About the speaker

As the CEO of Theia Strategies, a technology integration agency, Lucas helps companies that provide their customers with tailored solutions to optimize their marketing and sales programs. With over 10 years of experience in the oil and gas industry, he’s developed a unique perspective and a proven track record of solving complex problems, building effective teams, and managing interpersonal relationships.

Mark Lerner (Host): So why don’t you tell me a little bit about the agency and how you came to be?

Lucas: So we started these strategies about five years ago with the intention of becoming a marketing company to be that entire back office for someone who needs a marketing problem.

In the process of doing that, we learned that a lot of customers’ data and infrastructure were so noisy that they couldn’t get to their end objectives.

So we ended up doing a lot of data cleanups, a lot of migrations, and implementing new systems for them so that they could actually build an effective marketing strategy and campaigns 

They had a sales system over here where the right hand didn’t know what the left hand was doing,  so it created a lot of challenges as far as building an effective marketing strategy. Ultimately, it’s not an effective growth strategy.

Mark: How did things shift for your agency and customer base over the last few years, especially heading into COVID and the shifts that happened around that?

Lucas: What changed was that it upended how every business functions. If you were B2C and your goal was to help people with that – B2C but in-person relationship, which was a lot of our initial clients – your business dissolved overnight. You had to restructure and figure out how to sell your business. 

This brings up a case study where we had a wedding photographer that we work with in New York, and his competitors saw an 80% drop in traffic – search was 80% down – we tripled his business that year just because we paid attention to what were customers wanted, how they wanted to interact with his information and bringing in good data when they were coming in online. 

So we started tracking all of those metrics and building a really good data model, and when you get to the bigger scale clients, it ended up being that you would have marketing over here that was doing one thing and say, I was over here doing another and we had this business function in the past – business analytics and business intelligence – and that function was great but what it didn’t account for or didn’t do is have any authority to make decisions. 

The business analysts say “Hey, marketing here’s what’s working” or “Sales, this is what’s working,” but they didn’t work together and there was no line up to leadership to make an effective change. 

Mark: They didn’t have a seat at the table.

Lucas: They didn’t have a voice; they had nothing. 

And now, I think the reason why we’re seeing so much – in RevOps and now AI – of this data machine is that we have to understand what’s going on, and we’re no longer in this space where we can wait and see. AI, overnight, changed how quickly we have to react and if you don’t have a good data mode. lf you don’t know what your information is across the entire business, you can’t make an informed decision fast enough. 

I subscribe to three different business models: You can either be an innovator, you can be a fast follower, or you can be a late adopter. 

An innovator you’re a startup, you’re coming up with something that nobody’s ever seen; AI’s the innovator. You can be a fast follower which is “Hey I came out with this how do I adopt it today” and that’s a lot of the people we’re going to see here today. 

But this week at inbound is, yeah, fast follow a lot of the companies are smaller they’re able to move quickly, and they’re not going to wait. 

Then, you have the late adopters who wait until it’s matured and it’s in place they’re not the late adopters anymore; the Microsoft’s, those big companies are saying we need to be on this AI bandwagon now, right, so we’ve lost this gap in the market to get things going we really have to get everybody moving now and if we don’t move quickly enough we’re going to lose you’re going to lose your business you’re going to lose to the competitor because somebody else filled that gap before you decided.

Mark: You bring up a good point. So we’re here at inbound 2023 and that brings up the overall HubSpot ecosystem one that’s obviously growing considerably. 

Over the last few years HubSpot has done very well, it seems like there are more and more people who I think at one time would have, you know, kind of said there was only one CRM they would ever work with, but there’s a lot of those folks starting to move over into the HubSpot ecosystem. 

So, from the perspective of Revenue Operations, Business Systems, that kind of the data working getting all the systems working together, in the work that you do what are some of the challenges that you’ve seen with customers and have you seen any ways in which HubSpot product and ecosystem has evolved to meet the meet those needs?

Lucas: I mean, as far as HubSpot goes in the last several years, the depth of the platform that they put in place has really made it a very strong competitor to a lot of the plugins that you would bolt into Salesforce, right; Salesforce is a sales tool, it’s got it in the name; everybody knows Salesforce, everybody has to have Salesforce. and I don’t say this saying Salesforce is going the way the dinosaur, but it’s a very old clunky heavy system it does a lot that HubSpot can’t today.

But, in six months, a year, in five years, if Salesforce isn’t as nimble as HubSpot is – and Hubspot’s made a tremendous stride in the depth of the platform to where we have clients that we’ve pulled out most of their Salesforce tech stack, they maintain Salesforce for some data compliance issues and reasons but the entire marketing services the sales outreach functions all of that is managed through HubSpot, right? Because it’s just more efficient for the team to use. 

I can train a team to use HubSpot in five weeks to get them up and running and they’re confident in the platform. Yeah, we’ve trained people in Salesforce in those five weeks people are still like “Where do I go find this?” 

And that’s not just my organization. Those are my peers as well struggling with similar challenges time-to-functional is five weeks with HubSpot and the cost of my onboarding services is marginal where your time-to-live with Salesforce how complicated do you want this to look right you could be six months or a year before you stand up Salesforce. 

So having platforms that are nimble and can adapt to your workforce not only does that nimbleness that HubSpot gives you the ability to adapt your business process your pipelines your sales motions all of those things with a small team you can make changes quickly right but Salesforce Marketo the big monsters you need a dedicated team to make that happen and it’s going to take time.

Mark: I wanted to swing back to one of the things you mentioned: AI has been a huge thing over the last however long. Obviously, it existed well before ChaGPT kind of blew on the scene, but I think most non-techy people- myself included – didn’t really know what Open AI was until November 2022, like most people. AI is not only language models, I think the thing that people talk about is ChatGPT and other large language models, but do you see the kind of openness that even today during partner day when the CEO talked in a meeting laying out a very kind of clear AI strategy, but do you see that from do you see kind of additional AI tools being involved in the work – the data processing.

Lucas: I’m pretty passionate about AI yeah from the standpoint of like we call it Artificial Intelligence right in my mind intelligence implies it has creativity it can come up with something new, okay it’s at its core it’s fuzzy logic heuristics and statistical analysis of data whether that be in the form of an image and creating custom graphics with mid-journey uh data with open AI and ChatGPT or Bard or other systems that are doing it in the healthcare space and make the list goes on and on yeah the technology behind that’s been around actually since Alan Turing invented the first computer right he came up with this process we just now have the ability to process the data that’s a long-winded say way of saying AI has been here forever we’re just getting much more effective with it and that’s the risk that I see coming to market right it’s a risk reward no, Skynet’s not coming to kill us and anything like that, but it’s a different podcast okay that being said it is going to change how everybody gets everything done yes you are no longer going to be able to be an A player B player or C player; AI is going to take the B and C jobs.

The A players are going to be able to plug data in and get information out, so they don’t need the researchers the base get knowledge gathering skills that typically you would have yes so a good example would be a paralegal, a lawyer, an artist is going to argue the case the paralegal is going to do all the research well the paralegal is not going to be necessarty 

The state of New York just passed a thing – or some state, I forget – a lawyer submitted a brief that was completely done by AI reference legal cases that didn’t even exist. You’re still going to need someone with that creativity and that knowledge to know what’s real and not yeah, but it’s going to change how we work yeah, and every interaction you have, you’re going to be questioning like today we get emails half the emails we send are from a workflow or an automation or some recipe that does it without human intervention. 

Mark: Yes, we’re not going to be questioning, “Was this written by a human?” Yeah, I mean, I think—I don’t know—I’m already there. I don’t know about you, but there are certain tells I have already. Yeah, and when you get good at finding them, there it is. In summary, or something like that, I’m like, “I know that that’s a tell.” 

Lucas: Or like the images, if the shadows are off or there are extra fingers, yeah, I mean, the AI and the images. 

But that’s today; these are still in their infancy. Yeah, think about the growth they’ve had. 

I mean, the language models, these machine learning algorithms, really kind of started at, say, five or ten years ago, and like, they were machine learning. That’s what they were called when I was in engineering. 

Right now, they’re AI. Right? Where are they going to be in five years? I 

Mark: There’s a concept where technology doubles every time it improves. Well, there’s the—what’s his name—the guy from the chips, um, the Intel dude. Yeah, yeah. 

But I think what you’re talking about is an even faster kind of model. 

But yeah, it grows so much faster than even tech development. Right? Because we don’t need the time to build infrastructure, to manufacture. Right? It’s a few lines of code and tweaking and learning. Right? 

So, I guess my question is more along the lines of predictive analytics and how that plays into it because I guess that would be—that would fall under—I would call that AI. 

I don’t think that has hit the mainstream the way, you know, ChatGPT and generative AI and language models have become in the mainstream discussion. But do you think that that’s coming? 

Lucas: From the data perspective, I don’t think it couldn’t. Right? I mean, I hate to pull Hollywood references but think about every movie that’s sci-fi. The computer’s telling you something’s failed in your system, when your spaceship, your communicator, or something. Right? That statistical analysis of information—we’re not far off of that. 

And really, as humans, those data analysts I talked about in the beginning that created rev ops, right? They’re going to become the AI masters, right? 

They’re going to feed the machine the information to create those answers, those responses. And until we—yeah, there’s a grappling here of ethics as well, okay? Individuals have Tesla self-driving cars. Yes. And there are TED Talks, some individuals, you know, TED talks and things that go down this, over and over again. What is the value of a human life? 

Mark: Right, so, you’re basically playing out a philosophical debate that has been going on for centuries and letting the machine make the decision. 

The trolley car problem, exactly, over and over and over again. 

Lucas: But now we’re leaving that to a computer to make that calculation. Right? And we’re feeding information to it. 

At what point do people decide to pull that button away and say, “You or I hit the red button to shut down the system?” Right? When do we pull that away? And at what point… It’s going to start in our world; it’s going to start with marketing and sales. The marketer is going to say, “It’s good enough. Let it go.” Yeah, we moved quick; we don’t—the repercussions of a failure in our space are very, very minor. You hit engineering, you hit self-driving cars, you hit these more complicated systems.

Right? The repercussion of the failure is much more significant. And so, defining how we use these in our world, creating that environment that we can utilize them safely and effectively, and not compromise, do you see? 

Mark: Like, okay, so obviously, like you said, in a sales and marketing perspective, using AI the wrong way is probably not going to cause, you know, all the nukes to be fired or whatever. 

Lucas: Yeah, no ‘War Games’ movie, right? 

Mark: Yeah, there’s no thermonuclear warfare happening as a result of sending a completely automated workflow.

But, do you see a hesitancy in a lot of people or maybe even just kind of waving it off as though it’s just another trend? 

Lucas: I mean, I can tell you from my internal staff, there is a hesitancy. 

Yeah, I have some staff that will use AI models to help them do their job more effectively. Yeah, and they’re bought and sold on it. I have other staff that up and quit when we said you need to use this, because they were so strongly polarized against the fact that it could take their job.

Well, yeah, I think that’s a challenge. 

Mark: I think those kind of two things are playing out a lot, and I would imagine I haven’t really had the chance to check the pulse of the audience in the event, but I suspect there’s that dichotomy is playing out at scale.

Lucas: Well, I mean, earlier today, we had Jen and Jasper from Jeeves talking. They have a generative AI platform. Yes, and she put it very succinctly: ‘Do you have original thought? If you don’t have original thought, AI is going to take what you have. If you have original thought…

Mark: It’s a deep question; do any of us have original thought? or are we just kind of regurgitating a collection of our inputs?

Lucas: Fair, but that still provides an opinion. And what I think she said best of the whole thing was it raises our value as a human. If you are an original thinker, your value, your stock goes up. Yes, because AI can’t do that. If you don’t have original thought, then you’ve got to figure out, ‘Are you the person that fixes the robot, or are you the person that sits on your couch because the robot took your job? 

Mark: Or do you get plugged into the Matrix.

Lucas: Well, I’m thinking, however, the manufacturing robots that took the auto workers’ jobs away… yes, tons of their jobs away. But it facilitated tons of new jobs, right? That’s the argument, right? 

That even if it takes away some jobs, additional new jobs that we can’t even really fathom will become available. So like, we don’t know what AI is going to recreate.

Right, right. But we’re at such an infancy of AI that we don’t know what it’s going to create. And kind of going back to the beginning of the conversation about the data piece, we’re stewards of this machine right now. The data that we feed it is what it learns to create its predictions. If we feed it garbage, we’re going to get Skynet. 

Mark: Unfortunately, there’s a lot of garbage I’m sure it’s been fed.

Lucas: Oh, I’m sure there is. But if we only feed it gems, it’s not going to sound real. Right, so we have to feed it… So, we just need to figure out how to help it learn what’s a gem and what’s garbage. Because my kids have used it, and it’s asked questions about different science facts for school. I’m a science nerd. I’ve read the facts. I’m like, ‘That’s wrong. That’s a lie.’ 

Mark: Right. It tells lies very well. 

Lucas: But they didn’t know that.

Mark: Exactly.

Lucas: And so we got to figure out how to go from four, which will lie to us where we believe it very, very well, to five or ten or whatever it is, that’s truthful. 

Mark: Yeah.

I think it’s interesting, as this wasn’t the direction I expected the conversation to go, but I think it’s probably the direction that most conversations go this week. Because it’s such a… it’s kind of blown up everyone’s world over the last few months, six months, a year.

Lucas: And they kind of go back to the point you started earlier when we started, in the pre-roll, ‘How does this influence RevOps? How does this influence the revenue, the structure of your business?’

It’s going to be that core data architecture. Making sure you have good information is going to be the initial backbone of any AI that is layered on top of it. 

Right. I think in her message today, the CEO (of HubSpot was talking about their AI, and she said content, context, and customer were part of their… Right. 

And so the context, that’s the data that the AI uses as its baseline. 

So yeah, “Garbage in, garbage out.” The infrastructure, the data that you feed it, is going to be… You know, the output is only going to be as good as what you put in.

Lucas: Yeah. And like, so there is that side. But you also have to run data analytics and reporting and making preemptive decisions. That information coming in is going to help rev ops. What is this leading indicator? Do I have enough deals in my pipeline? Am I doing what I need to do to generate effective business? Or am I not? You’re going to have those indicators.

Right. But if you have one deal a year, you’re going to get garbage. Right? One point doesn’t make a line. If you have a hundred points a year, maybe you’re getting enough data to have a meaningful output. Like, we have yet to find what is that threshold of data in to mean that we’re getting quality out. 

Mark: Like, what’s the sample size that’s big enough.

So, I guess to kind of wrap things up, I wanted to ask about what you’re looking forward to most. I’m pointing over there because I think that’s the way the convention center is. Over the next few days, what are some of the speakers or events that you’re looking forward to the most?

Lucas: Really, I like the keynote speakers and what they’re talking about. Yeah. I’m kind of privileged this year. I’ve got a booth, and it’s going to be right by the stage, so I get to see both. The speakers, and our data partner that we’re actually sharing a booth with, Syncari. We’re kind of helping champion this AI, data quality, and integration of your tech stack. Yeah, so you get to be in the driver’s seat of helping decide what is that ethical decision? Interesting.

And then also hear what the other thought leaders in the space are going for. 

Mark: Yeah. Yeah, I think I’m very much looking forward to just kind of getting a pulse check on the way people are feeling about some of the conversations we’re having here, just based on kind of seeing online discussions. But I’m actually interested to see how much of that dichotomy we were talking about plays out in real life over the next few days.

I’m gonna find out tomorrow. I mean, I’m by the stage, and our banner says “AI get ready for the future”