[00:00:00] Keith: I have everyone do when they're involved in a buying cycle is. [00:00:03] I'm gonna ask you to just spend 10, 15 minutes in a doc and give me your must haves. [00:00:09] You're nice to haves, write 'em down for me. [00:00:11] I now want you to record yourself narrating what you wrote down. Two things are gonna happen, you're going to add context Or you may introduce something that didn't make it on paper the first time. [00:00:23] And now I have two really rich sources of perspective. [00:00:27] and then I take both of those and I give that to GPT and I say, okay, [00:00:31] Go ahead and synthesize all of that together to come out with a really cool list of requirements. and that's what we then hold vendors against. [00:00:40] ​ [00:01:06] In This Episode --- [00:01:06] [00:02:02] ​ [00:03:51] Phil: Keith, thank you so much for your time today. Really pumped to chat. [00:03:54] Keith: Hey Phil, really glad to be here dialing in from an easy breezy 59 degree day here in New [00:04:00] York City. [00:04:03] Phil: What's the weather like for you, Darryl? Right now. Nice and nice and hot. Snow is melting here in Ottawa, Canada, so can't complain on my end. [00:04:09] Darrell: No, it's like 60, so perfect. You know, a little bit chilly for us, right? So it's always 72 and sunny in Southern California. So I have to put, put on a light sweater, but I'll survive. [00:04:23] Keith: We're just coming out of third winter here in New York. We had our spring of deception, third winter. Now I think actual spring might be here. [00:04:30] Phil: Mm. [00:04:31] Darrell: Um. Wow. All right. Um. [00:04:36] Why Sales Ops People Who’ve Actually Sold Have the Sharpest Knives --- [00:04:36] Darrell: Yeah. So Keith, you've had like a lot of different roles in your career and I'd love for you, for you to like riff a little bit on 'em. It sounds like you were an operator, then you went into sales, then you went into systems and sales ops. I've also been in sales, um, so I can kind of, I know that the pains of being in that role, but it's also, you know, it's also really great too to be in those, in those shoes and kind of, it gives you empathy for, for other people that you're supporting. [00:05:02] Um. Can you tell us a little bit about that riff on it a little bit like, like did you like being in sales? Did you like being in sales ops? What kind of brought you to really fall in love with the type of work that you're doing right now? Tell us about that. [00:05:15] Keith: Yeah, absolutely. I'd love to talk about that. So I always like to, uh, tell this story a little bit with, uh. That I kind of, I fell into ops by accident. I was given a sales role when I probably didn't deserve one, and then found my way back to ops where I belong. Um, I was hired very early on, uh, at a, a rapidly growing healthcare company, and I didn't even know if you had said, Hey, you should consider a career in sales operations, I would've. [00:05:46] Responded with what is sales operations? And I would've said that question weeks and even months into that job. Um, I, my title was Business Operations Associate. I was doing, uh, forecasting in [00:06:00] Excel on behalf of this business, on like our bookings and revenue based on like shipping projections. I got a chance to tailor a couple of different processes, things like that. [00:06:08] But that same company put me in a sales role. Um, they, they wanted to. Put more people in sales from inside the company, which is why I got a shot at doing that. And so next thing I knew, I was this, uh, field sales rep working for this healthcare company, meeting with CIOs, um, and Chief and Chief nursing officers, um, in the southeast. [00:06:29] I moved to Atlanta for that and. What I love about that role and later on when I would, uh, make my breakthrough into what I just kind of referred to as startup land, my first startup where I was employee number eight, also hired as an account executive there, um, is that I just, I I gained a lot of empathy for, for sellers throughout that, that, that journey. [00:06:47] Right. Um, it was at the healthcare company that I got my first exposure to Salesforce and CRMs and really like lightweight deployments of sales tech. Um, and that kind of [00:07:00] started this. Uh, journey in my mind of just like, I, I, I see a way forward to, to make a really cool career here doing this. It took me a little bit while to really kind of get my footing, if you will, but I'll always remember how hard that job was and how much I had to try to command my own destiny, right? [00:07:23] Like, yes, I was given tools, yes, I was given quota. Yes, I was given. Some direction. I had a manager and a one-on-one and expectations and make sure your deals are in, in CRM and all of that. So like, you know, not completely out on my own, but definitely not given a ton of training or a ton of like real time guidance or continuous education. [00:07:45] And many go to go-to-market Organizations are much better than the ones I had, uh, the opportunity to work in, uh, today, but. It is a really hard job to do. Um, and so I, I always try to keep that at the forefront of my mind. Now in [00:08:00] systems when we're trying to design these solutions and walk that balance between what the business needs versus what the seller needs, right? [00:08:07] Because those. Those two things can sometimes be, uh, you know, uh, competitive in a sense. Um, so yeah, it just, you know, it was a really wonderful experience. I, I, I love it that I got to do both. I'll never do sales again. Um, but, you know, it's, uh, I, I, I've, I stand by one, two, adjust managers. I think the best operators are former revenue influencers, salespeople, customer success. [00:08:33] Like, it, it, you just can't replace that empathy. [00:08:35] Phil: Yeah. Very cool. So cool to hear. I, I, I feel like the job was even harder like 10, 12 years ago when you were doing that. When there wasn't chat GPT and you couldn't have someone to bounce ideas off or, or have like a sales assistant, so to speak. Like I, I am full-time on the podcast now and I have a big sales role to, to kinda like pay the bills now and find sponsorships.[00:09:00] [00:09:00] I rely on chat GPT so much for having, uh, like I've got a custom GPT that's just like my sales director and he's giving me advice on scripts for like my media kit, how I chat with folks. And the job is so tough. It's really, really hard. And yeah, like I totally get what you said, like building that empathy for sales. [00:09:20] It's, it's a really tough job. No kidding. You were sales too, Darryl. I actually didn't know that. [00:09:25] Darrell: Yeah. Yeah. And I don't know if this resonates with, with, uh, you Keith, but I felt like when I was selling one-to-one right. I would always feel like, man, I. This deck sucks, or, man, this, this process sucks. And like, I wish I could just fix it for everyone. And like, that was like the, the thing that kind of got me to like, I, I, I, I couldn't just, um, I wasn't satisfied just talking one-to-one with a person. [00:09:58] I wanted to help many people. [00:10:00] That was like sort of my transition into marketing because marketing is, is uh, one to many versus sales is one-to-one, and then ultimately the system. Does that resonate with you at all? Um, [00:10:09] Keith: 100%. And it actually, um, funny enough, that actually is what, um, organically resulted in the first step of my transition back into more pure operational and systems roles. So, uh, I'm at this early stage startup, um, data as a service offering in Atlanta, Georgia. Hired as employee number eight. I'm like the second or third account executive I can't quite remember today. [00:10:34] Um, and I got to know the product pretty quickly. I'm, I'm not the most technical person in the world, right? I can't code to save my life, but I do have, you know, a bit of a, of, of the brain, of, of an engineer, right? Like, I can, I understand the concepts. A lot of it does come rather naturally to me. And so. I got to know the product really well, and before I knew it, as we were hiring more and more salespeople at this startup, I was finding [00:11:00] myself teaching people about the product. [00:11:02] And I loved that. And like to your point of like, I didn't want to just be good at speaking to marketers who we sold to at a technical level, what our, our data product did. I wanted everyone to be good at that. What do you know? It CEO pulls me aside one day, pulls, uh, says come to my office and he goes, Hey, listen, I need to tell you something. [00:11:21] You're really bad at cold calling. Um, you're terrible at it. Like, and you don't even do it very often. And I was like, you're right. I, I hate that job. I'm all school. I. I'm also really introverted, so like that makes sense. Um, and he's like, but you know the product better than anyone else on the floor. He's like, so here's what I'm thinking we should do. [00:11:42] You should be our first ever sales engineer where you're gonna actually do the technical demo and do the proof of concept. Um, how does that sound to you? And I'm like. Absolutely, let's do it. And he goes, oh, and by the way, our Salesforce admin quit. You wanna learn Salesforce too? [00:11:59] Darrell: you go.[00:12:00] [00:12:00] Keith: So that's how I got back into sales ops and sales engineering. [00:12:02] So yeah, totally resonates. [00:12:04] Phil: Very cool. [00:12:05] The Difference Between GTM Ops and GTM Systems --- [00:12:05] Phil: So GTM systems now at OpenAI, and you were GTM in that kind of role when you're at mural as well. You got to get a whole overview of the landscape when you were an analyst at Gartner. What is GTM OPS versus GTM systems? How is it different for marketing ops and sales ops? Just curious your take there. [00:12:25] Keith: Yeah, absolutely. So, I mean, the, the easy answer is that it depends, right? Depends on the organization, depends on the philosophy and perspective of the leaders overseeing those respective functions. But if I had to give like a, a dictionary definition for them, I would say that, you know, the ops side of the equation is more about the. [00:12:47] Field facing, field enablement, process innovation, continuous education of like making sure that the field and its leaders, and when I say the field, I'm talking about folks in sales, maybe the individual [00:13:00] marketers, product, marketing, whatever have you, that they have what they need and that you're working with them in partnership to be that operational counterpart. [00:13:06] Right? But then in go to market systems, we're more of the folks that, for better or worse, are kind of behind the scenes, right? We're behind the pain of the. Third and, uh, first party technology that's being used by those individuals every day. But that's a very different kind of work and requires deep focus and, um, you know, needs to be done in a really methodical way. [00:13:31] I mean, you can get by by doing it quickly in haphazard, but eventually, especially at scale, it will come back to bite you. And so I think it's a perfect balance at the end of the day. My team, for example, both at Mural and now here at OpenAI, we rely heavily on our colleagues in the operational teams to feed us the guidance of what the field needs so that we can then build the right solutions on behalf of them. [00:13:59] [00:14:00] Um, a systems team could do that top to bottom, but the reality is if you wanna do the technical work as well as you need to. You can really benefit from this sort of yin and yang, if you will, of ops versus systems. Um, not really a versus not a competition, you know, if they're we're side by side opening ai, we were side by side at mural. [00:14:16] That's my favorite combination. [00:14:18] Darrell: Yeah, there's a, I think there's also like a, a amount of, or like amount or volume of work that is, you know, component of this. Because at smaller companies sometimes you can just do both. You know, like, oh, I can, I can run the CRM, I can run the map and I can talk to. Salespeople [00:14:38] Keith: And I, and I, and I did both for three jobs in a row, but all of those companies were just a few hundred people. And at one case, I had a sales team that was just low double digits, right? So like, you're right, they're all like, you can do both. But if the volume of work and scale is there, eventually they have to be separate functions. [00:14:56] Darrell: yeah, yeah. And you, the, the ops role, I think if you're [00:15:00] separate, if you're taking the tech away from it, is like really critical. And, and it doesn't really, I, I, I agree with you. Like it doesn't really matter. It's not like. Uh, you know, you're not competing. Um, to be honest, the ops people could sit on the team on the te on the systems team, just maybe like as a separate group. [00:15:20] You know, like, it doesn't matter where they sit. The, the more important thing is, you know, do they understand what the business needs are and do they understand what, what the people in the field need? And can you translate that to technical requirements, you know, and, and working together. Yeah. I think, um, you're right on. [00:15:37] That's, that's, that's a, a really good way to put it. [00:15:41] Why Data Dictionaries, Syntax, and Prompt Engineering Will Reign Supreme --- [00:15:41] Phil: Well, let's chat about agents eruption in GTM. Keith, we'd be remiss by not having someone from open, from open AI and, and not talking about AI agents as opposed every day that someone's writing about this, this eruption and tons of confusion around this space. Lots of [00:16:00] excitement and, and definitely a bit of FOMO when it comes to AI agents. [00:16:03] But I'd just love to give you the floor and let you share what you think is happening in this space with agents where you think we're going specifically, like what's vaporware, what should folks know that is kind of not really serious and, and what do you think is, is legit? [00:16:19] Keith: Yeah. So I think you know where we are today in terms of the, the technology landscape, right? Um, the, the agent offerings that are on the market or are coming to market that are really of further investigation, if you will, at least this is my own personal perspective, are the ones that are. Being set up, sold, configured to do, let's just call it the more robotic tasks that us as humans and those in sales and marketing have had to do for a very long time. [00:16:56] Right? The things that are, are very clearly if this, then [00:17:00] that or turnkey perhaps, right? Because that's what agents are gonna be really good at right now. As the technology becomes more and more sophisticated, and there's anyone who's followed. Opening AI has seen, right. Like the, the models just keep getting better, right? [00:17:15] That, that paradigm will continue to shift. Right. And eventually the, the agents will have near or even better human level intelligence, but at the same time, they won't always have the same context. Right. So, and I've, I've seen this take on, on LinkedIn and elsewhere, and I, I really do agree with it that. [00:17:36] If you're going to try to do anything involving some version of agentic orchestration, whether you're trying to build it yourself with technology like what OpenAI offers, or through a third party vendor who's, you know, making a big push for this sort of thing, you have to really confident in what I'll call the less sexy elements of any [00:18:00] deployment, right? [00:18:00] All the things that. If we are all being honest, we know we need to do and we probably could do better, but especially if you're in hypergrowth or early stage, you might count them more of a, a luxury than a, than a necessity. Right. And I'm talking about things like having really clear, understandable definitions of what the different data means in your organization. [00:18:25] Having a consistent level of syntax. For that data, right? When I say X equals Y, I need to say that every single time. And then in addition to that, you need to make sure your data is as consistent, tagged, categorized, codified as as much as possible because again, the, the agents you know, aren't there yet where they can. [00:18:48] Fill in some of the gaps that the human brain is capable of doing today, right? And so you've gotta make these things more mechanical in nature, if you will. Um, the models [00:19:00] will get better, they'll be able to make more and more educated guesses or informed decisions as as we go. But, um, where we're at today is that you've gotta have your stuff in order and if you want to deploy anything. [00:19:13] So, you know, I, I wouldn't say that there's, you know, a. All be all dictionary definition of this is vaporware. This isn't. I would just say you have to have these things in place and you should be trying to deploy them to do these types of things. If you're trying to do other things, then like all power to you, like the innovation will come and, and I think it'll be here sooner than we know it. [00:19:34] Um, but that's where it's a little more conceptual at this stage, if that makes any sense. [00:19:40] Phil: Yeah. Yeah, I think that's the great advice and, and music to the ears of ops folks who have been trying to get those foundation pieces in place. For a decade, and now everyone's screaming about AI and we need to get AI in place, and the ops team finally has a really good case to build for data quality and data dictionaries [00:20:00] and, and all this stuff you, you kind of talked about there. [00:20:02] I'm curious. [00:20:03] Keith: real quick, you know what's a really interesting point though, is that even where the technology is today can help accomplish a lot of that, right? Because some of that is somewhat, uh, menial or mundane in nature, right? It's a lot of, you know, read through this, analyze this, come up with definitions. [00:20:21] We built our entire data dictionary using our technology because. You can trust the models to make elementary sort of level assumptions of, oh, I see this field is called X, Y, Z. I think it's this right? And you can shortcut the journey to having all that stuff in order with that technology [00:20:41] Phil: Very cool. Yeah, I'm sure there's a lot of. Dog feeding from the chat. GPT tech that you guys are using internally? Uh, they're and our, our power users. But, um, I, on this like agent thing, like there's, [00:20:54] Every Martech Vendor Wants to Be the AI Orchestrator. You Still Have to Be the Operator. --- [00:20:54] Phil: there's one question I have in my head that's like around the orchestration layer. Like I chat with a lot of listeners and folks that work in GTM and in ops and. [00:21:04] There's like, everyone is keeping tabs on this like race for this orchestration layer. And a lot of folks are confused about where this is going. And I'm curious to get your take on this, like every MarTech vendor is integrating agent capabilities inside of their tooling. Whether they're connecting directly to OpenAI or recalibrating, like their own value props and trying to build something homegrown. [00:21:27] Like what advice do you have for MarTech or, or systems, uh, pros that you know, are thinking of this future or. May, may, maybe even like current reality for a lot of folks where there's a ton of complexity exploding with AI agents from different vendors with their own orchestration logic, like some folks have said, we're almost gonna become like AI referees where we need to decide to turn it on in this tool, turn it off, and that tool all for customer success, we'll use this tool like. [00:21:57] Maybe chat about like, how do we work in an enterprise [00:22:00] environment when every tool in your MarTech stack is saying like, turn on the AI for this tool. [00:22:05] Keith: Yeah, for sure. So, uh, I'll say, you know, in initially that. It ultimately will come down to what you wanna do. But the guidance that I think I would offer this audience is, I would say that the, this game isn't all that different than what we've been doing for the last 10 years on trying to get these technologies to talk to each other. [00:22:25] Right? And it's about coordination, right? If I've got. And let's just use the, the timeless example of marketing automation and CRM, right? Uh, anyone who's listening to this podcast has probably done any number of integrations with any number of maps and any number of CRMs, and we can probably all have a horror story or several of trying to get those two to talk to each other, right? [00:22:48] But if you do take the right amount of time and understand the, the timing and order of operations. Between the two, right? Like, uh, you know, you know, obviously my, my career has been [00:23:00] pretty heavily with Salesforce, right? That's the CRM I've used most often. Um, Salesforce has a lot of things about it that are pretty, um, routine, if you will, right between the way that it runs its processing queue to the amount of volume it'll handle at any given time, right? [00:23:16] Um, and in order to deploy Salesforce at scale. You need to know those things like the back of your hand, right? Because otherwise you might start building a solution, whether it's declarative or custom developed that isn't really compatible with the way the the core system runs, right? So I would say that there's really not a lot of apprehension that's warranted to say, like, don't turn on the AI or don't. [00:23:42] Use that vendor's orchestration layer, but you need to understand it intimately because at the end of the day, what you're doing as a technologist is you are playing translator between vendor A and vendor B and the agents or orchestration layers within them. Now there's [00:24:00] always gonna be space for maybe some sort of master orchestration layer. [00:24:02] That's then maybe essentially doing that job for you eventually, right. You can't teach something to do something if you don't know how to do it yourself. So I would just say that you've got to know how these layers are, and Ag agentic workflows are operating at a very intimate level. [00:24:21] Darrell: Yeah, I, I completely agree and [00:24:23] ​ [00:26:38] Darrell: kinda like what I'm, what I'm, what I'm hearing a lot is, you know, we have to be careful to, and, and, 'cause what I'm seeing out in the market now is people are using AI and companies are using AI just for the sake of it. You know, it's just like, oh, you know. We should use ai, but we're not really sure what we're gonna use it for, but we should use [00:27:00] it. [00:27:00] Otherwise we're gonna like either miss out or something like that. And so often, um, the core foundations of, you know, uh, sales and marketing plays, they're for, for many, for many companies, they're just not that good. You know, and they're missing some of these critical things. And that's what, what we really should be using AI for, like what you're saying, the data organization, you know, the infrastructure, uh, making sure everything's formatted correctly so that we can run when we need to. [00:27:30] But I think that, you know, in so many places, you know, and I get, I get approached by founders all the time and they're like, Hey, can you let, let me show you this new AI tool. And it's for something that like I didn't even wanna do in the first place. You know, and I think that's kind of the theme that I'm kind of hearing is like, you know, you're, you're harping on like, let's get all of this, you know, uh, orchestration layer correctly, let's get the data correct so that we can do things. [00:27:57] Agents Work Best After You’ve Translated Instincts Into Clear Steps --- [00:27:57] Darrell: But it seems like companies are just trying to, you know, [00:28:00] maybe sprint before they can even crawl. You know what I mean? Is that kind of. [00:28:04] Keith: Yeah, I think there's, I think there's a lot of, uh, a lot of, uh, validity to, to that premise. Um, I'll, I'll offer, uh, I think what is a, an apt analogy. It's not a perfect one-to-one, but it, it's a good example of a really, of something that we probably, um, I'm gonna say in a moment, and anyone listening is gonna be like, might even roll their eyes for a moment, but I'm gonna compare the, the future and the right way to approach ag agentic workflows to lead scoring. [00:28:31] And specifically how you operationalize lead scoring. So I'll tell a quick, a quick story. Years ago, I'm at a sales kickoff early stage company. One of those ones where I could do ops and tech at the same time. Right? And I'm on stage at the kickoff, unveiling our first ever lead score. 'cause we were, that early stage, we didn't, we hadn't had one yet. [00:28:52] We hadn't needed one. Right. Um, and I got several questions during that session of like, why do I even need this? I, [00:29:00] I, I know what a good lead is, and I'm like, you do know what a good lead is, which is why we met with you and why you gave us all this great insight to inform this lead score. But tell me real quickly again, what do you do to determine what a good lead is? [00:29:13] Right. And of course, the seller very good at what they do is like, well, I check A and B and C and D, and I'm like, great. Well, the lead score did that. It, it, it outputted the score based on A, B, C, and D and the values that we found there. And now you have a list from highest to lowest, so I can have you spending more of your time at the top of this list working your way down. [00:29:37] Right? That's the same approach to the A agent workflow. It's like, Hey, let's replace. Something that I do today as a human that frankly I do automatically or I do somewhat subconsciously, right? Because if I do that correctly and do it at scale with an nagen approach, then now think about all of the time that I am freeing up for myself, my [00:30:00] colleagues, my operations, my, my, uh, field team. [00:30:03] That's where the unlock is in my opinion, because. People who are much smarter than me speak to this in really wonderful ways. But I think that the, the future of AI is really about unlocking new levels of human creativity. I think that is the real promise of what we are bringing, um, with this technology. [00:30:22] Darrell: Yeah. [00:30:23] Phil: Very cool. Great, great analogy. Anyone listening probably has played a role in, in lead scoring. It was one of like my, my earliest intros into sales management and sales process between like MAP and, and CRMs there. So let, let's say Keith, like someone is. Your mentee and they're asking you like, you're so close to this tech. [00:30:46] The Ritual of GTM Ghost Stories That Turn Pain Into Pattern --- [00:30:46] Phil: You're in the leading company in this space right now. What advice do you have for marketing ops GTM practitioners to future proof their careers? All the stuff is moving so fast, where should [00:31:00] they be going to learn? Where do they, where do you go to learn? Curious your thoughts there. [00:31:05] Keith: Yeah, so I think I'll, I'll start by responding to that with more of a, a framework or a philosophy, if you will, to it. Um, and my very first boss ever, the guy who was, uh, maybe dumb enough to gimme that ops job, uh, I. You know, I, I think I did an okay job, he said so at least, so maybe, maybe it wasn't a smart decision. [00:31:24] I, I'm grateful for it. But he, he had a phrase, uh, that we, we lived by, uh, and it was do it, try it, fix it. Which it feels like maybe it should be Try it, do it, fix it. But no, it is, do it. Try it, fix it. Because we would just do the thing, right? And then we would keep trying and keep trying or iterating and improving over time. [00:31:46] Right? So I would say. This technology is so new, so paradigm shifting, changing so rapidly, and every vendor in the market is constantly trying to reinvent their version [00:32:00] of the wheel with it. So you have to have this iterative. Foundational and build and build and build and build and build approach, um, as part of just your core being of how you approach every business problem. [00:32:14] Um, and that may sound rather routine to others, but like it's really easy to like get presented with a business problem and try to boil the ocean, right? It's like, no, take the baby steps. Take the baby steps, validate that you're stepping in the right direction and when you know that you are then tar start to pick up the pace. [00:32:31] Right. Um, so I'd say that's, that's the beginning. And then I would say in terms of, you know, just some things that have been really beneficial to me in terms of shaping my career, I. Honing my skills as both a technologist and a leader of other technologists is first and foremost, like a community is a big thing. [00:32:48] Like you just need to find your community. And I'm not saying that it means you have to go out and join, you know, the, the flashiest loudest MarTech or rev ops community on LinkedIn, right? I mean, find those [00:33:00] humans. That are doing the same job as you that you can have a bond with, that you can share notes with, that you can bounce ideas off of. [00:33:10] It's, it's really no accident. Um, even though it is somewhat organic in nature, that a lot of my close friends are fellow operators, right. And it's 'cause we speak the same language and we're constantly getting value from each other. Right. But at the same time, we're also individually and collectively learning some of the repeatable, best practices that we can all, um, speak to, right? [00:33:31] So that's the next piece is that you gotta just, there there are good ways to do certain things, um, that, that most people understand, languages that people speak, right? So learn those things so that you can repeat them over time. And then the last piece I would just say is like, don't be afraid to fail. [00:33:46] Like it's scary. It's really scary. But if you do that iterative approach, going full circle here, right? The failures in the beginning are smaller, the risk is lower, but the education is just as valuable.[00:34:00] [00:34:01] Darrell: Totally. Yeah, I, there's like an uh. Another, another way to say this is kind of like, it's like kind of like dis declarative knowledge versus experiential knowledge, which is declarative is kind of like when you read about a thing, when you read about how to do something, but then it, you, it, it only helps to some to a point. [00:34:22] You know, it's kind of like reading about how to ride a bicycle. You know, you can read tons and tons of books about how to ride a bicycle, but like. You really haven't ridden one till you, until you've tried it. And I, I, I think that with a lot of people that are kind of, um, I don't know. S maybe more academic or maybe more like f fear of failure. [00:34:44] They'll study something so much, but they'll never actually go and try it. And that's why I love your do it. Try it, fix it. Um, framework. And I think the other thing too that's really great about that is a lot of times the skills that you learn along the way are transferrable to other things. So when you're [00:35:00] fixing your initial. [00:35:01] Model or version or platform that you're launching, you know, those skills that you're getting that's like experiential knowledge that really compounds. Um, so I, I, I think that your advice is right on, and, and people listening to this podcast should really like, you know, take that to heart and, and try. Um, but, but, uh, like you said, um, you know, it, it takes courage. [00:35:22] It's fear of failure is a fear. Fear of failure is a real thing. [00:35:25] Keith: it is, it is a real thing. And, but I want to like encourage the audience also with this, um, caveat, which is that when you do fail, the, the memory of that failure and the lesson learned from it is visceral. [00:35:38] Darrell: Hm. [00:35:39] Keith: will go with you everywhere you are. I, uh, a couple weeks ago in a like team weekly wrap up, the, the prompt that gave everyone was, we're gonna do go to market system ghost stories. [00:35:51] We're gonna go around the horn and no judgment. We're all gonna talk about a time that we built and deployed something that blew up on us. Um, and [00:36:00] guess what? Everyone on the team had a story. Immediately. Right. Um, and I remember the story I told, then I remember other stories that I have, and guess what? I don't make those mistakes anymore, right? [00:36:13] Because they were so visceral and so painful in the moment. But guess what? I'm still here. I'm still growing. I've advanced in my career. Um, so like you gotta, you have to embrace the failure to a certain extent, or at least the potential for it. [00:36:27] Phil: Yeah, great advice. A anyone who's worked on a marketing automation platform has a story around sending an email out to the wrong list or an unfiltered list. It's a rite of [00:36:38] Keith: Forgot about the suppression. Yeah. [00:36:42] Darrell: Yeah. [00:36:42] Phil: Classic. Classic one there. Um, [00:36:45] Why AI is Rewriting, Not Killing Martech --- [00:36:45] Phil: Keith, I wanted to ask you about the death of SaaS or. Like a lot of folks are talking about the dark age of MarTech that might be coming because of ai. You probably read a story or two about this. Lots of folks are claiming that AI agents, uh, will usher in the death of SaaS. [00:37:03] Obviously MarTech is a big category of SaaS. Some of the arguments made are that a lot of the SaaS MarTech vendors and sales tech vendors are tools that address specific needs, like niche, comprehensive, uh, like niche solutions versus. Comprehensive platforms like Salesforce or HubSpot are often sometimes overkill or seen as overkill from some of these folks. [00:37:25] Like their arguments is that AI power, DIY solutions, or agents that are being built internally are gonna replace a lot of these SaaS tools that address some of these, like, not minor, but like smaller niche, uh, tactical needs. So. Not necessarily that like engineers are gonna go off and build all these like in-house MarTech solutions, but that in a couple of years or already today, a lot of these teams are equipped with more power, more capability to replicate in-house solutions from these MarTech and these SaaS tools without needing to buy them. [00:37:59] What [00:38:00] are your thoughts here is, are AI agents bringing on the death of SaaS or do you think that's kind of bullshit? [00:38:06] Keith: I, I, I don't think it's bullshit, but I will say that I think the version of SaaS that many of us came up in is not going to be here for much longer. So I'll say that like software as a service. Buy, buy a thousand seats do only X, Y, Z with it. That is definitely going away. Right. But at the same time, you know, innovation only happens at the pace of its slowest individual to understand the value of that innovation. [00:38:36] That said, the pace is getting faster thanks to ai, right? We're able to do more with less time, which is great. Uh, but what I think is ultimately where we're headed is that these platforms will be changed. Forever, and it may mean that we need fewer point solutions that could happen. Right before chat, GBT broke the internet, we [00:39:00] already knew that consolidation was happening, right? [00:39:03] Vendors were buying each other or co-developing, you know, products and features that mirror their competitors or their adjacent solutions like. That's just capitalism at play, right? Like that's gonna happen. Um, but I think agents and, and agentic workflows and age agentic orchestration. It, it will be, you know, the battle of the platform that does it best like that is, that's a reality, right? [00:39:31] The, the vendors, uh, that maybe are being perceived as overkill today that don't do it as well as the one that does, probably will lose the battle to that, that player. That's, that's a reality that I believe in. Um, but at the same time, I'll say that, you know, you still do need these platforms to an extent. [00:39:50] Um, there's, there's no getting rid of them forever. At least not anytime soon. [00:39:55] Darrell: Yeah, I, I, and I think that it, the platforms [00:40:00] themselves, my prediction is they will just, you know, consolidate the AI capabilities into the platform so that the work that humans do is more seamless. And that's like, that's the way that I can see, I see it. I don't know if I've ever shared this prediction, but like the, I think that everything, um, starts to converge to convenience. [00:40:25] So, you know, my prediction is actually in the future, probably like way in the future, but your phone just kind of does everything for you. You know, and I don't know if you, if you, both of you watched Westworld. But, uh, in Westworld season three, uh, it's like a futuristic city. And Dorothy Dolores is walking around and she has like an agent in her ear and she just tells it what to do. [00:40:47] You know what I mean? She's just like, can you buy this apartment for me? And, and, and it just buys it. So like, that's how I see, um, um, AI playing, like, like playing in the future. And, and, and why? [00:41:00] I, I don't think that I agree with you, Keith. Like I don't think SaaS is going anywhere. It's just gonna be improved. [00:41:06] Buy ai, you know what I mean? So, and, and I also don't buy this idea of like, there's so many different agents and like the, you know, there's this joke, it's like, have my agent call your agent. You know, and like, I just don't buy that. Like, it's just not convenient, it's confusing. Um, so I really do think that the SaaS players will really embrace and, you know, our work will become easier and it's not gonna be this. [00:41:31] Let's all connect everybody's agents together like Legos. Um, and I know you, I know you like Legos, Keith, [00:41:36] Keith: I do, I do love Legos, that's for sure. Um, I will, I'll offer one slight counterpoint to that. I, I, I do think that there is a, a layer really far down on like the, let's just call it the micro transaction layer, if you will. Where it may make sense for one agent to talk to another, right? Uh, especially from one platform to the next. [00:41:56] Like, I think that is going to happen to a, to a degree, [00:42:00] um, but not to the level of SaaS is dead forever. Um, there's a big difference or big, big delta between those two [00:42:08] Phil: Mm-hmm. Yeah. It, [00:42:10] Startups Built by One Founder and a Network of AI Agents Are Already Emerging --- [00:42:10] Phil: it'll have to happen in some shape to have a future where the CEO is running a one person company and their job is essentially just managing agents. A lot of folks are talking about this. Uh, curious to ask you, like, how many years away are we from having the first startups be just a single CEO managing a bunch of agents? [00:42:35] Keith: Yeah. Um, I, this is a fun one. Um, I, I think, you know, to start off, you know, an easy territory, I would say that it depends, but specifically, I would say it depends on the nature of the CEO. Right. So if your question is how, how many years or how long till, this is a more common reality, right? I would say that, you know, the, the sales slash [00:43:00] revenue slash customer focused CEO, right? [00:43:03] The, the person who's really good at getting on stage, delivering value propositions and meeting with their customers and, uh, conveying, you know, their need to. The solution they offer, that person probably is a little bit further away from being able to do that because they're going to need the agents and the technology powering them to get good enough to the point where they can be that one, that one person company. [00:43:28] Right. But then if we think about more of like the operational focus leader. You know, they're gonna be able to see the, the building bricks, if you will, to these platforms a little bit more clearer just by nature of their expertise, right? They're all about plugging A to B, equal C, that sort of thing. And then you have the technical leaders, right? [00:43:48] Those are gonna be the people who do it first. Um, because they're the ones that, that, that's, that's how, what they're thinking about with their technology, right, is how do I multiply the force of human or [00:44:00] humanity, um, on behalf of the technology that I'm building. Um, and I think those folks will be doing some version of it much sooner, but there will be a sliding scale of adoption, uh, just like there is anything, uh, in, in, in the market and in society in general. [00:44:21] Phil: I was letting you, letting you ask the next one [00:44:23] Darrell: Oh yeah, of course, of course. Um, let's talk about, I like. This question, let's talk about the qualification of ideas, um, which is a really interesting topic. Um, so there's a, there's a, [00:44:35] An Operator’s Framework for What Gets Done, Delayed, or Silently Dropped --- [00:44:35] Darrell: there's a ton of projects that we have to do there. There's, at any given time, there's probably, I don't know, 50 plus things that you can do when it comes to GTM systems. [00:44:45] How do you think about qualifying, prioritizing, um. Resourcing these projects. What, what comes to mind when you think of, okay, this is a, a must do and then this one, you know, we can actually put on the back burner. Can you tell us a [00:45:00] little bit about your qualification framework, [00:45:01] Keith: Yeah, absolutely. So I would say that there are three dimensions to this. The first is. What you, you have to have, which, um, and you really can't get anywhere to any degree of repeatable, reliable prioritization, uh, without it. But you've, you've gotta have a true north, you gotta know what your, you know, what are your existential problems or what are the. [00:45:23] Really well-defined goals of your team, your business, the organization that you serve, right? Um, you gotta know what, what direction you're walking in, um, or running in maybe. Um, but after that, I would say it comes down to understanding the degree of need or maybe what I often like to refer to, uh, at least in the realm of like buying third party technology product need fit. [00:45:47] So does the product fit the need? Right. Um, that's a whole practice onto itself. Um, and I have a lot of. Ideas and uh, and [00:46:00] processes and methodologies on how to determine that. And I could probably do a whole podcast about that. Um, but I would say after that point, once you determine if product need fit is there, and product could be the internal solution as well, right? [00:46:13] It doesn't necessarily have to be third party. Um, next you have to understand the complexity of it, right? So if you know that I have this problem, I have this product that meets the need. It's an insanely complex thing to do. Then guess what? You get to do that thing and maybe nothing else, right? But you may find that I have this need and I have these four products available to me that are of lesser complexity. [00:46:42] Maybe they each offer cumulatively more value than that single product that's insanely complex. So I'm gonna do all four of these versus that one thing. Um, and then there's a, you know, not necessarily a fourth overall dimension, but I do tend to ask this question like, what happens if we [00:47:00] don't do it? Like what, what happens if we don't buy the thing or we don't build the thing? Like, will we be okay without it? Um, you know, there are a lot of things in this life and in my professional work that I really want to do. Right. Um, I'll give you, you know, a really practical example. Um. I've, I've been at opening AI for a little around 18 months now. [00:47:22] Right. I was the, the founding member of our go-to-market systems team. I've had the privilege of building this team up, and while we've been doing that, we've been walking a really fine balance of how much process to embrace for ourselves, right. I love process. I love spreadsheets. I love method, methodology. [00:47:41] So like I would love to do every ounce of process to have a beautiful board of knowing where everything is. And if you, if you select this is at risk, you do this and all of that. But guess what? We're okay without most of that. We're doing really good work at the same time. So, [00:48:00] uh, you do need to ask that question and be really truthfully honest with yourself and your team and your organization. [00:48:05] What happens if we don't do it? Will we be okay if we don't? Um, and that's just, you gotta, you gotta answer that. You gotta ask and answer that question. [00:48:15] Phil: That's such a good point. I. [00:48:16] Darrell: I assume, like, and, and this is probably, uh, you know, just like laying this, this one up, but I assume you're also like iterating with chat GPT when you're doing like the prioritization, um, and the, and the risk analysis. Yeah. [00:48:30] Keith: Yeah, well, it's like when, when you do find that that thing that is a little bit more turnkey and you're like, okay, this is how I approach this particular question or problem each time you're probably in, you know, again, being selfishly promoting open AI technology, I say, you're probably in GPT territory. [00:48:45] You know, like that means if, if you're feeling pretty routine or repeatable about what you're doing, then go build a darn GPT and tell it how to do that, and what inputs to ask of you and what outputs to provide you in return. [00:48:56] Phil: Yeah. Yeah. Love that. I actually have a custom GBT for helping [00:49:00] me define the effort that takes for a certain task. Like this is something I've always sucked at in my career, and a lot of folks like. Deciding prioritization frameworks, like the effort factor in this is so hard to gauge when we've never done it before. [00:49:17] And so what I'll do is I'll like give a list of tasks and the effort that I think each task takes, and then I'll ask you to tell me like, what am I not thinking about? And then there's always a couple in there that. I didn't think about and then I just like repro it, like based on these things that I didn't think about, what else am I not thinking about? [00:49:38] And which of these tasks do you think could take a lot longer because of external factors that we're not thinking about And like this whole like exercise of iteration and backing forth I think is really helping. Effort estimation because like any engineering team or like MarTech technology team knows that you can always two x the amount of effort or time [00:50:00] you think it's gonna take to implement something, right? [00:50:02] Keith: You know, it's so interesting, Phil, 'cause um, I've, I've heard people say, and I'll, I'll spoil alert. I agree with this statement that like ai, even in its current state, is at the same scale of. Benefit, if you will, that the introduction of the internet was right, because with the internet we were connecting humanity and we were storing the collective knowledge of our species somewhere that, and obviously there's tons of problems to solve about access to the internet and, and increasing that connectivity. [00:50:33] But that was a paradigm shift, right? But now we've made that thing conversational. Right now we have something I, I can't, I wish I could give credit to the person who I saw this posted, but they were like, why aren't you using chat GPT? It has almost the entirety of human knowledge at its disposal. Ask it any question you want. [00:50:54] Ask it to double check its answer. Like gut, check it yourself. The same way when you read a news [00:51:00] headline, maybe you should check the source of that headline, right? But like it, it's all there. So like, just ask the darn question. 'cause worst case scenario, you get an answer you don't wanna use and now you know not to use that answer. [00:51:12] Phil: Hmm. Yeah, such a good point. I, I also love your product need fit that you kind of outlined there in, in the qualification process. Uh, maybe two last questions for you, Keith, like on this product need fit. I think that there's a nice transition to buying MarTech and [00:51:29] The Buying Process Isn’t a Spreadsheet. It’s a Cognitive Extraction Exercise. --- [00:51:29] Phil: something that you've mentioned in our pre-interview is like you've really honed in and become an expert at buying. [00:51:35] MarTech and rev tag, and it's almost an art in and of itself, right? Like talk to us about what makes your buying process different, unique. What prompts or CTAs do you have members on the team filling out? Like walk us through that. [00:51:47] Keith: Yeah, I'm, I'm such a big fan of this process and I can't take full credit for, for it, for the record, I have been inspired by my colleagues and peers, um, who have helped me hone this practice over time. But, [00:52:00] um, I'll, you know, for the sake of time, I'll, I'll highlight one particular aspect of how I like to approach buying technology that I think is particularly fun because it's a. [00:52:08] Beautiful, I think intersection of AI and humanity and a really cool cognitive exercise to make sure that you are to my earlier point of like, you need that directional, uh, you know, directional guidance. You need to know where you're going so you can assess product need fit. Um, but what I, uh, what CTA that I have everyone do when they're involved in a buying cycle is. [00:52:32] I say, okay. First of all, we're gonna frame conceptually the problem that we're approaching here, right? Like what? What business problem are we trying to solve? Like, let's just make sure we all understand that in some really concise bullets. Once you have that, I'm gonna ask you from your perspective, whether you're a marketer, a sales person in rev ops systems, whatever have you, from your professional perspective, I'm gonna ask you to just spend 10, 15 minutes in a doc and give me your [00:53:00] must haves. [00:53:01] You're nice to haves, write 'em down for me. Okay, that sounds simple. That sounds routine, but this is where the fun part happens. I say, okay, once you've done that, once you've written it down and you probably, you might've been talking to yourself a little bit, you might've been just thinking silently. [00:53:14] While you're doing this, I now want you to record yourself narrating what you wrote down. Two things are gonna happen, at least two things. One, you're going to add context when you read what you wrote. Some other, even if it's just an extra 2% is gonna come out of your mouth. Or you may introduce something that didn't make it on paper the first time. [00:53:38] Right? And now I have two really rich sources of perspective. I have the, the, the more, uh. Static exercise of just write the thing down. And then I have the thing that came out of your brain while you were just like, thinking about it. Right. And, and reading what you yourself wrote. [00:53:57] Darrell: Like video record or [00:53:58] Keith: Yeah. Uh, [00:54:00] audio's fine. [00:54:00] Uh, I like video just 'cause you know, you, um, if you're on video right, you're gonna be a little bit more mindful of your inflections, of your things like that. But I, I think audio's fine too. It, it works either way, but. So I have everyone fill out a, a doc like that. It's, it's a template and then I have them record themselves narrating it, and then I take both of those and I give that to GPT and I say, okay, here's Phil's doc, here's Phil's video, here's Phil's background. [00:54:28] This is the role that Phil is, uh, is, uh, playing in this buying team. Go ahead and synthesize those two and also do that for everyone else who's in the same role as Phil, by the way. And then further synthesize all of that together to come out with a really cool list of requirements. Um, and that's what we then hold vendors and, uh, zero party and first party solutions against. [00:54:52] Phil: So cool. [00:54:53] Darrell: Yeah. That's why I love these interviews, like always learn something like a new way to do something that I never thought of. [00:55:00] Record yourself. That's great. Yeah. Yeah. [00:55:03] Keith: again, like I can't take credit for this, right? I, I was building a rubric for our very first evaluation in open ai and then one of our sales leaders slacks me and she says, Hey, uh, could I just like send you a video of, of me like talking about this? And I was like, uh, yeah, please do. And then that's when the light bulb happened. [00:55:22] Phil: Keith. Love that answer. Uh, I had a super short stint@atwordpress.com. [00:55:26] We are async written communication for a lot of stuff, and what I found sometimes is that you don't get the full context when you're just writing stuff. So really like your, your answer, Darryl happiness question. [00:55:38] Darrell: Yeah. Okay. So [00:55:39] You Can’t Go Hard Every Day and Expect to Stay Standing --- [00:55:39] Darrell: Keith, you're a GTM systems leader, an advisor, a well, a well traveled speaker. You're also a Lego fanatic, avid cyclist, severance super fan, and a newer runner. One question we ask everyone on the show is, how do you remain happy and successful in your career? How do you find balance between all things that you're working on while still staying happy? [00:55:59] Keith: Yeah, that's [00:56:00] a really, really wonderful question. So I'll say in terms of like, how do I make sure that I'm staying happy and I would actually say and fulfilled, uh, in my career. Uh, so two things. One, I need interesting problems to solve. I. Right. Um, most humans need some sort of stimulus, if not all, humans need some sort of stimuli to like keep them interested. [00:56:21] Right. Um, and it, it needs to be interesting to a certain degree. So I would say first and foremost, I just try to make sure that I have interesting problems to solve, that I feel rewarded for playing a role in solving. Right. Um, but I also have to make sure that I can have some degree of balance right. I know work-life balance is a, a hot button issue, and I'm not gonna try to go all the way down that route. [00:56:46] But, um, you know, you said I'm a newer runner, right? But I've also been an avid cyclist. That means I'm an endurance athlete. That means that I know what happens when you go too hard, too long, you crash and you [00:57:00] burn, and you can't do the, you can't get up the next day or get back on the bike or back on the road, right? [00:57:05] I'm gonna do a nine and a half mile run tomorrow. You know what that means? I'm not doing on Sunday. Running nine and a half miles again. So it's a balance for me of making sure that I'm interested, that I feel invest, I have a reason to invest, but also that I can push myself to a point where I have, I have time to pull back and recover. [00:57:24] Right. Um, and then in terms of like, how do I find that balance? It's just, it's intentionality. At the end of the day, I have to carve that out for myself. So, um, I'm a calendar freak. Um. I block my calendar off, right? Um, I, my calendar is blocked off until, uh, 7:00 AM Eastern and after 7:00 PM now I make it clear like, Hey, if you really need me, I'm not gonna say no, I never take a call after seven o'clock, right? [00:57:51] Um, I took a 7:00 PM call last night, but I also don't do that every day, right? So it's about knowing when to sprint, knowing when to jog, [00:58:00] knowing when to walk, and knowing when to lay your ass down and get some rest. [00:58:04] Phil: Awesome, awesome advice. Keith. This is interview has flown by. Really appreciate your time. Folks are gonna have to get a ton of value out of this. We'll link out to your, your LinkedIn. I know you've got some upcoming talks, uh, so folks should follow you. Thank you so much for your time, Keith. [00:58:19] Keith: Thank you for having me. It's been a privilege.