Announcer (00:06.018)
Hey, hey, listen up. This is Prompt This, all right? The podcast for those business leaders who are just fed up, you know? Tired of all that AI hype? Well, guess what? You've come to the right place. Clint and Greg, those two sharp cookies, they cut right through all that noise. They're here to give you the real scoop analysis, playbooks, you name it. They're gonna show you how to use AI to launch fresh ideas, scale your biz, and for crying out loud, stop getting left in the dust. So.
Let's get this show on the road. Here are your hosts.
Hey everyone, welcome to Prompt This. I'm Clint.
And I'm Greg. My co-host Clint and I started this podcast to explore how business leaders are using AI in the real world. We found that our friends and colleagues aren't just talking about AI, they're building with it. So we decided to bring them on one by one to share their stories. And that's prompt this.
Hey, Greg, what did you get out of today's prompt this episode?
Greg (01:05.644)
I think I finally heard someone articulate an actionable place to start if you don't have AI in a business.
You know, part that really stood out for me was when our guest talked about how he's deployed AI on the shop floor at manufacturing companies. I thought that was a really cool new approach.
You know what else I liked about this episode is when he talked about training salespeople with avatars and other uses of the LLM within within within the organization.
Yeah, we did talk a bit about the future of work and a little bit of the science fiction aspect of AI towards the tail end of the episode, but the front part was chock full of practical, real advice on how to get started. Let's listen to what Jason has to say.
Clint (01:57.678)
Hey everyone, welcome to Prompt This. Today's guest is someone who's been in the trenches with startups, scale-ups, and enterprise teams as the president of FAY, a global customer relationship management and customer experience technology consulting firm. Jason Green is a go-to-market strategist, business growth advisor, and frankly, one of the smartest guys we know when it comes to operationalizing AI.
We got another smart guy on here. All right. All right. All right. Welcome, Jason.
Thank you for having me. Glad to be on the show.
All right, well, let's kick it off here. And you've worked in scrappy startups, big teams. What's the wildest AI pitch you've ever heard?
think the wildest AI pitch I've ever heard is related to dead people. people? Yeah.
Greg (02:52.122)
I want to hear this one.
zombies it gets even better. All right, but anyways back to it.
wanting to clone people who have been dead for a while, voice and their video to create live interactive avatar experiences with people who have been dead. And I would say the adjacent story to that is kind of this emerging idea of like, about older people? Do we preserve them in a way that...
Right?
That's like that TV show, Carbon Black. Yeah, or Black Mirror, exactly. I don't know, Greg. When you're gone, I'm ready to, I don't know if I want to bring you You're gone.
Jason Green (03:25.698)
or black air.
Greg (03:33.774)
I'm done.
Yeah, I remember seeing early on one of the first articles I saw about what AI was doing was bringing your relatives pictures kind of back to life. You know, they could blink, right? Look at you. And I thought that was the coolest thing. I ran over, show my wife and I'm like, look at this is the best as she just turns you go. That's creepy. I'm not ready for that. So just people's ideas on these things are just vastly different.
Yeah, I don't know. I don't know that I want to do that. think I feel like for me personally, I would lose control of who I am if this future AI version of me is what family is looking for. I don't know. I think it's a cool idea. I think it's going to probably have a place in other areas, but bringing back the dead, nah. Let's leave that in the zombie shows. What do you think, Jason?
I mean, I think it's a generational question. think it's creepy for me, but you know, the notion of privacy, which is important to, think, people our age, for younger people, they don't think of it the same way. So, you know, will my kids, will my grandkids think it's creepy? I don't know. Maybe not.
I have the feeling that when you've passed away, it's not your choice.
Clint (04:56.642)
There you go. That there's a point. don't really have
A whole lot of control over it. That's true. That's a new paragraph in a will. But anyways, let's move on to some business topics here. Yeah.
Yeah, let's jump into the business talk, Jason. let's say to kind of just get things rolling a little bit and maybe dig into what you're seeing through your company, helping companies, helping your clients figure out AI. When you're helping those companies implement AI, what's the most common challenge that trips them up?
The common challenge that we're seeing is organizations basically trying to understand where do they start. I think once people get beyond just what is this AI and they start to understand the possibility of it, it becomes overwhelming. It's like a fire hose. And so we just see a lot of companies, executives like, hey, where do we really start with this?
Do you find that they're listening like they're having you come in for advisement or are you walking into a room with maybe five leaders in there? And it's a, it's almost an argument of where it should be. And everyone has a different opinion. And then you're in there as a specialist trying to lead that.
Jason Green (06:09.006)
Yeah, that's a great question. think the only point of difference in a room has been, do you think this is like electricity, 1882, first homes are wired type of change up, or do you think it's 1994, the internet? so early on, maybe a year ago and prior to that, yet people are like, this could be hype, it's not even the internet. And now everyone we talk to on an executive team, it's...
94, 1882 or somewhere in the middle. And so because of that, we're seeing companies, executive teams really wanting the advisement, advisory, but also just the synthesis to help them even understand how to ask the right questions.
Let's unpack that a little bit. I'm kind of wrapping my head around that. So the 1882 electricity versus 1994 internet, I'm getting the impression you're saying electricity was this entirely new thing that just really changed everything, whereas the internet was just kind of the next evolution of computing. Is that what you're?
So we can go back to 94 prior to the internet and the internet was a big change up and it allowed for things like Uber or Airbnb or Netflix to come, these business models that simply couldn't exist without the internet. But I would still make a broad argument that our lives are still kind of similar to how it was before the internet in many ways. However, if you go back to something like electricity,
The difference between being before electricity and after electricity is significantly more profound.
Clint (07:48.92)
So you're kind of laying out is AI just the next step of computing or are we going to just completely change the way we live day in day out?
Right, is it a step in technology leap that has new business models or is it a societal leap? I you could even think of something like sleep. mean, humans supposedly prior to a couple hundred years ago, they slept in, a lot of them had two sleep cycles. You'd sleep for four hours, you'd get up and then you'd go back to sleep for four hours. That was normal. But then since the light bulb, we don't live like that anymore. You're expected to sleep, you know, seven, eight hours through the night. so what...
type of things like that would happen and that's where if you start thinking it's 1882 it gets into okay is this changing the nature of work what does this mean to be human what is things like death how does it all this type of stuff starts kind of
I've talked to all the, you know, our peers and business leaders. think everybody has this feeling like they're have to go forward with AI no matter what. And no matter what their, their feelings or, or thoughts are concerns, it's a must move forward. How do you get past like the debate part of coming into a new engagement and then you can probably see the path forward. How do you end the debate part? And then
start turning it into you know into business readiness.
Jason Green (09:15.842)
Yeah, mean, for us, what we tell organizations is it starts with the top. It starts with developing an AI strategy and having a posture and understanding how the org chart will be impacted, how the cultures and values, or the culture and values need to be potentially modulated within the organization. And then from there, it's a matter of usually jumping into an initial workshop and then oftentimes starting
with a fractional AI strategy offer to augment their team. Because the thing is, is someone has to lead it. And so if you don't have someone to lead it, that's falling on the CEO, the COO, sometimes the CFO, we've seen that, or the COO. they're not, I mean, they don't have the time to do that. And so they need someone to help. And so we kind of come in with a turnkey set of offerings that...
Here's an easy button, let's start.
Are you finding the head of IT is becoming that lead AI person or is it maybe a head of a business function that's becoming that lead AI?
So far, it's a business function because it's really affecting the nature of work itself. Now, what's interesting is when we come into an organization that has already been doing AI at a technical level, like machine learning, predictive analytics, now into the generative. So there's some companies that have had an insurance space, for example, where they have technical, they've been doing AI.
Jason Green (10:51.052)
maybe in their own product that they have in the market or back office things as it relates to technology and IT, but this is like a whole different thing. That's what we're seeing.
So what's the point of view that you're bringing into companies? Is it electricity or is it the internet?
Point of view, what I say is I go, look, it's at least the internet from my perspective. And if it's anything even close to 1882, you have no choice, I mean, you have no choice but to do something. And so I think conservatively it's the internet and we kind of start with that. And even with that, that requires some pretty big shifts of thinking, education, et cetera.
Let's get practical for the audience here. I'm sure people want to know what are those most common things that you're solving with AI right now.
Yeah, so any type of manual entry where people are reading off of a PDF or other live document, having to take that data and put it into a system. So that's a big area to take away that manual entry. If not in the totality, at least to remove 99 % of the manual work done there. So I'm thinking
Clint (12:06.658)
being invoices and contracts and things like that. Is that the type of
More like stuff that's like intake forms. Things that are filled out. Yeah, things that are filled out by hand. Those type of things. A lot of business still has a lot of forms. Another big area is just around the whole quoting and estimation process, both in like manufacturing and construction. Oftentimes as a sanity check against what the estimator internally is saying.
sanity check against what subcontractors are saying it's gonna take and then making sure you're actually including everything you're supposed to and so the ability to kind of sift through building codes and compliance documents, state legislative laws and all this type of commercial codes to make sure that if you're quoting to do something, did you remember all the things A, you're required, okay, to have when you're doing it to pass some sort of code.
And another thing is B, have we included all the things we could include that are reasonable to include that a client should have that we just happen to forget little bits. And so that's kind of a really interesting set.
So when you, when you say it like, manual forms, that's a great way to kind of digest where, you know, as a first step, where this could go, you know, I think early on in my career, I was in document management and you know, how did we scope a customer? Where's the paper, you know, it kind of reminds me of that when you say the forms and now you're, I heard you say it was like, I heard finance, I heard legal, I heard HR, I heard sales quoting.
Greg (13:53.044)
I heard purchasing and they're getting, you know, so now I'm starting to see, yeah, you're probably, you can walk in and see where it can fit everywhere. So you're having to see what a client can bite off in the first round and where they're going to save the most money to be excited about this. That's what it sounds like to me.
Yeah, because you want to get people excited. You want to get momentum in the organization with these type of projects. You know, the second part to it is you want to get everybody trained organizationally because there is a lot of things that aren't necessarily a project or initiative that you would do. And then there's some rollout per se or technology around it. A lot of it has to come from the ground up. And the individual worker
has to learn and understand what AI could do so that they can augment little bits of their job on a regular basis.
Coming back to those different use cases that you're describing, that you're helping companies solve, lot of them being kind of starting with what's paper manual. Are you finding off the shelf software solutions that you just need to configure and deploy? Or are you building stuff from scratch for people? Is it a customization project?
Yeah, that's a great question. In general, it's a combination of existing technology tools and then using our skills and knowledge to integrate and build custom application frameworks to support it. I mean, you have all the LLMs as a baseline. That's going to be dry.
Clint (15:33.934)
GPT and Claude and that sort of thing.
GEMINI, you know, you could get with the llama, you have perplexity. So you have some of these LLMs that you're going to use as a baseline. We have a product that we built that takes those all and makes them accessible in one interface. And then you can have different agents you build that are using the different LLMs and you can build out.
because I was going to ask you, you have a preferred LLM that you want, a large language model that you go with?
Yeah, and that kind of, origin of that is that we've seen and experienced that different types of tasks or asks of the AI produce different results. And so one LLM might work better, Claude works better for some than GPT and vice versa. And then the other thing is because the LLMs, they keep coming out with new models, it's...
It's really good to be prepared. So maybe today you have an agent created for some manual entry and you're using Claude, right? And then GPT has a new model that comes out in a month and you might want to test that. Like what if that model is now better? And now if you're kind of doing this in a framework where you can switch the models out without really having to change much else, it kind of sets you up pretty well to take advantage of the latest AI.
Greg (17:00.846)
That's interesting. You know, if we look, take a look at like your own stack of tools and frameworks, and obviously you're very knowledgeable and been working with a lot of them. What's your favorite? What's your favorite one right now? And which one do you think is made like overhyped, you know, you know, as you're digging around and playing with these.
I mean, we really like, there's kind of two different arenas. I mean, we really like HeyGen and Eleven Labs for audio and video cloning. I really like the kind of output you're getting there and then its ability to take some of the content and recreate blogs and other multimodal outputs and then having live interactive avatars. We think that's super cool. There's a lot of really interesting kind of open source or infrastructure components.
such as OpenRouter, which lets you route through a bunch of different models in a very efficient way. There's other tools like...
I to route through different models.
Yeah, I want to learn more about that.
Jason Green (17:58.24)
Yeah, so like there's a lot of models out there beyond, there's a lot of LLMs and models out there beyond just the base ones and there's new ones being created and there's like libraries of these different models that some of them are like really specific and very specific kinds of analysis or research or math.
So like beyond chat, GBT, Gemini, Grok, the big names you hear, there's specialty.
specialty type of little models and then like they kind of show like here's your result for this type of thing and so open router lets you and we have that connected into our framework and we can then route to these different models and some of these extra specialized ones in a fairly easy way.
Give us a real world example of where you've taken something that may, the LLM that just couldn't get the whole job done. Now you've put this together. What does that look like?
Where would you Yes, so I mean, you might have, well, an example would be if you're trying to take schematics, like manufacturing schematics, and you need to produce a quote, like your tool in die shop. It's great use case. Schematics drawn for pieces of equipment from the 1950s, hand drawn. And so one, we tested out both Claude and GPT. We liked Claude better. And then we had to train it. We had to then basically,
Jason Green (19:19.278)
provide a bunch of schematics and we told it what it was so it kind of understood it. We trained it on saying, what's the quotation for this? Oh, you forgot this or you forgot that or don't forget this. And then we also kind of laid out what the shop floor was. This is the kind of lathes we have. Here's the machine or here's like what machines we have. And it was amazing how like even just the bass Claude is able to understand.
a machine shop at like a police
Just blown you away. That must have just been great.
It's just completely crazy.
Because everything I hear about seems to, at least in my world, focus around strategic planning or marketing and using LLMs for that. And here you're describing something very practical for a nitty gritty industry.
Jason Green (20:09.068)
Yeah, and it's like, vertically, you have horizontal lathe, and it understands the difference. So we did that. So we trained it up, and we built out a knowledge graph, which is kind this vector database that sits on top of the LLM. So when you're querying it, it's using that as part of its augmented retrieval.
Does that, do have any other underrated examples of where people aren't thinking about how to apply AI, but they can get real value out of I like that one you just gave. Is there any more like that?
Yeah, mean, one example I like that I think is pretty interesting is for sales enablement and like objection handling and training. The ability to create interactive avatars to do different types of objection handling with like a new salesperson I think is a really interesting use case. I think translation is interesting. So a lot of businesses, you have a lot of people speaking a lot of different languages and you might...
need to speak, you might need to do something in Spanish or Romanian to the workers and you need a way to quickly translate. So the AI is able to translate. Yeah. I think it's pretty cool.
Do Translate.
Greg (21:25.384)
voice voice voice handling of the objection training
Better watch out, Greg, the AI's gunning for your job there.
They can have that one. Cause I've always found that to be the most time consuming piece of training. A large sales organization was, the role playing, you know, usually I have like, you know, six to eight management heads for 90 people. That's a lot of role playing.
And you're taking them off the front lines if you're doing all that back end training.
man, let the AI take that every day. I want to get back into it now.
Jason Green (22:02.894)
Yeah, I mean, you got all that kinds of stuff. then, you you mentioned on the strategy side, and so I think we've seen this with some of our clients, there's, want to develop a marketing strategy, being able to get initial swings at like, what would a broad strategy by a top, you know, GSI look like, a top consulting firm look like. And even if that's not then used as the strategy, I think for a lot of,
Companies it would really help them kind of broaden and make sure hey, okay. These are the questions We would need to ask when we're in this type of strategy
can't even think of one other thing. I know in some of the smaller companies I've worked in, the management has usually been around for a long time, but the workers coming up, even the knowledge workers coming up are pretty new in these positions. And you hear the management say, bring me a strategy, bring me a plan, bring me the quarterly plan. just the fact that you could have it pump out something that looks like a plan so that you know,
what to bring. can't tell you how many times I've seen the junior person just go walk out going a plan and yeah, yeah, the output that comes back. You're like, wow, that's not a plan. Let's sit down for a while. Imagine AI taking that training. That would be to the to the new person. That would be amazing.
Give me an outline of what that plan might look like.
Clint (23:34.956)
So let's shift gears, let's pop it up a level. Let's talk big thoughts here about where AI and society... What?
I've heard off the shelf FOMO pop it up a level. We got about 1994. got a modernize your vocabulary.
No, no, no, no, dude. These are old, well-tested idioms here. I can roll these out all day long. OK, thank you. All right, so let's talk where this is all going. And one of the things that you and I have chatted about once or twice in the past is your vision, or at least your questions, of how's a company structured in the future?
Old. Yeah.
Clint (24:23.328)
If you've got this super intelligent AI assistant sitting at the elbow of everybody, what does that mean to how a company's actually organized?
Yeah, it's a great question. It's a topic that comes up. Like what does the org chart of the future look like? I mean, right now, at least up till now, you have an org chart that has boxes of roles and they're filled with humans. And so if we think about the future, is the org chart a place where you have boxes that are human filled and boxes that are agent filled or AI filled? And then that starts to bring up interesting kind of
questions around the architecture of an organization. And so one example would be, you know, you'd have a financial analyst and that would be a role. Well, where does that sit? Okay. So a lot of times that could sit under the CFO or within that, within that group. But like, if you can make an agent that's a financial analyst and you can make agents that are business analysts and psychologists and all these types of things. Do you now have these boxes that sit underneath?
your account executive, does your account executive who's an IC suddenly have boxes underneath there of these agents? And even if those agents aren't reporting there, the notion that you could have digital agent, digital roles and human roles, well obviously is a new idea. I think it's new, or at least it's becoming real or.
Well, if I think about it from a sales standpoint, you know, an org chart, you know, if you have a manager and you've got eight to 15 reps under a manager, maybe now it does look a little bit different, you know, maybe it's one rep and then how many agents
Greg (26:22.69)
Are we putting under this rep to execute across one territory? And now maybe instead of eight people across the West, it's the one West rep with the, with the eight agents under it. And you'd still have to, you still have to, or chart out the cost per, per agent, I'd imagine. So I'd probably be charged for eight of them, you know, and one human as opposed to 15 humans.
I could see that. it make sense to have AI, all agents and everything, just be a service center over here? And then we all go get all we need out of these people.
think it depends just how evolved this AI becomes. So you're already hearing about things where there's AI girlfriend, AI boyfriend. Okay, well, that's something. So you're starting, people are using, know, there's a therapist, AI therapist. And so I think the difference versus the automation or software of the past or technology is what if you can't tell the difference that it's not human?
thing.
Jason Green (27:32.972)
Right, if you're on a Zoom or a call and you cannot tell it's not human, you might know it's not human because it's your digital agent, what does that mean? And so that opens a whole new set of possibilities and perils. So that's why I think it's bit different, right? The interface between human and machine was really hard 50, 60, 70 years ago. Punch cards and okay, then we had a UI and now the UIs have become so much better and you can...
talk to the UI and talk to your car and all this stuff. But now if the interface just becomes, I don't know, something different, what happens?
Greg (28:19.47)
Hey Clint, what did you think about the AI voice intro?
You mean at the very beginning of the podcast year where the guy with a different accent? didn't quite follow that accent. introduced New York accent this time. Wait a second. Isn't our guest Jason from Chicago?
New York
Greg (28:40.097)
yeah, guess those guys, those guys don't go together, do they? Well, I'll rethink it. I'll rethink it. Next week. I got another one for you.
All right, I like how we're doing a different one each week. It really just kind of shows the full range of what you can do with AI voices these days.
Okay, now it's time for this week's AI challenge. This is where we challenge the audience to get their hands on AI and start using it so they get comfortable bringing it into their business environment.
Today we heard from Jason about how you can use AI for variety of things, such as a salesperson, learning how to be better at answering objections that you might hear. We also heard from Jason about how he envisions the user interface to computers totally changing in the future with voice interaction with AI. So we're going to explore both of those today. We're going to go back in and learn more about a sales custom GPT.
and we're going to do objection handling and you're going to actually do it by talking to the computer as opposed to typing anything out. That sounds like a pretty cool and different AI challenge, right, Greg?
Greg (29:46.07)
yeah. So take a look down into the show notes. You'll find a link to the blog article with all the instructions and everything you need to complete the challenge. Once you're finished, share your experience in the blog comments and we'll service the best experiences and highlight them on the next podcast episode.
Greg (30:09.954)
Well, Jason, I got to tell you that was a great session and we appreciate you coming on the podcast. And for those folks that want to follow you and learn more about what you do and continue the conversation, where can they find you?
They can go to our website, fadigital.com and learn more. Perfect. Awesome. Thanks for having me. I enjoyed the show.
We appreciate your time. You have a great day.
Hey, hey, thanks for tuning in with Clint and Greg today, alright? You wanna catch all the Prompt This episodes and dive deep into some articles? You gotta head over to www.promptthis.ai. Trust me, it's worth it. And hey, don't forget to smash that follow button down below, yeah? We can't wait to see you back here, alright?
