Subscribe
Share
Share
Embed
Future-proofing the humans behind the tech. Follow Phil Gamache and Darrell Alfonso on their mission to help future-proof the humans behind the tech and have successful careers in the constantly expanding universe of martech.
[00:00:00]
[00:00:00] Phil: ops roles in organizations are increasingly about finding ways to use AI to automate tasks. Speed them up dramatically. Find ways to increase productivity for a bunch of people in the company, ops folks are the new resident AI tech implementation experts and are very well placed for this future of ai.
[00:00:19] What do you think?
[00:00:20] Darrell: I love it. A hundred percent agree. Uh, AI operations, I dunno what we should call it. AI Marketing Operations. A-I-G-T-M operations. Maybe?
[00:00:28] Phil: GTM engineering.
[00:00:30] Darrell: I know. No, you know my thoughts on that one
[00:01:00] Phil: What's up folks? Welcome to episode 1 68 of the Humans of MarTech podcast. Today, Darrell and I are chatting about AI's talent crunch marketing jobs on the brink and those set to thrive. And as episode we'll start with marketing jobs that AI could wipe out, like campaign ops, content creation and data analysis, marketing roles with strong AI survival potential like AI implementation experts, data orchestration and pipeline management.
[00:01:26] Product marketing, community building and change management, all that, and a bunch more stuff after a super quick word from two of our awesome partners.
[00:01:36] [00:02:00] [00:03:00]
[00:03:26] Phil: What's up everyone? And Darryl, we, uh, been itching to do a solo episode, just the two of us. Uh, since we got this thing going, we did one episode, uh, one 50 at the end of 2024, just the two of us. This is our first time without a guest. Uh, naturally you had a bunch of ideas for, for topics usually out when you're walking your dog, Stella, like, uh, hey, I got an idea for one episode.
[00:03:48] It's not, uh, too hard to come up with ideas with like all the shit that AI is changing today. So the topic we had is around Martech talent crunch. Who's a demand, who's at risk? So. [00:04:00] I wanna start off maybe with like the risk stuff. So what jobs in marketing, maybe a bit more Martech, but like marketing too.
[00:04:08] We can chat about like more generic marketing roles, what's more at risk, and then we'll chat about like what's more in demand. And maybe there's a bucket of just like, I don't know, like, uh, curious to get your take on some of these. But, um, yeah, you wanna start, uh,
[00:04:22] AI's Coming for Your Campaign Ops Job (Unless You Evolve Now)
---
[00:04:22] Phil: what's the first one that comes to mind for roles in marketing that are most at risk?
[00:04:27] Darrell: So I think the traditional campaign operations role is at risk. The one where it is these people are. And I kind of feel bad because, you know, my team is a bunch of, but I, you know, I, I'm gonna, I'm gonna come, I'm gonna swing around and, and, and say, say the bright side. But I think that the role of configuring marketing automation tools and email tools to spec right, meaning that we already know what the images should be.
[00:04:59] We already [00:05:00] know what the text should be, we already know what the call to action should be. And we already know where everything should go within the lead routing tools. I think configuring the systems to reflect that brief will be, I, I don't know if it will go away completely, but that one is definitely at risk.
[00:05:17] Primarily because AI and, and I think the things that we, we've been seeing can do the configuration for the marketer already. And you know, to be honest, there's been whispers of these tools. There's, there's already been tools that can kind of do this. Right. And I think with AI it's just gonna go crazy. I mean, I've, I've seen demos of, you know, what was once a dream where people just, where marketers just type in what they want and the marketing automation tool produce it for them.
[00:05:46] I've already seen demos of early, early stage, um, startups and, you know, add-on products that do this. Uh, so that's my first one. What is your first one?
[00:05:56] Phil: Um, yeah, I'll comment on the, the campaign ops first. [00:06:00] Uh, I, I actually had this one in like my unclear bucket and, and I, I know you've been meditating on this, like, uh, one of your recent posts was just like unpacking all of the tasks that are bundled into campaign ops. Like, I think when a lot of people think of campaign ops, like, Ooh, who's the person pressing the button in Marketo or the email tool to send out the actual email?
[00:06:20] Campaign ops is so much bigger than that. There's so many different tasks that are bundled on there, and I think your infographic like, we'll, we'll put it up on the screen for the folks on YouTube there. Like there's, there's so much involved. Um, I had it in like unclear because there's like two categories that I think are a bit more at risk and some that are.
[00:06:39] More likely to survive. So the more obvious ones, like more likely to be automated by ai. You kind of touched on them, like the, everything that's like reporting execution, campaign analysis, uh, like performance tracking, when we include like more, uh, paid stuff on the campaign front, like bid adjustments, uh, conversion rate optimization.
[00:06:59] [00:07:00] Everything in email. That's like automation, nurture, nurture flows, like using NAC and tools like that for like landing pages and forms. All of those things are getting way easier. Like there's already so much less hands-on stuff to do than there was like back in the day. I had it like unclear because there's a box of some of the things under there that I feel like.
[00:07:20] Are still pretty likely to survive. Like a lot of the automation and some of the tools that you mentioned, I've played around with them and not all of 'em kinda like hit the, the, the, the, the bullshit like stiff test or whatever. When it comes to like campaign office, one of the main rules that you had in that infographic was.
[00:07:38] Objectives and KPIs like coming up with what is gonna be the goal of this campaign. And I feel like AI can kind of help with that, but it requires like so much business understanding you need strategic alignment. Like there's a lot of battles politically about should we do this, should we not do that? I think that whole like objective goal setting thing is a strategic, like human part of the, the [00:08:00] process.
[00:08:00] There's the whole like, uh, value and, um. Like what's gonna be the main CTA that we decide for the campaign. There's a lot of creative and strategic thinking that's involved there too. Uh, another one I have in like the unclear bucket is budgeting. Like budgeting is a big part of campaign ops. There's a ton of cross-functional negotiation and strategic decisions there.
[00:08:22] So yeah, there's like a bunch of other stuff like qa, checklists, um, you know, best practices. You had that as one of the things too, like. You know, AI can help with some of the best practices, but it can only spit out stuff that is already very common when it comes to like new best practices, like insights from human experience, context judgment, like all that stuff is, is pretty good.
[00:08:45] Uh, AI replacement. But um, yeah, it's an
[00:08:47] Darrell: Well, I, no, no, I, I, I interesting that you put it in unclear. I think that, and, and I, I totally agree with the objectives, KPIs, like, uh, and the alignment. I will say that I think [00:09:00] that the specific role of campaign ops building to spec, and what I mean by building to spec, I mean. Yeah, they actually didn't participate in the brief creation and like the requirements, you know, they get a ticket and the
[00:09:14] ticket is full of everything that they need.
[00:09:17] That specific part of the role I think is gonna go away. And I think that what it's going to evolve into is more of what you're talking about where they're actually looking at the objective and KPIs and they're, they're actually changing the requirements and changing the brief and they're also like, you know.
[00:09:37] What I'm hoping marketers do more often are the concept of always on programs, like programs that are just, that are on continuously. And I think campaign ops is a great, is great to oversee and continually optimize those versus what we see most, many campaigns doing, uh, campaign ops, doing is doing like webinar after webinar, [00:10:00] event after event, and it's just like a net new idea each time.
[00:10:03] That also I think is like it. It'll never go away, but I think we're gonna see a lot less of that.
[00:10:09] Phil: Yeah. Yeah, no, that's a great point. So I think there's some elements of what you described. There are campaign ops that are, are pretty strong resistant to ai. And maybe we can say like, 'cause I actually had like basic campaign execution as one of the main things. At risk, like will you make like for paid stuff like top of funnel, like automated media buying, like mid bid adjustments, tactical implementations there, everything, marketing operations, like AB testing, execution, like all of those things.
[00:10:38] Maybe some strategic elements to that, but I think that. What you just described, like the strategic function of marketing ops, the stuff that's less like the hamster wheel, repeating new ideas, same repurposing execution there. Um, yeah, it's, it's an interesting one. It's funny that you say most of my team is, is campaign ops.
[00:10:57] Darrell: Well a lot. Yeah, a lot of my camp and that's why I wanted to say [00:11:00] my, the hopefully saving grace is that I think the roles are gonna change. I think we're a little bit for enterprise especially. I think we're a little bit further away for, from that kind of change. But I will say I do have high hopes for what the role will involve into.
[00:11:15] Phil: Yeah. Um,
[00:11:16] AI Will Eat Generic Content Creation (But Experts Will Thrive)
---
[00:11:16] Phil: the other one I had under most at risk is a bit of a layup here. I think this one's pretty obvious. Most folks will agree. I had like generic content creation bucket in there. Like I think a lot of folks are. I. Really excited about Gene AI and using gene AI to create like basic landing pages and not editing any of the text.
[00:11:36] Um, we're already seeing a ton of like, um, disruption with copywriters and content marketers who aren't very unique and are producing more generic content be replaced by AI roles. Um, I think like. Explaining it as more of like routine content creation without having a distinctive voice or even like cultural nuance.
[00:11:59] If you're [00:12:00] like a global org, that one's pretty obvious though, right?
[00:12:03] Darrell: Yeah, I, I, I think that bad content. Is going to become obsolete, like the creation of bad content's come obsolete. And what AI does is really just accelerate and amplify the, the bad content, like the crappy content, and what I think is still gonna be unique and rare. Is really good content, like good content that speaks to people.
[00:12:29] What might be interesting is, you know, the rise of subject matter experts, honestly, as content creators because they don't really have the skills or like the, the desire to spend all this time writing. I. But now, you know, and I'm sure like both you and I do this, we we're both content creators. It all, all it takes is like the spark of an idea and a few bullet points.
[00:12:52] And you, you have a full post and it's gonna be way better than someone, you know, like a, like a marketer for example, that doesn't really [00:13:00] care about the product or about the industry and is writing like crappy content. So a hundred percent I agree with the, the content creation one,
[00:13:06] Which Data Analyst Jobs Will Survive the AI Revolution
---
[00:13:06] Darrell: the other one that's at risk for me.
[00:13:09] Is the data analyst roles that are primarily pulling data. So I, you know, in the past have worked with a ton of people where they really just kind of ask you what you want. You know, Hey, what, what data do you need? What are your requirements? Again, the whole, whole like built to spec thing. Um, and I'm gonna make a report for you.
[00:13:34] And then they would bring, they would produce a report or dashboard and go like, here it is. What do you think? Do you like it? Do you not? What do you wanna change? So you're really just telling that person to do stuff for you and they're like an order taker, a glorified, highly paid order taker. Right? Um, that role I think is completely at risk.
[00:13:55] And I think that, um, when we get to the in on demand part, I'll, [00:14:00] I'll, I'll, I'll share like the, the flip the reverse part of the coin, but, um. With, with, with ai, you should be able to, given you have a good data set that's, uh, normalized and standardized, you should be able to just ask it to, you know, pull the data for you.
[00:14:16] And I think we're, we're seeing that already. What do you think?
[00:14:18] Phil: Yeah, a hundred percent agree. I had, uh, this one on my most at risk also. I called it, Hey, I wasn't sure what to call this bucket. Like in some of the companies I worked at, those roles were called analysts, and they actually sat on the data team and they worked with marketing ops and a lot of cases, a tiny marketing ops team.
[00:14:36] And so the analysts like owned the Looker, the BI tool, and they were the one just like, Hey. What, what report do you need? A new dashboard? And, and oftentimes it was us asking to put this new report on their roadmap. And so I agree, I think that is getting easier for like non-technical, non analysts to build.
[00:14:56] There's this whole like automated dashboard and performance [00:15:00] tracking idea to this. Uh, we can probably even like include some type of attribution reporting, uh, and basically like ROI stuff in this, my gut says that like. This is an easy one in a sense that it, it will maybe not fully go extent, but change big time.
[00:15:17] Like this whole idea of reporting. I think it's already massively shifting out of the marketing ops department into the data team department. Um, I think like. What you kind of called out, like the, the front end dashboard data viz part of, Hey, do you need this report and I'll create it for you is super easier already today what's changing is, um, NLP Natural Language Processing, allowing business users to interact with data, uh, have conversations with like bi chat bots, having a whole experience without having to download any data.
[00:15:51] You don't need to be a SQL expert. You can ask. Plain questions and receive plain English charts and ask the AI to like, [00:16:00] help you. Um, I, I have a client, a tiny little customer that I'm helping with. Uh, they, they're a big data company and they've invested in ThoughtSpot as, um. They're like, they have a really impressive search based interface with NLP that allows users to turn questions into a dashboard.
[00:16:17] I think they call it like spotter. And I've played around with it and it's really impressive, not perfect, like sometimes you're just like, ah, it wasn't exactly what I was looking for, but this whole thing of like someone coming to you and saying like, Hey, what do you need me to build for you today that's being replaced with you.
[00:16:32] Interacting with a chat bot that sits on top of your data warehouse or your BI tool and asking it questions and being like, what were sales in the last six months? What was traffic in the last six months? Basically like changing that whole category in a sense.
[00:16:47] Darrell: Totally. Totally. Yeah. I, I, I completely agree with that one. And, um, yeah, that natural language processing, I think I find that fascinating. And I think that, my guess is we're getting really [00:17:00] close already. You know, I, and I'm just speaking from like, just from personal experience, you know, I create diagrams on my, on LinkedIn and I have chat GPT, and I'm just using sample data, but I have chat.
[00:17:14] GPT continually iterate on the, on the, um. Visualizations. So I'll just say, uh, can you like, turn it into a horizontal chart instead? Can you add like different dimensions to it to like, so I'm, I'm, I'm talking to. Chat, GBT to do that. I don't think we're very far away from, you know, having companies like, uh, ThoughtSpot or others where you're actually just like continually just asking questions until it, until it gets it right.
[00:17:44] All of this, given you have the right data, so can't ignore that one.
[00:17:49] Phil: That, that's the beauty of this, this space being like bi dashboard tools that ideally sit on top of your warehouse. Like there's a couple of them, like I had a, a long stint earlier in my [00:18:00] career portfolio, which was a BI dashboard tool, and initially the vision was. Instead of sitting on top of the warehouse, we're gonna connect all of your sources.
[00:18:09] And so it basically just became like an API connector product that you could download data from GA and Facebook and product data, put it all in one spot, and then you could build VI visualizations on top. And there's other startups like polymer doing this. I think the future is when your BI tool sits on top of the warehouse, there's gonna be a whole new category of tools that instead of buying a BI tool like a Looker and spending millions of dollars on that.
[00:18:35] Our little startup can directly sit on top of your warehouse. And once we get to a stage where it understands the tables and like we get to a good structure, you can have a conversation with a bot about what exists in a data warehouse. Marketers don't need to know the structure, the table, the right term for this and for that.
[00:18:53] So I don't know. That's, that's exciting to me.
[00:18:56] Darrell: Yeah. Yeah. And then do you think that like it's going to be a new [00:19:00] startup that sits on top of these data, data warehouses, or is it going to be the companies like Snowflake or even the visualization companies that build this ai, they've gotta be thinking about this stuff, right?
[00:19:11] Phil: Yeah, that's a really interesting question. Like, uh, census for example, they're, they're trying to like pitch this like unified data layer idea where they sit on top of the warehouse and they have like a, a segment type of builder that. No code allows you to just have a UI and interact with what is in your data warehouse.
[00:19:31] Like marketers can quickly build segments that, you know, would take like hours to finish loading and Marketo, looking at like loading screens. But you can bring in like GPT columns and, and enrich that data so they're crossing into like what Clay is doing in, in like a separate side of things. So anyways, that, that whole area is, is really interesting.
[00:19:52] Darrell: And then do any of them do data normalization and hygiene? And, and if they don't, I feel like that's like [00:20:00] big, you know, could be a whole like prediction in itself of like, because AI is so important. And, uh, top of mind for everyone. I think we're gonna see huge increases in companies that just specialize in fixing people's data, whether it's consultancies or, you know, there's a handful of, um, what's it called?
[00:20:25] Data hygiene companies like, uh, RingLead, I, I'm forgetting now 'cause I, 'cause I'm, I'm, I haven't been doing it very often, but like, RingLead, I think, um, open Prize I think focuses on, you know, like I, I, I'm, those, those companies, like if the, if any of them are publicly traded, their stocks should be going like skyrocketing.
[00:20:45] Right? Or, I don't know, maybe I'm not seeing something.
[00:20:48] Phil: Yeah, there, there's like a lot of overlap with tools that are calling themselves, like enrichment focused versus, um, even like transforming data. Like DBT has a really interesting opportunity here. [00:21:00] Also, everyone is using DPT to transform data before it goes in the warehouse and after it gets out of the warehouse.
[00:21:08] Um, like no code ai, power data transformation. I think a lot of folks are gonna be investing in that, like even building. Workflows without reject, without like formula fields and being able to push that data and, and activate it somewhere.
[00:21:23] Darrell: Yeah,
[00:21:24] Phil: Cool. What else? Uh, what else do you got on? Um, most at risk,
[00:21:30] Darrell: I think that that's pretty much it. I'm trying to think if there's anything else that, um,
[00:21:36] Phil: I.
[00:21:37] Darrell: I prob that I probably wouldn't. Uh,
[00:21:40] Phil: Cool. We'll,
[00:21:41] Marketing Ops Will Shift to AI Implementation Experts
---
[00:21:41] Phil: we'll go to the areas with strong survival potential and, and if there's a couple that you could disagree with me there, we can, we can debate about putting it into the earlier category. Do you wanna start? Do you
[00:21:52] Darrell: Yeah, no, you start this time and then
[00:21:54] Phil: Um, so this one is an easy one, like lay up for, for most of the folks in the audience, I think they're [00:22:00] working in this area.
[00:22:01] I didn't know what to call it initially. It's probably just like a way to call it marketing operations, but I think specifically. The subtask of implementing AI tech. This one seems obvious, but you know, there needs to be someone at a big company that isn't just it. Who's responsible for being that like resident AI tech implementation person for go to market tools, specifically marketing sales, right?
[00:22:27] Like marketing ops folks who can automate tasks using AI tools. Are gonna be highly in demand. Um, when John Taylor and I, a couple years ago had an episode similar to this, um, we brought up a, a tweet that was really interesting from, uh, Pete Laya who, uh, thinks a lot about this space. He will put up the, the tweet on, on the video version of this.
[00:22:49] But he said, so ops roles in organizations are increasingly about finding ways to use AI to automate tasks. Speed them up dramatically. Find ways [00:23:00] to increase productivity for a bunch of people in the company, and he thinks ops folks are the new resident AI tech implementation experts and are very well placed for this future of ai.
[00:23:11] What do you think?
[00:23:12] Darrell: I love it. A hundred percent agree. Uh, AI operations, I dunno what we should call it. AI Marketing Operations. A-I-G-T-M operations. Maybe?
[00:23:20] Phil: GTM engineering.
[00:23:22] Darrell: I know. No, you know my thoughts on that one, [00:24:00] [00:25:00]
[00:25:39] I think that.
[00:25:40] These AI operators not only are gonna be responsible for the implementation, but also the training and enablement. I really think that because ops folks are more technical in nature, they're well positioned to teach other people how to do this stuff. And this is something that I'm finding like over and over [00:26:00] again of like marketers that are, you know, doing work.
[00:26:04] And I'll ask them like, have you. Done three different versions of this copy in chat, GPT or, or you know, in, in Claude or perplexity. They're like, oh, no. Or, or like, Hey, for your idea, for your call to action, or for your idea for your event, did you, did you go through 10 different versions? You know? And oftentimes the answer is no, they don't have it.
[00:26:26] I don't know if it's like a habit, you know, that, that ops people just have, but like I think that you and I are both the same. Where, where whenever we have a project, the first thing we do is start talking to chat GBT about it or, um, perplexity. Um, which one do you use all the time? All Is it Perplexity.
[00:26:43] Phil: Uh, I, I actually have both. I, I pay for, for, for chat, GBT. Uh, I have the pro license, but I also pay for anthropic code. I, I prefer some of the writing outputs, uh, a bit more on code, like when we convert the, the transcripts into like blog posts, passages, I. GBT is so [00:27:00] much more powerful. Like in terms of the whole scope of
[00:27:02] Darrell: Like the reasoning or,
[00:27:03] Phil: yeah.
[00:27:04] The reasoning is, but I use, uh, perplexity for like more search based stuff, um, like coming up with ideas. For this episode specifically, I used perplexity and it's funny, like I was gonna bring that up earlier. Like I, I posted about this last week, um, and researching like the future of marketing ops and ai, asking that to perplexity and finding me like legit sources for it.
[00:27:24] It pointed me to an article on LinkedIn. I forget this person's name is now, but I read through the article and I was like, ah, like not impressed with it. Like most of it was AI generated. And it's funny that an AI generated piece of content was one of the best sources surfaced up by an AI search engine.
[00:27:43] They're just like feeding each other, but.
[00:27:47] Data and API Services are the New Content
---
[00:27:47] Phil: This is the second one I want to call it. It's perfect transition here. Like Scott Brinker talks about this all the time. One of the areas that is really well positioned in this collaboration between marketing and AI is like a bucket of data plus [00:28:00] API services.
[00:28:01] So creating composable systems that integrate AI with your company's proprietary data. Also exposing data from your business to let AI assistance leverage them in a bunch of other workflows. And this is, this is something we're gonna talk about in a lot of upcoming episodes, but there's like this reckoning within software.
[00:28:22] Um, uh, rich Waldron and this CEO and founder of Trade Out AI talked about this in the episode that I did with him.
[00:28:29] Darrell: Hmm.
[00:28:29] Phil: we have it scheduled out like he's building iPASS, right? And he says that iPASS is the future of orchestrating AI agents. But then you have CDP tools that are saying that the CDP is the future of orchestrating AI agents.
[00:28:43] And then you have marketing automation platforms that are rolling out AI agents, and you have content like everyone is legitimately working on AI agents. Things in their Martech vendors at some point. And rich as this like marketing operations folks, um, like PEEP calls it [00:29:00] like residents, AI tech implementation experts.
[00:29:03] Rich is calling it more like AI referees for marketing tools. You're gonna be the referee deciding like. I'm not turning on the agent in that tool or that tool. I'm gonna turn it on in this tool because governance is better. There's a good like area for super proofing, like the, like hiding sensitive information, like confidential stuff that we don't want to, so we're gonna become like referees and, and implementation experts like that.
[00:29:28] That whole thing is changing really fast too.
[00:29:30] Darrell: I like that AI referees. That's a good one. I think that, so my on my list was like a role that's similar. I called it either Solutions Architect or data operations. And so I can't, maybe, maybe they're both separate roles, but when I think of data operations, I think of all the behind the scenes stuff that.
[00:29:53] Is required to make sure your data is set up and connected and formatted in the right way and ready to [00:30:00] use for AI capabilities. So like all of that is a job and whatever we, you're calling that data operations, what have you, I think that there's gonna be, that is gonna be an ind, demand in demand go-to needed role.
[00:30:15] Um, and it might be the most needed role in, in my opinion, um. That's like another one of my predictions, or solutions architect, which is like, Hey, I wanna connect these various tools that we're using and making sure the data flows between them. Um, and now does a solutions architect more of pick the tools or make sure that they work together perfectly?
[00:30:40] I don't really know. So, but I, I will say that, um, that if that is one role where like I'm intentionally selecting tools to accomplish a task. And making sure the data goes through together. That also is gonna be, I would say, the most in demand, you know, survivor role, you know, very, uh, [00:31:00] very sustainable job.
[00:31:01] Um, and I'm, I'm excited about it.
[00:31:04] Phil: I think part of that overlaps with like the API services data area. I actually had though in my like unclear, uh, bucket and I was curious to get your take on that.
[00:31:15] AI Can't Replace Human Orchestrators of Marketing Data
---
[00:31:15] Phil: This idea of like pipeline management specifically. I think that's like part of your answer there, like getting data ready and like formatting it so that you can surface it or whatever.
[00:31:25] Like there's like part of me that's unclear about it because like. ETL reverse ETL processes are becoming increasingly automated because of tools like census, like basic data cleaning and prep is so much easier, and there's like self-serve, no code tools to help you do that. I, I think that like, we'll see DBT get into the more marketing operations use cases and, but I don't know, like it, it's tricky because.
[00:31:52] It's easy for us, like non-data operations people or, or like, not data engineers to say like, oh yeah, like data pipeline management is, is [00:32:00] at risk to be automated by e AI because we're not deeply into that space. But the thoughts that I have are, data complexity continues to increase as like. You know, Martech stacks are growing more sophisticated complexity of data integrations, blah, blah, blah.
[00:32:15] So that's part of the side that makes me think, you know, there's like a bit of future proofing in there. Um, but there's like system maintenance that needs human attention when pipelines are breaking and stuff. Custom business logic remains like really tricky across like nuances and everyone's like a little bit different.
[00:32:34] I don't know. There's like this thing that makes me unsure about it because of how fast this space has changed and how easier things like, you know, collecting data, putting it into warehouse and sharing it across tools that like, I don't need to hire a team of data engineers to do reverse ETL for me anymore.
[00:32:53] I could have one person doing that with census. Like I, that's the part that like, makes it a bit unclear to me, you [00:33:00] know?
[00:33:00] Darrell: Yeah. So, yeah, so let's like unpack that a little bit. So I, I agree with you. I don't think that, you know, if you don't have like reverse ETL or these, um, what do they call it when. What do they call just normal ETLI, I dunno. Just ETL then,
[00:33:16] Phil: Yeah. ETL.
[00:33:17] Darrell: yeah, so if you don't have that, then you need a data engineer to actually write scripts and calls to pass from data from one system to another.
[00:33:28] And with the, the tools that are coming out, the platforms coming out today, yeah, you don't need that. So I, I completely agree with you. I think that for me, when you, when, when I'm, when I think about data pipeline. I do think that that's gonna be a necessary role. You know, it, it won't be a team, it'll be a necessary role because I do think someone needs to oversee the connection of, of, you know, the reverse ETL tool, like census and, and you know, connecting [00:34:00] it from all of the different activation tools or, you know, upstream data sources over to the data warehouse.
[00:34:07] I don't think it just. It's definitely not just plug and play, and it can be, it can maybe in the future become plug and play if there's some standardization. But like, so for example, this is the, this is the only way that it could go away is, you know, I think the, the listeners are familiar with, like Salesforce or Marketo.
[00:34:27] They have big app exchanges and with tools that connect to the app, uh, that are part of the app exchange. It literally plugs in. Like all you have to do is enter these tokens and these API calls and these secret keys. That doesn't require a person. You can actually have a marketer or whoever, you know, you're an an admin person, enter those keys and the tools are connected.
[00:34:50] But I don't think the data warehouses is like that. Like we're not even kind of close to that. Even like Snowflake, I think they're trying to go there. But you still need a person to [00:35:00] say like, okay, how are these gonna work together? What are the data points? You know, do I wish that they connected together like Marketo and Salesforce?
[00:35:08] You know, they just like, it's like a handshake and then boom, you're connected. I, I, I think we're still far away from that, and therefore, I think the data pipeline role is still gonna be critical, but it's not gonna be a whole team and it, it doesn't need to be an engineer.
[00:35:19] Phil: Yeah, yeah, totally agree. Uh,
[00:35:21] Product Marketing and Customer Marketing Are Extremely AI-Resistant
---
[00:35:21] Phil: the other one I had in like strong survival potential, maybe stepping a little bit out of, of marketing ops is product marketing and customer marketing. So in a lot of teams, customer marketing is like also lifecycle marketing, and sometimes they overlap a lot with, with marketing ops.
[00:35:39] I, I've worn both of those hats, but I think product marketing and customer marketing roles that require this like leap of faith concepts. Derive from customer empathy, like really knowing your customers, not just like asking them surveys and looking at like high level quantitative data. Like the qualitative people that [00:36:00] really sit down with customers.
[00:36:01] Understand the use cases like growth marketing that prioritizes experiments. Beyond what AI could potentially test. I, I like those areas of like future proofing beyond ai because like, you know, maybe at some point in the future with like a GI, it kinda replaces this, but the role of a product marketer, especially in like complicated technical, we sell to developers type products.
[00:36:27] That's hard for, for AI to replace like a marketer doing. Use case testing, building use case maps, like designing positioning that stands out, like speaks to people. That's really hard for AI to replace.
[00:36:40] Darrell: Yeah, that's a really good one. I actually didn't think of that, but now that you've mentioned it, I agree that that's a very, you know, critical role. I think it's going to be turbocharged by ai, so it's gonna be product marketing again, like, like the same way that, that we do content creation. [00:37:00] Like, I need to know, like the different ways to position this product in front of customers, in front of different segments of customers.
[00:37:08] You know, come up with a different, different message for each segment. You know, for enterprise, for small business, for like finance, for healthcare. Um, and the same with customer marketing, you know, uh, using AI to, to help accelerate that. I. The one that I had that I listed was strategy operations, which is, I, I'm working on a post about, and to me strategy operations is like the facilitation of GTM strategy from end to end.
[00:37:32] So like, you know, at a typical organization, everybody's kind of coming up with their own strategies, but they're rarely connected together and they're, they're rarely, um, done so, and like operationalized through and through. To actually the, the campaign. So in a typical org, especially at Enterprise, the leaders will come up with the plans and it'll be like filtered down and, and game of telephone changed all the way down to the [00:38:00] marketers who are like building the campaigns.
[00:38:01] And then all of a sudden you have campaigns that like literally don't even talk to each other. So I think strategy, operations of the facilitation of that. Ver not necessarily like coming up with a strategy itself, but making sure that it's implemented. So I'm, I'm bullish on that one. I'm also kind of biased because like that's kind of what my team does.
[00:38:19] But, um, anyway, that, that's, uh, that, that's, that's one that I had for sustainable.
[00:38:24] Phil: I, I love that one. Uh, that one's pretty similar to, like, it, I, I can see like a whole new. Like area focused on this. Like, I think like part of it is, is campaign ops. Like when I said like the, like what is the goal of this campaign, the kr, but you're saying like more high level company strategic stuff, like how do we align all of these different sub-teams, especially in, in bigger teams?
[00:38:48] Yeah. That's, that's a really good future proof one for sure. Um,
[00:38:52] AI-Proof Jobs in Marketing and Community Building
---
[00:38:52] Phil: the other one I had kind of out of the box here a little bit is. Community building and something that's like the, [00:39:00] the, the role, um, like we chatted with, uh, Mac Redden, um, his whole idea is building like the GTN network, like go to network movement.
[00:39:09] And he's getting like sales folks to think more about building relationships and community with prospects and people. And that's like so hard to have AI replace that. I think AI comes in and helps out with like. Community management, like having automation. When like someone asks a question, you make sure it's routed to the right person.
[00:39:28] But at the end of the day, like the whole point of community is building relationships between humans and maybe AI and robots will help with that facilitation itself. But at the end of the day. There needs to be a human thinking about the strategy and, and how we go to network and create those communities.
[00:39:47] Not always like inter overlapped with, with marketing operations, but there's always like a piece of technology involved in that. Like, do we just create like a Slack community? Are we creating a separate part of our site? People can log in. So that whole aspect, you know, [00:40:00] still ties into Martech, but yeah, I think community is, is feature proof by ai.
[00:40:04] Darrell: That's a good addition. Yeah. Yeah. Again, I, uh, I wish I thought of that one too, but no, I don't really have anything to add. The com community, I think, uh, you know, if I had to say another one is like brand probably. Um, I wrote an article one time and it was around how we're getting way too connected to revenue and the, the things that I mentioned that we, we, things have gotten lost because we're trying to attribute everything to revenue is things like community, is things like, um.
[00:40:32] Brand and um, um, product-led growth type activities, like things where like, uh, um, however, maybe AI is helping with a lot more with that, but brand and community, not so much. I.
[00:40:49] Phil: Yeah, that one's, that one's tricky.
[00:40:51] The Cultural Complexities of Global Marketing AI Cannot Solve
---
[00:40:51] Phil: the other one I had on my list is localization and. This one definitely overlaps with, with marketing ops. [00:41:00] So this idea of, um, you know, globalization is being able to speak. Culturally informed with different local areas and there's like an aspect of culture and understanding other countries and even like sub regions within those countries that is so human and it's changing constantly that, you know, like GPT models are, are based on a lot of older data.
[00:41:26] You can't trust. That it knows the new norms. Uh, like our kids are gonna grow up one day in high school and they're gonna come home with like these random terms and you know, they're just gonna like, brought up culture changes so fast, just like tech does. And I think localization, speaking to that market and not saying the wrong things, like being culturally appropriate.
[00:41:45] Um, oftentimes, I don't know, like if you've worked with like teams like that, but, um, I chat, I'm trying to think of her name, localization. Fuck, I keep spelling it wrong. [00:42:00] Okay, there we go. So I had Natalie Kelly on the show in November of last year, episode 1 47. And, um, she works at, uh, Zappy, but before she worked at Zappy, she spent, uh, what was it, eight years at HubSpot, and she was VP of marketing and one of her main focuses before that. And even as part of that role was.
[00:42:26] International ops and localization. And so when HubSpot expanded to a bunch of different countries, it wasn't as simple as just saying like, yeah, yeah, we're like, we're gonna offer it up to a bunch of like German cities and like Asia Pacific cities and countries. Like there's a whole like thought process and it's not just translation in translating.
[00:42:47] Language. It's translating concepts and ideas. Sometimes single words are totally lost in translation. So anyways, I thought that one was a really good one. That involves humans.[00:43:00]
[00:43:00] Darrell: So, okay. That's a really smart one. So I've got a few things to say about this. I actually did for almost a year, oversee the globalization team here at, at my, at my work. So, so the couple things that I will say is a hundred percent AI does not get you where you need to go. Like it can maybe do some slang or like.
[00:43:21] Cultural sayings, but there's just like the overall, like what's taboo within a certain culture that you like, you just shouldn't say, so it's so important. What I will say from overseeing this function is that it's hard to convince everyone else of its importance because a lot of people are very like their own country centric and. Um, one way that you can solve for localization, because lo localization assumes you're creating content or marketing initiatives. First in ano in a like us based, like it's a US based, kind of driven [00:44:00] campaign and then carrying it over to the other market. It, there's another way to do it, which is you hire a team in the other market and they do it from, they do their own thing.
[00:44:09] And that's what I kind of see a little bit more often. So I would actually bucket, even though globalization localization is important, I would probably bucket it in unclear, just depending on like, you know, number one, if the AI capabilities start to actually understand culture, which I think they, they kind of starting to are.
[00:44:27] And number two, you know, uh, depending on, um. Can we just solve for it by like using the local talent there? I, I don't know. I guess we'll see.
[00:44:37] Phil: Yeah.
[00:44:37] AI Bias Creates Demand for Human Ethics Guardians
---
[00:44:37] Phil: What do you think about ethics, privacy slash responsibility? Where, where would you bucket that?
[00:44:47] Darrell: I would, I would like, this is just a knee jerk response, but I would also put it in unclear because
[00:44:51] Phil: Yeah,
[00:44:52] Darrell: of the same reasons. It's so important I. But like, how often is it number one or number even number two on a, on the [00:45:00] priority list for executives? You know, it's, it's often an afterthought and often only after they get their wrist slapped or something, or they get sued, that's when it becomes a priority.
[00:45:10] Not like, it's not the first thing they think of. Um,
[00:45:13] Phil: The companies that paid attention to, like the, the shit storm, the PR shit storm that happened with the PGA couple years back when they used mid a, I don't know if it was Midjourney, but they used an an AI image generator to change. The background in the body, uh, like the clothes that some of the, the golfers were wearing and like all of the white dudes all had like the same like classic background, but the one person of color was wearing like a worker's construction outfit and it was like a, on a darker background and they just like posted it on social.
[00:45:51] It's still live on Instagram. I had a whole episode with, uh, Brittany Mueller and we were talking about this, like there's the whole area of like. Ethics, privacy and [00:46:00] responsibility with AI that I think is well positioned for the future because you need to have a human point of view of underrepresented groups when it comes to using ai.
[00:46:11] Most AI models are fed, I. From things like Wikipedia, like a lot of common sources that are predominantly biased by white males in their thirties. Like if you're interested to learn more about that, like Brittany Muer breaks it down really, really well. Um, and so yeah, I think that like this whole area of like, Hey, we're using ai, we're moving faster, we're coming up with cool content, that's great, but we need to have this like regulation piece, like this last check in the process of a human that says, okay.
[00:46:42] We're about to publish this, you know, we need to think about underrepresented groups from this one idea. Are we, you know, showcasing the POV from everyone here? Is this highly like white male biased here? Like are we including other point of views? So that angle makes me think like [00:47:00] that, that's a really interesting role.
[00:47:01] There needs to be someone on the marketing team that like, does that check before we go to market to avoid shit storms. The PGA, but also to like make your customer base feel included and like know that the marketing team behind this thought, like empathetically about this campaign.
[00:47:17] Darrell: Yeah. Very needed role. Very needed role.
[00:47:20] Um, so in the last few minutes, do you want to cover unclear? Did you have like other, I had like one unclear, uh, that I think we haven't talked about yet.
[00:47:27] Phil: Uh, yeah, I don't know if, I think we touched on all the ones that I had unclear. There was just budgeting, but I feel like budgeting kinda goes into campaign ops. The one that I wanted to get your take on was, I think we didn't touch this well, we kind of touched on it, but like,
[00:47:42] Why Change Management and Collaboration Will Survive When AI Eats Marketing Jobs
---
[00:47:42] Phil: this one is highly future-proofed in my change management and planning.
[00:47:48] I feel like that touches on like strategy ops a a little bit on, on what you talked about, but like. In bigger companies especially, and even startups like change management requires human empathy, [00:48:00] relationship building, understanding organizational psychology from one to the other. There's like collaboration and governance.
[00:48:07] You have to like negotiate with other humans. You have to find consensus, give and take project planning, scoping like. All of those things that are like behind the scenes in marketing off that don't get a lot of love, I think those become even more important because AI's never gonna have a, a, a meeting with like a counterpart in another department when you're negotiating about like picking that tool or that tool or this campaign or that campaign, like those are human change management things,
[00:48:36] Darrell: Oh yeah, that's like, I think the perfect one. Uh, you know, I don't, I don't know if you put it in unclear, but like, I may, maybe it should even be, yeah. Like, um, um, a survivor, I, there's so, there's so many examples where, um. People just need another person to talk to, to reassure them about what's happening.
[00:48:56] And there's, there's countless, you know, when I've done change management, there's [00:49:00] countless times where I've held a meeting telling people what changes are coming and all of the things that I've said, and all of their questions were already answered in an email or in like a wiki, but they just were just like, what about this?
[00:49:14] So you literally just have to be a person. Saying like, Hey, it's gonna be okay. Hey, don't worry about this. So a hundred percent change management. I think that that one's staying. I didn't have any other ones in unclear. I think we covered all of them. Um, but, uh, that's, that's a, that's a, a smart choice is change management for Yeah.
[00:49:31] A hundred percent survivor.
[00:49:33] Phil: Cool. This was, uh, super fun.
[00:49:35] Closing
---
[00:49:35] Phil: Yeah, so I'll, I'll close out with these thoughts. Uh, so in, in my research for this episode, I actually discovered a, um, the Marketing AI Institute. Um, not sure why I hadn't heard of them before, but their founder, Paul, uh, Rader, not sure how to pronounce that, but he has a brilliant post about the feature of work, uh, on LinkedIn.
[00:49:51] That was an inspiration from the book. Predictions, uh, prediction machines. Essentially he's talking about knowledge work and how it fundamentally involves [00:50:00] making predictions. Even though as humans, we don't think we're making predictions, but we're making a lot of guesses sometimes. In, in our work jobs are essentially bundles of tasks, many of which involve, uh, predicting outcomes based on data that we have right now, like medicine, law, marketing, um, primarily make predictions all the time, like in marketing.
[00:50:19] Stuff that even we think is complex, like e even back in the day, this is changing already, but like we're sending out a series of drip emails for marketing. We're predicting what the best subject line is, the best send time based on previous data. Like we're still making those predictions. Um, but Paul is essentially saying, and kind of like coins are our discussion here, uh, that humans remain essential in these like prediction centric roles for two critical reasons.
[00:50:45] One. We will always need to direct AI on what to predict. Uh, that's like where AI agents are making this a little bit fussy, but we still need to decide what it is, what's the thing that we're trying to predict, like setting the [00:51:00] goal and the task itself. But two, we also need to use judgment to decide what actions to take.
[00:51:07] Based on the predictions and the recommendations that we get, and oftentimes that's where like the fear comes with letting AI agents loose and go off autonomously and do a bunch of stuff for you. Like do you trust that their recommendation is always the right thing to do? So. This is really interesting 'cause it creates like this new paradigm where AI excels at generating predictions from data.
[00:51:30] But human expertise and intuition is always, or at least in the short term, gonna remain vital for guiding AI and LLMs and AI agents in converting its outputs into effective decisions and actions. Based on your company, based on the nuance, based on the context that you know of that company, instead of just applying it blankly.
[00:51:52] When evaluating AI's impact on jobs, which is what we've kind of done this whole episode, the key is to break down [00:52:00] roles into tasks, understand all the different subtasks that come into that, and identify which of them are prediction based, and recognize that AI can absolutely handle some of these predictions.
[00:52:12] The humans of Martech are always gonna provide this essential framework and judgment to make those predictions useful. We'll catch you guys next time.