From the Archive: Building an AI-First Marketing Team

0:00Chris Marin: So AI becomes a way to get to market faster, better, more cost-effectively than anybody else. And that's how we think about it. We don't put any constraints on it, to be honest.

0:18Nick Panayi: Welcome to the award-winning Health Tech Marketing Show. I'm Adam Cherinas, and I'm actually away this week taking a little bit of a break. So we're pulling something special from the vault. Before we get into it, thank you for listening. And if you're not following the show already, you can follow it on any of your favorite podcast application. And we're also on YouTube. So please do subscribe. This episode also comes to you thanks to our friends at Healthcare Now Radio, the 24-7 healthcare-only radio network that syndicates this show and many other healthcare shows all around the world, all around the clock.

0:55Nick Panayi: Big thank you to Roberta and Carol for making that happen. Today's From the Vault Pick is my conversation with Nick Panai from about a year ago when he was the CMO at Inovolom. And when we did this, Nick walked me through how he built, how he and his team built an AI-first marketing organization from the ground up. and how he got buy-in from legal, from compliance, how he addressed his team's fears about AI taking their jobs, and how his team started using AI-generated podcasts to turn dense research into some content people would actually consume. So since we recorded this, Nick has actually moved on from Inovilon and is getting ready to get started on something new and exciting.

1:50Nick Panayi: So keep an eye out for that. But the playbook he laid out here, an AI-first mindset, hiring for AI fluency and treating experimentation as a leadership mandate is arguably even more relevant today than when we first aired it about a year ago. So here's that conversation with Nick Panai.

2:12Chris Marin: Well, Nick, welcome to the show.

2:15Nick Panayi: I'm so delighted you're here and wow, we got a lot to talk about. Let's get started. Tell us a little bit about yourself and your journey.

2:24Chris Marin: Thank you, Adam. First of all, thank you so much for having me. It's a pleasure to be with you. So my name is Nick Panay. I am the CMO here at Innovolon. I've been at Innovolon for two and a half years. I've been in marketing much longer than I care to admit on this podcast. So it's quite a long time. I've been in technology marketing my entire career. And my last stop before Innovolon was actually for a conversation, was working for a conversational AI company. So I was the CMO there. And that was my initial kind of introduction to AI. And of course, I brought that passion to Innovolon. And Innovolon, of course, is in the healthcare space.

3:04Chris Marin: We're a data and solutions company that serves the entire healthcare ecosystem. And our focus is to improve healthcare outcomes and economics across the ecosystem.

3:17Nick Panayi: You're a pretty large company. Give us a sense, maybe just in terms of the size of your marketing team, how many sharks are on the team?

3:25Chris Marin: Yeah, the marketing team is a good size. It's about 95-ish people. The company is approximately 2 ,000 employees.

3:34Nick Panayi: That's a big team. That's great. That's really...

3:37Chris Marin: It's a size team. I'm not complaining. I'm not complaining. Good for you.

3:41Nick Panayi: That's great. So let's get into talking about AI. In the study that we completed in the last month or so with the health tech marketing community, we had, I think, 110, 120 respondents completed the survey. So it really is, you know, it's Nick, your peers across the industry. and the thing that is remarkable i think is that seven maybe seven percent are anywhere near the level of ai adoption and the way that you've got it integrated strategically and so we're going to go deep today in understanding your approach but yeah let's go back to the start how did you get this started what did that look like why and how did you get it moving so aggressively

4:28Chris Marin: Yeah. Thank you. It's been a fast journey.

4:32Nick Panayi: I will tell you, based on what I've experienced in my last company, I already had a lot of passion for AI coming in to this position.

4:42Chris Marin: So there was no question in my mind, AI is here to stay, and it's here to impact pretty much every profession, including marketing. So I was always kind of AI curious and AI expert.

4:54Nick Panayi: I am not, but they're definitely an AI lover.

4:58Chris Marin: and as soon as I landed one of the first things I did and it was a fortunate thing is I brought with me a person to run my digital marketing function that has worked with me before in two companies ago, DXC Technology that person's name is Chris Moran and Chris is my guy you know how we always say I got a guy

5:19Nick Panayi: yeah, you got a guy it's in New Jersey, what's it?

5:23Chris Marin: yeah, he's my guy And Chris happens to be one of those rare individuals that brings, obviously, digital marketing expertise. That's what he was hired for. But he's also deeply passionate and very much an expert in AI across the board. That's just a passion of his. And AI, specifically, AI uses in marketing. So I saw that as a plus. That wasn't the reason to bring him in. The reason to bring him in was digital marketing expertise. But, you know, he brought along AI expertise and that really helped me go from, wouldn't it be nice to use AI to sitting down with Chris and saying, hey, think AI first.

6:04Chris Marin: Anything that we do, you know, what can AI do for us before we even go from glimmer in the eye to execution? You should be thinking AI first. And when was this? Last year, two years ago? So you joined me about two years ago. The reality is the world of AI has accelerated dramatically in the last two years. So it wasn't the first conversation we had. It was one of the first conversations, but it was definitely, hey, let's keep an eye out. Let's have an AI first mentality. We have gained enough credibility in the company to be given the flexibility to experiment. Yeah. So thankfully, I'm in a company where the leadership team, my CEO, trusts us and trusts our judgment and knows that we're going to do whatever we can do and gain a competitive advantage.

6:58Chris Marin: If that means using AI, then so be it. So we do have the freedom to operate and the freedom to experiment, the freedom to make mistakes. And honestly, that's all we need is we just need the license to go at it and go at it fast.

7:14Nick Panayi: So explain where it fits within your marketing strategy. So just to maybe unpack your marketing strategy a little bit to start with.

7:23Chris Marin: Yeah, look, the way we organize is probably not atypical. We have a shared services model. So the way we go to market is in the healthcare space. So our product marketing function serves those segments by putting out a marketing plan. And the shared services model, including digital marketing, content, creative, ARPR, those are the shared services that we use to, in essence, activate a marketing plan in the marketplace. So from a strategy standpoint, we bring all of those tools and capabilities to bear to activate the marketing plan as effectively as we can in the marketplace. So AI becomes a way to get to market faster, better, more cost-effectively than anybody else.

8:15Chris Marin: And that's how we think about it. We don't put any constraints on it, to be honest. We don't think about, hey, let's only consider AI in these use cases or in these particular functions. We have the ability to experiment across all functions and across all tactics, across all approaches to marketing. That's the benefit of bringing in an in-house AI expert that knows the marketing organization well and knows what we can best use AI and at least knows how to prioritize some of those investments.

8:49Nick Panayi: That makes perfect sense. So overall, it's an accelerator and a way of increasing efficiency. What have been some of the barriers in this journey over the last couple of years?

9:00Chris Marin: I would say barriers, but a couple of speed bumps. One would be our legal organization, as you would expect. And I totally respect that, especially in a healthcare space. Anytime you think about AI, you start thinking about privacy and what information are you using? How are you modeling things? What are you doing with it? But thankfully, in the marketing space, unlike if you are developing a product for healthcare where you're using PII, we're not using anywhere near PII. We're using tools and data sets that don't really affect privacy in any way. So we're able to put those kind of, I don't want to say fears, but concerns to rest early.

9:44Chris Marin: Of course, we're able to say, here's the sandbox that we will play in, and we will stay in this sandbox. And it's all commercially available, widely available information. We just use it in a smart way. So that was the first, speak, Bombon would say. The second would be even helping our own marketers understand how AI can be used without necessarily affecting their value and their worth to the organization. It's a human reaction. if you tell a content producer hey i'm going to bring in a tool that will help with content creation the immediate feeling is one of hesitation of course it's like well what do you mean that's kind of my job right so you know are you looking to replace me and all of those things are human emotions that everybody understands every every time ai gets involved in a function or in a profession, you're always going to have the what about me kind of question.

10:44Chris Marin: Again, we were able to put those to rest very easily because we're not looking at AI as a way of replacing humans. We're looking at AI as a way of making our humans better. One of the best things I've ever heard explaining that relation between AI and human was a saying that went something like this, AI will not take the job of humans, but humans with AI will take the job of humans without AI, right? And that makes total sense. It's about AI being a companion and a partner in what you do. So we very much have a human in the middle and we very much value the human contribution, but it's just simply done better, faster, more efficiently with AI.

11:33Nick Panayi: That analogy makes perfect sense. And I can understand how that would go some way towards assuaging fears because everybody feels sort of a bit, you know, a bit insecure about the impact. Tell me about the roadmap. Have you got a, you know, we're going to get into the details of what you've done in a second, but I'm curious as like, you know, do you have a, do you treat it like a product roadmap or how have you approached this? We do.

11:58Chris Marin: We have the benefit of having a Chris, having a digital expert whose knowledge is very much up to speed and up to the minute and his knowledge of marketing as well. So it brings us a lot of these ideas so far, the beginning of it. We look at them as a leadership team and we decide which ones we're going to go pilot. They all started as a pilot. Unless it's a horrible idea, to be honest, Adam, it's like the time and energy that we're putting in a pilot is far outweighed by the potential benefit. So, you know, there were very few ideas brought forward that would say, ah, that just, you know, pass.

12:38Nick Panayi: Very few ideas. I can't even remember.

12:40Chris Marin: I remember a couple of discussions, but I can't remember what they were exactly. So the roadmap was our digital AI expert bringing in some ideas. We're looking at them at the leadership level. And he's putting them on a roadmap, balancing, of course, what we call keeping the lights on bandwidth, right? That he and his team have to have. We seem to have a website that's running. And the marketing automation platforms and email platforms and all that. So we don't want any of those to be kind of sidewiped for this. but we also want to invest in AI in a very significant way. So that's the only gating factor is do we have the bandwidth to do one more pilot?

13:24Chris Marin: That's so far been the case. Since we've been doing this and we've gotten everybody kind of on board and a lot of people excited, we've now formed an AI tiger team that has a person from each function. As a matter of fact, we're just putting this together. They're going to be meeting on a regular basis and basically giving the various functions an opportunity to go out and scout for themselves, bring in ideas, have those discussions, and then start to formulate a more formal roadmap based on all those inputs. Because the whole idea is that you don't want to have your innovation gated by one person as smart as he may be or she may be.

14:05Chris Marin: You want the rest of the marketing team thinking in exactly the same way. So we're looking to institutionalize that now.

14:13Nick Panayi: That's interesting. So essentially, it's start with one expert who's got a passion, who's going to drive this thing forward. And then as you scale it, then essentially clone his capability, or at least sort of get that capability replicated in the other departments, and then spread it that way. So it almost creates a network effect.

14:35Chris Marin: Yeah, for sure. And maybe less so capability, more so scouting and bringing in new potential tools. I see. I don't think anyone, I'm not going to be looking for somebody in product marketing to become an expert in AI, but I'm going to be looking at somebody in product marketing to let us know if there's an AI tool that's specific to the use case of product marketing should be something we should take a look at.

15:02Nick Panayi: Yeah.

15:03Chris Marin: is this one person i want to be able to fan out and there's more tools than hours on the day in the day for us to be able to scan everything so it's about somebody who knows the function and knows the company and go say you know what given our current posture this is something we should

15:21Nick Panayi: be looking at let's dive in a little deeper on this can you kind of go through some of the kind of killer use cases where the adoption is accelerated and you're really seeing an impact.

15:33Chris Marin: Yes. So let's start with content, right? That was the first area that was a natural area for us to look at. You think about generative AI, and again, you think about the fact that content, especially thought leadership for us, thought leadership content is probably the most important lead magnet, let's call it, right?

15:55Nick Panayi: That's what people are coming to us for is they basically

15:57Chris Marin: want to get smarter about some of these areas that we serve. So we have great thought leaders in the company. We have a great content team. We have the content team interviewing thought leaders and writing pieces, whether those pieces are white papers or producing webinars and things of that sort. Now, when the writing begins, we wanted some assistance for this content producer. So we put together a tool that is an AI-driven tool that takes in the persona that we serve. It's about 100 or so persona. The brand voice that we have, the brand positioning, and also our approach toward content development, which we adopted a model called Make It Punchy by Emma Stratton.

16:46Nick Panayi: Ah, okay. I'm not familiar with that. I saw something actually on, I think it might have been a LinkedIn post or one of your colleagues that said, I'm not familiar with that.

16:56Chris Marin: Yeah, it's basically, look, in B2B marketing, especially in technology, this is sea of sameness out there, right? And the same terminologies and acronyms and, you know, like best practices and best tools and all of these phrases that we've used for years and years and years. And it's very internal versus market and customer. So making punch is an approach that thinks about what the customer really wants to hear, what kind of solutions they're looking for, but also acknowledging the fact that they're human beings. And human beings don't respond to phrases like we've been using that are written really more internally than anything else.

17:41Chris Marin: So making punch is about just get to the point. Tell me in regular English what it is that you are trying to tell me. And then you can back it up with fancy phrases after that, but get my attention by telling me what the real deal is all about. So obviously I'm not doing it justice, but that's the general approach. But that is a writing style. There's some rules to that. There's some do's and don'ts. So what we did is we fed, we trained the system based on that content philosophy. and our brand voice and positioning and messaging, and the 100th persona that we have. And we basically created a rubrics cube that says, you're going to fit in your content.

18:23Chris Marin: It's going to rate it based on those variables. And then it's going to tell you what you need to change to get it to be 100 out of 100. So the writer starts first with the first draft. AI tells it, okay, given this persona and what he or she likes to hear and make it punchy and your brand voice these are the things that you may want to change to it so it actually gamified content creation yeah with the whole idea of making every writer your best writer very interesting that's the idea right yeah yeah in a

19:01Nick Panayi: lot of the content that we create we might actually start with a persona and possibly with a really detailed prompt, get the AI, particularly for something like Claude, to write the first draft. And then the human then intervenes with it. We've then got some GPTs that we've created that optimize it for things like search, particularly EAT for the AI answers. But the notion of actually taking a content creation framework and applying that in the process is really smart. In something that you shared with me, it sounds like you're doing quite a lot around personalization as well. Can you share a little bit more about that?

19:44Chris Marin: Yeah, sure. So again, you think about the jobs that AI can do better than humans, and there's many of them. But the one thing that they all have in common, those jobs, is that they're long and tedious, and the volume of things that needs to be done is just simply beyond human capability, or it would take many, many, you know, man months to produce. So how do we take something like that and work it with AI? One of those things, and this is as long as I've been around in marketing, is the ability to take every one of your pages on your website and tag them with tags that are useful for you and further downstream as you're working marketing campaigns and marketing tactics and you're trying to drive traffic to that page.

20:36Chris Marin: So we've in essence taken all 1 ,000 or so pages on our website, created an AI tool that basically scans the content and tags it with the funnel stage, the business unit or the segment that is trying to appeal to, the persona that would find this relevant, what kind of content this is, and other predefined topics and freeform tags that our team kind of came up with. And basically every page is now tagged with this. And again, imagine a human being going through every one of those thousand pages into it.

21:14Nick Panayi: It's just probably not a good use of time.

21:16Chris Marin: Yeah, it would happen at the end of my tenure, but it wouldn't.

21:20Nick Panayi: It wouldn't. And it wouldn't be done well.

21:22Chris Marin: Yeah, it wouldn't be done well. It wouldn't be done consistently using a consistent set of tools. Again, you can maybe brute force it by giving it to some outsourcer and say, hey, put 50 people on this. But then there's 50 slightly different interpretations of this. And how do you figure out the persona and all of that without AI? So, well, why even do that, right? Because once you do that, then you feed your marketing automation system with that information. so that when your demand generation function or people who are building campaigns are building a campaign, you can recommend what pages would be relevant for a particular campaign or a campaign can basically drive traffic to five pages.

22:08Chris Marin: You will know what five pages to go to because you want to optimize for a particular persona or a content type or something like that. And very importantly, you analyze your web traffic later. you try to figure out not only what pages people went to, but again, you look at things like what personas were attracted to your website and why. And you have a good match between what you thought the right persona was going to come to or not, and you make adjustments to that. So it's going to be the gift that keeps on giving. Again, this is not something that would be possible without AI.

22:40Nick Panayi: Yeah, I can see a million different use cases for that. I mean, for applications of that, For example, you could create personalized email sequences for different personas based on those tags. That's really smart. I've not heard that one. That's great. You also, I believe you guys are using agents to create ads. Can you talk a little bit about that?

23:03Chris Marin: Sure. So that was also one of those cases where you try to find areas where AI could really help. The art and science of pay-per-click advertising is one of those areas that until now, you kind of have to outsource that to an agency, right? I mean, there's so much there to go through before you determine what to write in that ad block if you're doing it right. So we thought that that was a perfect opportunity for AI, and not just AI in general, but agentic AI. So that was one of those first experiments of agentic AI for us. And I can tell you, it makes all the sense in the world. We're already starting to see kind of the difference between the before AI and after AI.

23:57Chris Marin: And we're now looking to scale it. We also want to have a human in the middle as well to make sure that we're doing it right. But just to kind of summarize, you have, in essence, multiple agents, which the whole idea is almost kind of a specialization of labor, right? You take one agent who is responsible to scan the whole marketplace for what competitors are saying. and bring that back. And you have another agent that's responsible to take that content engine we just talked about and apply that content engine to do a first draft of an ad. And you have another agent that basically critiques that ad.

24:37Chris Marin: And you have yet another agent coming in and basically figuring out what to bid for that ad, given the competitive environment. And then it goes on and on. It's like multiple steps of putting these ads together. And at the end, we want to have a human, of course, look at that and say, okay, are we sure here we're not going off the rails? Make sure that, again, we stand within our brand voice and we stand within our budget parameters and all those good things. And actually, the human starts with some initial content or copy that they write. And then the agents take over and then the human at the end does the final check and we go.

25:19Chris Marin: And that is a perfect example because, again, it needs constant refinement. The minute you put it in the marketplace, something can change, you know, either in the competitive landscape or some keywords can come in or out of favor. Something happens in government policy that impacts certain things and you want to be able to take that into account, especially in health care. So all of those things are good, strong reasons why a model like that could work for SEM because of how quick it turns and how reliant it is on science as well. It's not just, you know, you're creating an out-of-home advertisement is one thing.

26:04Chris Marin: This is about search engineering and now with AI as well, coming into the mix of search. that's making things even more complicated and more interesting oh yeah yeah i know so you're

26:16Nick Panayi: using it for paid search so and that i mean that makes perfect sense yeah google ads are you able to do that with you know maybe more brand ads that are sort of upper funnel they might be programmatic ads and display maybe linkedin ads are you using it with those where there's a little bit more where you might want the human to be much more involved i don't i'm not sure how you

26:40Chris Marin: how it's applying that we're not doing that there because it's not a burning bridge for us in that particular case the variations of advertisements that you're running are nowhere near the same yeah the same volume as it would be for ppc because you want so many different variations and you're only paying per click so you know you can just head at it whereas with display advertising and out of home and some other things, we take more of a human-based approach. Having said that, the humans who are writing the copy are using the rubrics that we talked about to write good copy, even creative copy as well.

27:16Chris Marin: So we're using AI in a different way, not in an agentic way, but we're using it to refine the ad development, but less so than in the PPC space.

27:28Nick Panayi: I want to double click on something. You've used the word agentic. I've heard this a lot and I'm, yeah, I know what an agent is. I know what a GPT is. Agentic is like, it also sounds to me like something where I think you asked 20 people what agentic means. You're going to get 20 different answers. What is agentic? But I've got two part question for you. What does agentic mean? And the other thing is, is in one of the documents you shared with me, you talk about GPT is a zero shot. I didn't understand that.

27:56Chris Marin: Yeah, so I'm going to give you the non-technical person's understanding. Good, please do. Because I'm sure I'm technically not correct. But I mean, if I were to just generalize, agents are basically software models with varying degrees of autonomy. And they're equipped with specialized tools and knowledge. By definition, AI right now, it's all about specialization. We don't have general AI yet. So, I mean, chat GPT is general AI, but most AI systems are very finely refined on what it is you train them on. So, Agentic basically takes that concept and says, great, I'm going to train five different agents on five different parts of the problem.

28:45I'm going to send them off to do their part, and I'm going to bring back the answer at the end.

28:52Chris Marin: So instead of trying to figure out one general AI system that will have the answers to every part of PPC advertising, I created an, I didn't create it, Chris did, created an agent that goes out and searches for competitors. That's all it does. It knows the competitors. It searches for competitive ads. It brings back the intelligence in a certain way. And then I created another agent, like Chris did, that goes and critiques the ad that was written. And that's all it does. So it's easy to do the specialization. But they're all working towards the same end. And so, of course, through multiple ends, they're working towards the same end.

29:34Chris Marin: You stitch together an answer at the end. And that's how you go at it. And ZeroShot is really the idea that ChatGPT, for example, and models like it, like Gemini from Google, they're built to be able to understand, comprehend the question, understand what it is, obviously, that you're trying to get to, and then use pre-existing knowledge to go synthesize an answer for you on models that they're on information that they haven't been trained on yet. So one of the best examples I heard was, well, you know, like, I don't know if you use Google Photos. I use Google Photos. Google Photos uses, obviously, AI to do face recognition, right?

30:16Chris Marin: and you basically actually train it. It's like, oh, this is Nick's face. And it goes and does some measurements and it looks for Nick's face and it kind of figures it out. So it's been trained on my face. Whereas with Zero Shot, I may ask it, you know, who is Nick Panath? I'm Nick. Defining Nick Panath. And they'll go like, okay, have understanding, have context of what you're trying to do and use other methods to find you without necessarily having been trained on who you are. That's probably not the best example, but in general, that's what it is. It's zero shot in that it hasn't been trained on this data set that you're asking about, but it's able to use pre-existing knowledge to go synthesize and ask it for you.

31:02Nick Panayi: Yeah, that's a great explanation. And the way that you described agentic versus a GPT reminds me of an analogy that I heard somewhere, which is, you know, chat GPT is like having a fantastic specialist consultant that you can only interact with through chat, maybe through voice, but, you know, basically through chat. Whereas an agentic approach is like hiring McKinsey to solve a problem where you'd have multiple departments. You might have the financial analysts analyzing certain aspects of competitive analysts, and they all come together to give you a solution. and it's a much more comprehensive approach.

31:40Chris Marin: Specialization of labor, in this case, is specialization of AI agents towards a common goal.

31:46Nick Panayi: Yeah, makes perfect sense. I don't want to get too technical here, but I was really interested in the architecture, some of the tools that you've built, and you've developed this resource called Amaru AI, is that right? Or Amaru? Yes.

32:03Chris Marin: Yeah, I don't know how to pronounce it. The name is, I understand, is an Inca god. And I think it represents duality of human and god, I believe, in one. Oh, okay. Something like that. I'm not doing any justice. Chris, who is my guy, he was also born in Ecuador. So he used some of his heritage for that. And that is a tool, that platform that he has built called Amaro AI. And we're actually in the process of going from, okay, that's great, Chris. How do we then make sure others who are not as smart as you, meaning me and pretty much the whole marketing organization, figure out how to use this and what does it do?

32:50Chris Marin: So what he's built, and again, he's thinking a step ahead. He's thinking, okay, so we have 15, 16 different AI pilots as we speak. and I can share some others as well. So amazing stuff. But if he's thinking down the line where we're going to need an AI for marketing place where all of the best knowledge is housed and all of that starts to share information among one another, right? So as opposed to just having a content AI thing and an image generation thing and a personalization web page tagging thing, if we were to bring some of these models together into one place, you have this great effect of the various models learning from each other as well, number one.

33:44Chris Marin: And number two, it allows non-technical people to have a place where they can create workflows, multi-agentic workflows of their own, like for example right so it's supposed to be a no code multi-agent workflow right so he's got the ability to create prompts automatic so as you know in ai search prompts are the most important thing absolutely right i was trying to explain people it's like what people are using chat gpt or gemini as a search engine that's like one one thousand of the value it's almost like find me this. You know, I can still use generic search engines for that just fine. It's still valuable in that I don't have to click from there.

34:31Chris Marin: It gives me the answer. But if my question is multi-turn, as comprehensive as it can be, the answer is directly aligned with the complexity and the validity of my question. So the more valid, the more comprehensive my question, the more valid, the more comprehensive the answer. And people haven't necessarily figured that out, is you can go in and have the question be like two paragraphs. They say, I want this and this and this, but not that. And make sure you do this, and then don't do that, but do this. I mean, very, very complicated. And a human being will be like, hold on, I'm still on like step one of your question.

35:09Chris Marin: The beauty of AI doesn't have that limitation. The more turns you give it, the better. but you have to know how to phrase the question correctly so chris thought of a way said look okay if you're a copy editor or a creative writer or the cmo or a marketing strategist tell me what it is you're trying to do and i'll synthesize the prompt and go ask ai yes it's almost like prompt engineering for example so he's trying to simplify the world for all ai all

35:42Nick Panayi: marketers regardless of function that they serve in come in and use the most out of the ai tools

35:50Chris Marin: that we have single place to sign in single place to go as opposed to every marketer having their own you know like credentials to four different engines and subscriptions to four different engines come to one place and get your ai powered marketing in one place in one platform so that's what he's working on i saw the first variation like anything else version 1.0 is i had more questions than answers but it seems very very powerful and he's working on the next generation

36:23Nick Panayi: as we speak very interesting i also was quite impressed and curious about the way that you're you've basically got it integrated into monday yes yeah that's another thing is we wanted to put it

36:38Chris Marin: in a place where marketers live every day. Yeah, so Monday.com for us is a workflow platform. You know, we are our own agency in many ways. We use agencies, we use contractors, but we have our own content team, we have our own demand team, our own product marketing team. So a lot of the workflows are similar to what you would find in an agency. So we have, at any given point in time, two or three hundred projects with various people having a part of that project. Like, oh, you need to approve this. Does this look okay? Does that look okay? So that's where they go every day. So it made total sense if you're writing content to put the content generation AI tool in there.

37:20Chris Marin: So we have in there, for example, the content tool. We have an image remix tool. Again, back to using multiple engines. Instead of you having to go to like four different AI engines and say, can you make me a picture with a nurse, you know, holding, you know, a tablet in front of an elderly patient seeming concerned? I could do that. I could go to an engine and do that. But if I'm a creative, I want to get maximum options. So he created this tool that you put one prompt. It goes and interrogates, I think, four or five different engines, comes back with images, and you can see like, oh, OpenAI did this, Google did that, that one did this, and you're looking at all of them in front of you and you can say, okay, take this one, regenerate this one and regenerate that one.

38:15Chris Marin: So it's almost like processing. Very cool.

38:17Nick Panayi: Yeah.

38:18Chris Marin: But imagine how many steps it saved from doing that. Absolutely. Yeah, absolutely. And what happens is this is pure generative AI. You own that image. Right. So you basically own that image. you make it available to other marketers, it goes into your content library right in the same spot.

38:36Nick Panayi: Right, right, right, right. Yeah, that's a brilliant idea. Wow. There are some other tools that you use. I think you're using, there's one that I haven't heard of called, is it Tiga AI or Tiger AI?

38:48Chris Marin: Yeah, Tiga. Yeah, I'll tell you about Tiga, then remind me to also tell you about SynthFlow.

38:53Nick Panayi: Yeah.

38:54Chris Marin: So you remember in the beginning, I said that, you know, we have a combination of approaches. One is things like you need a Chris for, right? Things like what I just described. You need somebody to be able to take all these AI engines and bring something together for you, for your people, for your company. And it has context of everything, which is great. That's what Chris does for us. Not everybody has a Chris. Some people can just go out and look for platforms that already exist. Tiga AI is one of those platforms. It was actually founded by the CTO of Eloqua, if you remember that, Brian. So he and I were friends because I was an Eloqua customer.

39:35Chris Marin: So he let me in on a beta program for TIGA AI, and we've been using them ever since. So TIGA AI allows you to do AI-based prospecting in a very specific way. You basically have agents that you tell the agents what it is you're looking for. Your ICP, your type of person you're looking for, what kind of, basically what the ideal prospect would look like and conditions that would make them an ideal prospect. They go out and scan the web for them. Mostly, they use other sources as well, but they do their best work in LinkedIn. They find them on LinkedIn. They begin a conversation based on information that you've given them.

40:20Chris Marin: They do the initial outreach that says, hey, wonderful post. Funny you should say that. We have a white paper that you may be interested in because it goes towards this. Person says, oh, that would be great. Go ahead and send it to me. AI agent passes it to the human. The human continues the conversation. So it's a great marriage of AI and human collaboration in prospecting.

40:46Nick Panayi: Clever. Very interesting. Yeah. It's also, it's doing the thing that we've been training salespeople to do for the last 10 years, which this is a basic way to use LinkedIn and very few of them do it.

40:59Chris Marin: Yeah. And it doesn't scale, you know? So, you know, you basically, that's where AI can come in and make it scale. Yeah. And of course, to be authentic, once AI does the first introduction, you want the human to get involved because, you know, you could carry this on forever. but at some point we're not yet at a point where humans are super comfortable going all the way with AI they're comfortable with AI doing those first kind of steps and they want to deal with the human right so that's the beauty of this but at the end of the day I can see this going further and further the better AI gets it also is very good at finding RFPs which in B2B RFP hunting is A very, very valuable exercise, going out there, scanning the horizon for us.

41:46Nick Panayi: Very tedious too, right?

41:47Chris Marin: Yeah. Yeah, exactly. Finding the ones that are most relevant for you, bringing them back. And again, the human is going to take all of them. The other company that is also in prospecting that is very promising for us, it's still early days, but we're starting to see some of the ROI. It's a company called SynFlow. So that's Y-N-T-H, Flow, SynFlow. They happened to be out of Germany, out of Berlin. They were relatively small when we started working with them a few months ago. They're growing by leaps and bounds. I mean, you look at their logo page, you'd be amazed at how many companies are using them already.

42:22Chris Marin: So what these guys do is conversational AI prospecting agents. So they have conversational AI competency, obviously, in any language. Obviously, we train them on English in our case. We didn't train them on English, but we used them for English language. But they reach out and they do the initial phone outreach, which, again, humans don't scale very well for that, right? You tell a human, hey, you're going to be spending your whole day placing 100 phone calls. You may get to speak to, like, two humans. One of them is going to curse you out. The other one may be interested in what you have to say, right?

43:04Nick Panayi: It's not a very fun exercise. And that's a good dad, by the way.

43:08Chris Marin: Yeah, and that's not a good dad. AI does this very, very well. It acknowledges itself. You know, I'm an AI agent from company ABC. I'd like to see if you're interested in signing up for this webinar. Or we have a white paper which we think you'd be interested in. Can we send it to you or something like that? In essence, getting somebody to raise their hand in some sort of way. After they raise their hands, they're in your system. Now, I've heard, you know, obviously we have video of this system. It's available to you. Everything is recorded. You can listen to these conversations. I was blown away by the concept because I came from a conversational AI company, and I was there three years ago.

43:51Chris Marin: The degree to which this has matured is absolutely mind-blowing. There is no way I could play that voice and you would know that it's AI. the only time you the only time you would know there's still some of the wrinkles being worked out is that you know there was some delay sometimes because like the ai knows not to interrupt the human so if the human starts talking it stops so you have some of these awkward moments where literally there was one where it was like 30 seconds worth of goodbye like okay goodbye thank you goodbye you know one would interrupt the other so you have some of those awkward moments But the ability of the AI to understand what the human is saying and take intelligent routes to answering those questions that come back is incredible.

44:42Chris Marin: So we taught it, obviously, enough to know about our products and a webinar series or multiple webinar series that we're running. It would place the phone call. I've heard conversations where the person on the other end would say, first of all, are you human or AI? And I've heard many of those. And he would say, I'm an AI agent here to help you. Of course, I can connect you to a human if you like. After that, the person would say, well, this would be great for Jane, but Jane is on vacation and she's not going to be around. The AI would say, thank you so much for this information. I understand Jane's on vacation.

45:23Chris Marin: How about I send it to you and you let Jane know? I mean, that kind of thing. You cannot script that. That's the beauty of AI. If I had to script every possibility when the conversation would go, I would be called God. There is no way that you can script any of these conversations. But AI is smart enough to know how you must communicate. And you give it some leeway. and you make sure it's polite and you make sure that it doesn't do anything silly. But some of those interactions were actually just mind-blowing.

45:55Nick Panayi: And that's like sharing some of them out.

45:58Chris Marin: Now, again, has this blown the doors open yet? Not quite yet, you know, because you're still going to get the people who hang up don't want to ever talk to an AI. But this is the worst it's ever going to be.

46:10Nick Panayi: Today, it's going to get better and better and better.

46:16Chris Marin: And if it's mind-blowing now, I imagine like a year from now.

46:20Nick Panayi: I will look into that. And for anybody listening, there'll be a link down below in the show notes. I want to switch gears and talk about two of the issues that came out of the study. So these were the two top issues. So the first issue is a lot of the marketers in the survey said the number one thing that they're getting pressure from their CEO about is how can we use AI to increase the top line? That's number one. And number two, how do we measure the impact of all this AI investment? They may be the same thing, but not necessarily. So I'm curious about where you are on the journey on those two topics.

47:00Chris Marin: Yeah, it's interesting. When I first presented our executive, my boss at the time, we've had a change in the leadership team since then, but I presented this to my president at the time. That was his first question. It's like, wonderful, what's the ROI? And we had already developed, in essence, a custom ROI calculation for each one of these, because it depends on the use case. Like, for example, something as simple as content creation. Actually, take image creation. I know what I have to pay Getty for, right? And I know what I have to pay for generative images, which is zero. So I can do that calculation very, very easily, and it's a cost containment calculation.

47:50Chris Marin: I can use those dollars to be something else, especially if I do that at scale, right? Same thing with content creation is, Well, I've just shaved X number of cycles for my content creation. That means my content developers can create Y number of assets more per year than they could otherwise, right? Those assets bring in leads. So I'm able to kind of tie the leads that I bring in because the content brings leads in. When it comes to prospecting tools and things like that, honestly, it's a no-brainer. I don't know if the one, the last case that I mentioned to you, the cost is per completed call.

48:30Chris Marin: It's not even per call made. It's per completed call. And it can be a voicemail. It can be like an IVR system. So the fact that I'm only paying X number of cents per completed call, and then I can basically look at one deal that came through. I paid for it like for the next 10 years. Yes. One deal. One deal. Yeah. So it's very, very easy to do those things because AI does not require significant upfront investment. It's directly related to your usage, you know, and even that direct relation to your usage is so small dollar-wise that it becomes an old-brainer discussion. But it's always good to document it regardless, you know, regardless of the tool.

49:15Chris Marin: So each tool has its own ROI calculation.

49:18Nick Panayi: But I think there's something really important learning that I'm taking away from this. And that is that the way that you've got this organized is into very specific pilots and work streams. And that for each one, there's a KPI, which is related to that particular work stream. So if the image one's a great example, there's a KPI, which is the cost of images. So there you go. That's how you can measure the impact on the outbound calling. You can measure the impact and then tie that back because then you can track the lead all the way through to deals. You can see the impact of that. So that's very helpful.

49:58Nick Panayi: You have to get really granular to do that, right?

50:00Chris Marin: Yeah. And I should mention another example that I believe is even easier to demonstrate, right? So one of the things that we do as marketers, right, is create video. As we know, the more video, the better, right? I mean, human beings are very much attracted more so to the visual content assets than writing. So video is very, very important. We produce a lot of video. Sometimes it's presentations. Sometimes it's just animated videos and things of that sort. Voice over is very important. In yesterday's world, you had to go hire voice talent to be the voice of your video. You don't have to do that anymore.

50:43Chris Marin: There is high-quality AI voiceovers that can give you that the fraction of the price and a fraction of the time that it takes you to go to an agency to get five samples of voices, to pick one, to do the reading, to play it back, to go change. oh by the way somebody says oh i'm sorry you know i'll just change this one line needs to change the talent goes back to the studio it's done with ai instantly the company we use 11 labs yes i use them too yep yeah yeah 11 labs is one so that's a very very straightforward ai i give you another data point i don't know how i could put any roi to this but we had a fantastic event our ceo was one of the presenters 99.9 percent of the presentation was spot on there was one phrase and i forget what it was maybe it's just selected like just got it out of my memory that he said their own phrase it was just the phrase didn't make any sense for the context we had something else in mind he said it so it wasn't like you know mispronunciation or something like you, which is the wrong phrase, you know?

51:59Chris Marin: So at that, it's like, what do you do? I mean, it was a live event. We wanted the video up there because he was getting hits left and right. So we used one of his, you know, AI voiceovers, framed it on his voice, and had it say that phrase correctly, put it back in the video. You wouldn't know. I mean, obviously, you know, and we took his face off the video.

52:25Nick Panayi: Right. So there's no syncing issues. Brilliant.

52:29Chris Marin: But it was literally 0.1 % of the presentation, but it could have been one of those moments that somebody said, what? That doesn't make any sense, right? So we did that. The other thing that we find fascinating, at least I do, is AI podcasting. So obviously, podcasting is so important now. Again, you can take like a 100-page white paper and boil it down to an 11-minute podcast. And you have yourself an amazing peripheral asset that sometimes even becomes the central asset. So we had one of those cases which became the impetus to use this capability. We use Google Notebook LLM for this.

53:14Nick Panayi: Yeah.

53:15Chris Marin: It's on our website. The first one we did, we did multiple. But it was one around, I started with it with Harvard. it had I think four or five white papers as much as I like the healthcare industry that I'm in you know I could not read all five white papers and retain that content I listened to the podcast and I got it if I got it right you know that was a good kind of gauge and it was it's literally two AI podcasters you know a male voice and a female it's unreal again you cannot tell their banter between the two of them and they're joking around we didn't have to script that you basically load you give some guidance to it yeah the good news is you can say make it funnier make it shorter make it longer make it more serious yeah you know focus more on the facts less on kind of opinion but it's amazing what you can do so we're using that for our podcast as well so i've got two

54:15Nick Panayi: closing questions for you yeah by the way i mean we've been talking for an hour and honestly i could talk for another three you have the time for that first question is a very personal one so i run an agency what's your advice for consultants and your service providers now obviously you guys have you're probably because of the size of your team you probably don't use agencies as much as other many other companies do but what would your advice be for a firm like me mind?

54:45Chris Marin: Yeah, I would say be the Chris that others don't have. That to me, knowing the art and science of marketing and understanding how AI can empower marketing is a huge value add to any company. And I can tell you, and I've done presentation, the deck that I shared with you, Adam, I presented it, as I said, multiple times. And almost every single time people say, how do I get a Chris? I'm like, well, either you get a Chris or you get an agency that has a Chris or is a Chris. So the agencies can play that. Many, many companies don't have the ability

55:19Nick Panayi: to have the size of marketing organizations that we're blessed with, right?

55:23Chris Marin: So, but agencies can play a significant role. It's not just about producing stuff. It's about doing things more effectively, more efficiently and producing stuff is part of the mix. You know, it's really part of the mix going forward.

55:39Nick Panayi: Yeah. I'm reassured to hear you say it because I completely agree with you. And any agency that's not figuring this out or is in denial about it, they're going to be in trouble. This is fundamental. Anyway, enough about the agency stuff. So, you know, as I mentioned, right, the vast majority, so you're on the sort of high end of the top 7 % of marketers in our category in healthcare technology in terms of the way that you're using AI. What advice do you have for your peers in what they should be doing as they think about their journey? In addition to get it, Chris, the second thing I would do is make sure your marketers are

56:23Chris Marin: immersed in AI. They need to be comfortable themselves. They need to understand the role that AI has to play, and they need to ensure that their head is nowhere near the hole in the sand. because it's going to happen to marketing, but it's not a bad thing. It's a good thing. It makes us, as humans, focus on the things that we do better, and it lets AI do the things that we don't like to do anyway and just frees up our mind to do the things we like to do more of. The explosive opportunities for creativity, I mean the fact that you can write what you want the video to be and then you see it play in front of you a few seconds later I mean what creative person would not love to have that which one of us has not had an idea says I know what the next video should be and of course you're not gonna sit there you know oh that's gonna take a few hundred thousand dollars to do so no you know It democratizes creativity as long as you know how to work with it.

57:27Chris Marin: So yeah, I would say free up their minds, make them comfortable and tell them to go experiment and make mistakes. I think that is a great piece of advice to finish on.

57:39Nick Panayi: Nick, I am blown away by what you guys have accomplished. I found it completely inspiring. I know as an agency leader, I should probably be terrified by it, but actually I'm completely inspired by what you're doing. and I want to start an agency called Get a Chris.

57:56Chris Marin: I'll send you a t-shirt. I think we're about to make one.

58:00Nick Panayi: Nick, thank you so much.

58:02Chris Marin: Thank you.

58:03Nick Panayi: It's been a pleasure. Thank you, Adam. So that was my conversation with Nick Panay, recorded when he was at Inovalon, and how he built what I think is one of the first AI forward marketing teams in health tech. It's really quite amazing what he and Chris Marin accomplished. And by the way, if you want to catch the most recent episode I did with him, where he talks about agentic lead scoring in a lot more detail, there's a link to that in the show notes. A few things worth pulling out here, especially since AI adoption has only really accelerated in the last year. Number one, an AI first mindset has to be a leadership decision, not grassroots experimentation.

58:51Nick Panayi: Nick's team asked what AI could do for them before starting any project. And that only worked because his CEO gave him room to try and fail. Second takeaway, addressing your team's fear of replacement head on matters more than the tools themselves. Nick's line was that AI won't take anyone's job away, but people using AI will take away jobs from the people who aren't using AI. Number three, healthcare's privacy caution is real, but it doesn't have to be a blocker. Nick's team drew a clear line around what data they would and wouldn't touch, which let them move fast inside a defined sandbox. And takeaway four, AI-generated content formats like AI-hosted podcasts his team built from a white paper are worth testing.

59:45Nick Panayi: Even now, it's a preview of the same AEO and content discoverability questions we've been digging into on other episodes in this series. The last thing that Nick left us with is his advice for agencies and consultants still holds. Be the person in the room who can bridge marketing expertise and AI fluency because that's becoming the differentiator. Nick has moved on since Innovalon and started his new gig. I think by the time this show comes out, he will have announced that. And in the meantime, if this conversation was useful to you, please share it with someone in your network who might be really thinking about wrestling with the issues about how to use AI across their organization.

1:00:33Nick Panayi: And if you want to keep hearing conversations like this one, follow the Health Tech Marketing Show wherever you're listening. We'll be back with new episodes very shortly.