different path. And I just thought that that was awesome work, innovative that I want to share with the audience as well as with other CEOs. Cause it really, for me, it was, and I being along the journey with you, was
Andy Sack (45:18.54)
emblematic of your own AI first transformation. you've taken, you've stealing the pebble from the teacher's hand and going off and really innovating and leading. And so I just thought it was worth highlighting.
Yeah, no, it was really good because once you started, you know, hearing these different opinions, it made it made me way, you know, why, you know, generally the reason, you know, we have lots of ideas in our company and the trade offs are always how much works are going to be to test it. Right. And so you could have a great idea that's going to take four months to build.
or you, you could have an okay idea that's going to take two days to build. And that might be better to do the okay idea first, cause you can get it out the door. And so what this challenged me to think about was doing this choice option, which I had initially thought was like, it's too much work. It's too much of a risk when I've been sure that this new thing, this big time bonus that we rolled out, isn't the best. And so it forced me to come up with a conclusion, which was, well, how do we, how can we test that without all this dev work? And so, yeah, the.
The innovation was basically just put up a path picker that doesn't go anywhere. It's accurate enough to get directionally what players like what, and then measure their results afterwards. And so it'll probably take only two days of work and we're going to get 85 % of the value instead of having spent months and months and months building something that, you know, after two weeks of testing could conclusively prove that we should have.
chose path A over path B. So now we're going to get the benefit of figuring out the right path or proving that both paths exist in equal value to different players all because of the personas giving me their feedback. So I love it.
Adam Brotman (47:09.41)
Hey, Dan, let me ask you a question. Did you, just out of curiosity, did you create, did you just like create like a custom GPT with the different personalities saved as like different, like I'll call it knowledge files that, that, you could, you could have actually, it was you chatting and saying, Hey, I want you to now pretend you're this. And it knew how to go into that role. And then I want you to pretend you're this. then, or did you actually create something where they were like talking to one another without you starting the prompt?
did both. I initially created them all as individual. And then I found it was super annoying to like, you know, I'd take one person's feedback. Yeah. So then I created a group one and then, and then I just kept coming up with ideas of additional personas. And so, yeah, eventually I created this big round table of like eight different personas that all had different perspectives. And then I put in the information and part of the, but it was a custom GP.
GPT part of the prompt was they would talk back to them between each other and give themselves feedback. And then it would come back with the kind of the collective wisdom of the crowd.
you would sort of, I get it, that was really smart, Dan. So you would, that was very clever. So you basically said, hey, I'm gonna give you the information you need so you can play the role of all these different people, if you will, that have these different sort of knowledge sets and perspectives and priorities. And then I want you to like have an internal conversation with yourself effectively and then come back and tell me like what each person sort of said and yeah, fascinating.
Isn't that cool?
Andy Sack (48:47.384)
Dan on a scale of one to 10, 10 being totally exceeded your expectations. One completely missed your expectations. How has the AI transfer, how, how has AI transformation been for you at Influence Mobile?
I would say it's an eight trending towards a 10, right? Because you asked about like what the next big thing is. And I think once we really start understanding how to put AI and machine learning type concepts built into our products, there's still plenty of room for improvement by moving there. I think we're getting there. I've got developers excited who are already talking to, I'm telling them about future products and they're already on the side testing.
how they can train models and do different things. So there's, we've got the water boiling, right? And I think there's lots more to come. So I've been very happy with it, but I do think there's still a lot of upside to be captured. So.
In terms of advice you have for CEOs who have not started the transition, do you have any advice for them?
Yeah, I I would say there's very little downside to doing this. know, I could imagine maybe in a gigantic, gigantic company, there's some risks that I don't, that I can't anticipate, but I don't see there being too many negatives to getting people to start spending their time both personally and on the, in the work, work day using AI tools, just to start to understand it. What I tell my employees is that,
Dan Todd (50:22.19)
there's this conversation that AI is going to replace people. there could be some edge cases where that's true, but I actually 100 % believe that people who are good at using AI are going to replace people who are not good at using AI. That's 100 % true. So if you're a developer or you're a content creator or you're a customer supports person, using AI to make yourself 20, 30, 50%, 100 % more effective,
is probably one of the best things you can do for yourself from just a career trajectory. And then I think smart employers are going to find a way to leverage that and find a fair balance. Like what I've told employees is like, I'm not just trying to squeeze every piece of juice out of the lemon. Like if you can become 50 % more productive, maybe you're working 10 % less and getting paid 20 % more and everybody's ahead, right? Over the longer haul. And so I want employees to understand that.
you know, their investment in this is going to improve their, it's going to improve their day to day work. It's going to improve their personal life. It's going to improve all these different things. And I think when people trust you and get behind that, they will invest their time.
Dan, thank you for checking in. can't wait to continue the journey with you and Influence Mobile on your digital transfer, your AI digital transformation. Adam, do you have any like highlights that you want to summarize from what you've heard? You've not been on the journey, but you've, you've heard me talking about it. Would anything jump out at you?
Yeah, two things that jump out at me, first of all, when you're outside the forest for the trees on this journey that you guys are on, it starts with you, Dan, and your mindset. You guys have been in it, and it's kind of amazing what you've been learning and surprised by and experimenting with, but if you step back, it's you. If you hadn't have said, I'm gonna...
Adam Brotman (52:26.222)
push on this and I want, I think it can make our company better, but I need to do it in the right way and the right way for my company. Like that, we talk a lot about that, but to see it in action is key. Cause if you weren't sponsoring this and doing it yourself, if you personally weren't getting your hands dirty and learning yourself, then there's no way your organization was going to push through and, get to wherever you're going to go on this journey, which I'm positive is going to be a net positive for you and your people and your company.
So that's my first reflection. And my second one is just, it's, it's, it's a good, you're a good reminder of like the need to.
the need to try things and fail and experiment. like that, think a lot of people are afraid of getting going on this journey because they don't know what they don't know. And it's just kind of overwhelming to them. and I love how you just were like, well, let's just start playing with this and let's see what's possible. And, and you're even self-deprecating and humorous about like, I wasted my whole life on this shit 4.0 when I could have been using 03 and like, but that's, that's
What we, what Andy and I found is that that's, that is the journey for, you know, the people that are really at the edge of it. That it's, there's not like some magic, you know, classroom you can go to to like learn all this. I mean, it's just moving too fast and it's too weird and wild. So I feel like your journey is pretty emblematic of what I think a really successful bumpy learning journey is going to look like in this space.
you
Andy Sack (54:08.302)
Greetings Dan Todd. Welcome back to the AI First podcast for the third installment of the transformation of influence mobile and super excited that it's the third and last installment of our series. And for those of you that have not listened to the first two episodes, we've been tracking Darren Todd's work with forum three and his transitioning.
of Influence Mobile into an AI-first enterprise. And today we want to hear how it went, what his learnings were, what was good, what was bad. And then we also want to hear about his own personal journey about becoming what I would claim as an AI-first CEO. So Dan, why don't you give us the high level like overview of the approximately 90 day journey that you've been on with your team and talk.
Talk about any results, any challenges, just give the audience a summary.
Sure, yeah. You know, so AI has been on my radar since, you know, it came out and I was certainly dabbling with ChatTPT last year, but I think the focus that I put on after you and I chatted earlier this year was to make sure that every employee in my company could benefit from the day-to-day implementation of AI into their workplace. And I would say we got an A plus in that regard.
The stories of people from the most technical to the least technical. I people who were self-described non-techies are now doing SQL queries with the help of chat GPT developers are cutting their work time down by 50 to 70 % to get tasks out the door. we actually have AI enabled technologies coming, that are in the works right now where we're, we're doing very substantial data analysis and, analytical work that
Dan Todd (56:04.94)
then moves into tactical changes in our product based on AI analysis and help. And so I've been really happy with the transition. We've used a lot more tools, but at the same time, I've become keenly aware of the massive mistakes that AI can still implement into the workplace. And so, you know, I think many people I talk to are fearful that AI is going to replace jobs and, you know, certainly going to reduce jobs at some companies.
But more specifically, people that are really good at using AI as a brainstorming partner, as a workflow automation helper, those are the people that are going to replace jobs. And so I've really encouraged my employees, not only for the betterment of Influence Mobile, but for the betterment of their career to make sure that they jump in and are at the front end of this technology. You don't have to get very big of a lead to start knowing a lot more.
than everybody else and to be able to keep up with the cutting edge changes, which are coming very fast. And so I've greatly appreciated my time with you guys and we're definitely way better off because of our investment in this area.
You've not been on up close and personal to to this transformation as I have. You have anything you want to ask Dan or.
Yeah, Dan, what, from your perspective, I mean, we, remember on our last episode, you, talked a bit about some of your own personal revelations of like, my God, I hadn't really been using three as much as I should be. And, and I, you know, sort of your own journey, but I want to ask you as a manager, like you got your team to use this. You just gave some examples, but like what's been
Adam Brotman (57:54.794)
Give us an example or two of where, like how it feels for you. Have you noticed your organization, I'll call it moving a little faster or producing, I'll call it emails or reports to you that you've noticed are a little higher quality or didn't take as long to get generated. Like how have you, I want to ask you as a manager, how have you felt the improvement of your organization by the fact that they're using AI?
Yeah. Yeah. I mean, every day I get reports that come from leadership people who are doing massive depth analysis that they just wouldn't have had the time to do before. Right. They've, I think people feel more confident pulling into spirit data fields and putting them all together and helping them do summaries. Like I said, I from, the actual productivity side of things, you know, we, have a smaller team than we've had in the past and yet.
some of those teams have said they are equal or more productive with AI than they were with a larger team. Right? So the application of AI at that particular time this year has really been helpful, like I said, to, whether it's the BI team or the dev team, across the entire organization, everybody is using it. People who were once skeptical are using it. People who were scared of it are using it. And they're seeing it as a really valuable tool.
And yes, on a day-to-day basis, it makes me very happy to see the individual improvements, but I still think some of the biggest unlocks are 90 days away where after we've got this ball rolling in this momentum, we've got three or four big projects that are implementing versions of AI slash machine learning that I don't think we would have prioritized as much as we are now. And I can't wait to get those things out and live and impacting the business.
Yeah, day to day as a manager, I've seen a big impact in my own personal work. It's a great thought leader and it gives a level of depth to the work that I do, which is a lot of times product specing that I never had in the past. you know, I would do a one page Google writeup of what I wanted to kind of generally see. Now it's super detailed with a lot more analysis and I think easier for the product team to turn into a high level.
Dan Todd (01:00:20.856)
you know, actually to a detailed level spec.
Dan, if you were talking to your friends who are CEOs or members of our audience who are CEOs who have not embarked on the AI journey themselves or with their companies, what would you say to them?
I would say the significance of what AI can do for your business is no less significant than the changes that occurred from the internet itself, which is pretty monumental. I mean, like there were just fundamentally things that we couldn't do. Like my business couldn't be in existence without the internet. Right? So I told my employees during our training, even though the internet displaced a lot of jobs, it
created more jobs in general, right? And so AI right now is, set up to give massive fuel to the fire for anybody who invests in it. And so I think falling too far behind gives your competitors and your enemies and whoever's working against your business, a headstart that you don't want to give them.
You know, one other comment that I think I heard you make earlier that I want to pull on for a second, which is really interesting is how it, I think, maybe I misheard you, Dan, but I think you said, you know, it's still, one of the things you've learned is what it's good at and what it's not good at, and like where it's helpful, where it makes mistakes, which is a really important lesson. In fact, I was just talking to another group of people this morning and saying that, look, you know,
Adam Brotman (01:02:03.104)
Is this eventually going to replace all of our jobs? Probably. But I don't know when, because the truth is it, it, as it gets better and better, you almost need humans to sort of be discerning more and more. So we end up all becoming kind of managers of AI systems in some ways, which is, which is really weird and interesting. Right? So I wanted to kind of ask you about that topic. Like do you, when you
Have you had an example, you don't need to throw anybody under the bus, but have you had an example of like getting an email report from one of your people where you wondered if they had actually read it and understood it if they just generated it by AI?
I haven't seen that. you know, I think at different levels, it can happen to any of us. certainly happened to me where the work that I was doing was several levels of data science, proficiency above my, you know, my abilities. I went several days down paths where I made one sentence. I said one sentence wrong to AI, for example, on this humongous data set.
And it came back with a really smart answer. And I'm like, this is perfect. and I kept working on that for three or four days. And it took me almost to the end of that next week to realize that one sentence a week earlier had cut off 90 % of my data set. So, and so I wasted a whole week of time analyzing a subset of data because I didn't take the time to go, you just, know,
I just kept asking it meta questions about the data, but it had cut off basically a million rows out of one and a half million rows. And so I'm, I am fearful that that can happen all the time. And so I'm constantly telling people like the more significant the analysis that you're doing, the more you have to. For sure. Double check with other AI systems and then try to find experts who know what, what they're doing, because it is very easy to misinterpret, especially people right now who
Dan Todd (01:04:13.752)
They're doing, I mean, we have a lot of different data points that inherently a lot of our employees don't understand. So AI makes it easier to go down the wrong path. So it's probably, like I said, the biggest dark shadow over AI is you have to remain eternally vigilant or you will just, you'll send out something that's really wrong.
Right. Because that's interesting, isn't it Andy and Dan? Like that's like that topic is so interesting where it's like, on the one hand, we almost seem like we're so hyperbolic and hyped about like, my god, you got to use this all the time. And it's amazing. And yet, it almost seems counterintuitive that we're like, yeah, but you have to stay very vigilant on this stuff. Like it's at times it's amazing. And there's a temptation to
just let it do a lot of work for you, but you've got to check it because it makes mistakes. there's sometimes are not sometimes the mistakes of reasoning or analysis. Sometimes it's like you said, it'll it'll get something wrong just factually or data set got cut off. And so it's a weird counterintuitive thing that we could be so so enthusiastic about it, but at the same time, be so almost cautious about it at the same time. You have to they go together.
But that's not a normal pairing, if you know what mean.
Yeah. Well, I can say, think where, where it has benefited me is where it, becomes a thought, thought partner for me to refine my own thinking within the scope of knowledge that I'm already proficient at, right? Because I can then control its errors and I can say, no, you don't know what you're talking about. so, but it helps me get deeper and deeper into projects and I can constantly kind of like click it back into the thing where, where I wish it was perfect and where.
Dan Todd (01:06:06.858)
I run into the biggest errors is like when I start doing massive data set analysis and it's like suggesting random tree regression and things that I zero, I wouldn't even know how to validate, right? I would need a, a, a super smart data scientist, which we don't have on staff to validate those things. So in those areas, to your point, Adam is like super promising and I love that that potential is there, but I've seen, I've seen it make so many mistakes.
that I'm extra, extra cautious of the data analysis end results. It doesn't mean that it can explain to me the methodology and I can go try to do that methodology myself, but I'm very skeptical of massive data analysis that I don't really can't verify.
Yeah. Yeah.
I want to switch gears for a little bit because we have limited time with you. Talk about your own transformation and really like, you know, the frequency with which you yourself use it, use AI and really what does it mean from your perspective to be an AI first CEO?
Yeah. So the, the usage reporting that we were doing with you guys to kind of show increased usage. The first month was kind of shocking to me when I realized that I had used AI on average every 10 minutes for one straight month. And so that, that, that, you know, apart from sleep time. so what I realized was, you know, 2000 interactions spanned.
Dan Todd (01:07:48.546)
just a massive, know, across personal work related random questions. I never Google anything anymore. Everything's chatty PT, right? And so it's, it's become a, as significant of a work partner as my laptop, right? Like I use my laptop every day. I couldn't do my job without my laptop. It is, I could do my job without AI, but I couldn't do my job without AI.
as efficiently as I do today. And so that just becomes, like I said, it becomes a new tool that you have. You're still, you know, people still built houses when they came up with, you know, the, the air hammer, they just did it a lot faster. And so AI is for me become this very beneficial tool that really lets me work at very fast efficiency. And it actually also
lets you be better reflector on what you should be doing because you can always ask, TPT, am I doing the right things in my job? Right? Should I be reassessing if I do want to do this? And so I think that's one of the most useful aspects of AI is you can help. It will help you use it more efficiently. And so I think if people don't, you know, I tried to just jump right in and I made a lot of mistakes and some of them I told you guys, but
The more comfortable you are, once you start understanding how to change the models and even using a bunch of different versions, Grok versus Chat2Qt, initially I didn't really understand why anybody would even do that. Now I see the benefits of using those different things. So for me, it's just been a lot of experimentation and trying to figure out where to fit it into my life to make my life more efficient.
And how would you say it's changed your job?
Dan Todd (01:09:39.47)
I don't know that it's changed my job. It has allowed me more time for certain things. So, you know, if I was going to send out a large email to investors or my employees, I might spend an hour and a half proofreading that sucker. Right. And now I do not have to do that. I speak into chat TPT. It comes back. It usually doesn't have the perfect tone or exactly what I want, but
It takes what might've been 90 minutes down to 10 minutes. so like that, that doing that a lot of different times for starts to free up a lot more time. So then that allows me to reinvest that time into product innovation. And then like I said, that helps me be a better, you know, cause I'm it's able to like pull in all of this broad set of information for me to be considering as a product person, which has made that part better. so like, as it starts to replace low efficiency work,
It lets you spend more time in deep higher value work and then it helps you be more effective at that higher value work. So, you I don't think it's unrealistic to think that people are 2x or 3x, not 20%, but two times to 300 % more efficient because you're shifting your time from the redundant sort of administrative things, which it's very good at, into deeper level analysis if you're using it consistently.
Can you while we're on the quantification of you becoming 2X, 3X more efficient, more productive as CEO, can you talk about really the journey of the last 90 days and what you saw in terms of the growth in chats amongst the team and how you quantified that or translate that into productivity gain for the company?
Yeah. So, you know, thanks to your guys' prompting, we upgraded to the enterprise version of Chat CPT so we could track individual. We started to do it manually and that burned people out pretty darn quickly. But between Gemini and Chat CPT, we have a fairly consistent ability to show total usage. Then we started rewarding people for the best prompts and the best uses of models and pushing out
Dan Todd (01:11:55.586)
People who are doctors within our own company, what they were doing. And that kind of caught a fire with people on top of me constantly bringing it up. And yeah, I don't have it in front of me, but we have, we have seen 20 to 30 % week over week growth of AI usage for three straight months. I don't think we have a single employee who doesn't use AI. There's still some who don't use it as much as others for sure. But we've, we've also, you know, I'd say the top 10 % of people probably represent.
half of all the usage, so you got some really heavy users. But we're also making sure that the 80-20 rule works in our favor. We're trying to make sure that everybody, the big long tail of people, are all using it more consistently. And so we've just seen good results from the methodology that you guys helped us apply.
You had to quantify that and just in your own mind comparing to the organization of 90 days ago, you're what more productive.
you know, if I just had to pick a number, I'd say at least 50 % more productive, right? Because, know, not every single person is using it the same degree that I am, but I definitely know, like we have a lot of developers and a lot of the developers are using it. And I've had developers tell me that something that you, well, I had one specific person tell me they were able to do a 40 hour project in three hours. And another person told me they were able to do a 10 hour project.
in two hours. that's, you know, between 80 and, you know, whatever, 60 and 80 % reduction in total time. And not everybody's doing that for every single project, but I think, you know, there's some people who are probably 200 % more productive and some people that are 20. So I would definitely say 50 % is a conservative estimate.
Andy Sack (01:13:43.432)
Adam, you get the final question and then we'll summarize our perspective on what we've heard. Do have anything you want to ask Dan before we transition?
One of your breakthroughs that you said you were to get the whole company to sort of be a little bit more excited and more engaged in using AI to help them be better is that you said you did it, did you do a prompt contest or did you just do like office hours? What was the specific tactic you used?
So we created an AI knowledge share in Slack and we had a weekly contest for a free lunch across four different things. Total AI usage, best prompt, best personal use, and best multi-model use. and so.
Yeah, great. How did you do the grading against the rubric? Did you do it or do you use AI to do it or how did who decided?
I created a rubric from AI. published that. So everything was graded on a scale of one to 10 on five different factors. And then I would pump everything in. It would come back with the score. I'd largely just review that and say, does that feel right relative? There was a couple of times where two people did so awesome and the score was slightly less that I would just award it both.
Dan Todd (01:15:06.2)
But yeah, I tried to use AI for all the scoring and then I published it all. Then we use AI to create little fun emoji pictures of the winners and we post that in the Slack channel. And then, like I said, and then we just kept trying to bring a lot of exposure to it in the company meetings. so, you know, thankfully people, people had a willingness to, you know, jump in.
That's great, Dan.
Yeah, that is great. I mean, with that, mean, normally at this point, I turned to Adam and ask him for a summary, but I think I'm going to actually start with you, Dan, to give your own summary of the transformation. Like you say whatever you want to our audience in conclusion. Yeah.
I would say that, you know, your investment in AI should begin sooner than later. And if you don't have a clear vision of what that path should look like, get people like Forum3 to help you, right? So you don't want to invest unwisely in AI and waste your time and bring people down the wrong path. And people like Adam and Andy have been working with, you know, substantial companies and seeing who's doing it right and who's not doing it right.
And so if you've got your own vision and you can help have, you know, have AI give you all the momentum that you need. Great. Go do that. But if, if you're listening to this podcast and you haven't invested in your company in that way, you're probably losing out on a very big ROI. so find someone like Adam and Andy or find Adam and Andy and have them start helping your business become AI first.
Andy Sack (01:16:43.842)
Did not pay you to say that, but thank you.
I was not paid. I was not paid. Like I said, I'm very happy with the results and they speak for themselves.
Yeah, awesome. Adam, what's your, I mean, you followed the journey, heard the, you heard Dan's answers. What's your summary of
You know, my summary is that Dan's experience is wonderful and kind of fun because it's not easy to explain. you know, that's the theme I've... Dan, you've done a great job of explaining. You've explained examples of how you've gotten people to use it. You've explained specific instances of where productivity and quality has lifted. And yet, if someone were to be like, what...
how does this work? What's the formula? mean, even what you just said, Dan, is right. It's like, well, you have to personalize it to your company and your culture and your leadership style. And if you dedicate the, if you just trust the process, it'll work out. But that's, you know, there's a lot of stuff in life that's really important where you just have to trust the process, right? Like,
Adam Brotman (01:18:03.266)
Like that's a life lesson. And I think it applies to becoming an AI first CEO and AI first company, which is you got to do it. You got to trust the process. And if you want to know exactly what all the use cases are and in a nice little bottle and packaged up nicely, you're not going to get it. And yet every single company we've worked with and talked to the CEOs of, they have a story like yours. They're all slightly different, but they all end up by saying, we're transformed, we're different.
We took our own journey and we're better and we're more productive. And it's unlike any other kind of transformation you've ever been through. And so that's my summary and I'm so glad you did it and shared it with us, Dan.
I'm happy to do it.
I have three things to say. One is the story that Dan shared about AI is like his laptop and I couldn't do my job without my laptop. How could I do my job without AI? That feels right to me. And that like, think summarize, I mean, other times people have compared AI to electricity. And I think, I think both of those metaphors hold. So I, that jumped out at me. The second thing that jumped out at me was the story of Dan's lost week and
The downside of like one sentence where he inadvertently told AI to ignore 90 % of the data. But that one sentence, which was, it wasn't an AI error that he highlighted. It wasn't a hallucinization. It was a, it was an error in prompting, but how that cascaded into a last week. I've had lost hours. I don't know if I've had lost weeks, but that's a new phenomenon that you can go down and you don't even know that you're going down it and, how you.
Andy Sack (01:19:47.182)
catch that and error check that is a real phenomenon of AI work that didn't exist before. And I think that that's a real, it's worth highlighting that and calling that out. And then lastly, I would just say having been close to Dan's leadership of Influence Mobile's AI first transformation, we always say that AI first transformation start with AI first leaders and
And Dan led by example, Dan really was and is and became an AI first CEO, whether he set out to do it or not, there was curiosity and he led by example. and he was the top of the leaderboard. like his usage and then turn and passion for using it led him to the top of the leaderboard. And you know, guess what? The employees were like, this is important to my leader. I'm going to start using it. And, and it really.
It was amazing to be on the journey and watch Dan really embody the kind of leadership principles that we talk about and seeing it happen was nothing short of beautiful. those are my concluding comments. With that, I want to say thanks, Dan. Thanks for letting us be a part of your journey. Thanks for calling us out and recommending us to our audience without being paid.
And I look forward to watching the journey as it unfolds going forward. I want to thank you all for listening to AI First with Adam and Andy. For more resources on how to become AI First, can go to our website, forum3.com. Download case studies, research briefings, and executive summaries. We have lots of amazing material on the website for leaders looking to expand their knowledge about AI First transformations.
We truly believe you can't over invest in your AI learning. Onward.