The AI Advertising Podcast: S2
Episode 8
AI, Automation, and the Real Edge in Marketing Ops
About This Episode
AI in marketing automation has reached an inflection point. The tools are powerful, budgets are rising, and AI now touches everything from campaign planning to personalization. But outcomes are still inconsistent.
Chris Bentley | Director, Marketing Operations and Analytics, StackAdapt
Wayne Coburn | Director of Product, StackAdapt
Transcript
Diego Pineda (00:00:00)
AI in marketing automation has reached an inflection point. Advanced capabilities are widely available. Budgets are growing. And AI now supports everything from planning to personalization. Yet results remain uneven.
StackAdapt research points to a structural problem: 66% of marketers say up to 30% of programmatic budget is wasted due to siloed or fragmented execution. Adoption has moved faster than coordination.
AI in marketing automation no longer delivers advantage through features alone. Performance now comes from consolidation, integration, and activating AI as a shared intelligence layer across fewer, more powerful systems.
To unpack this, I’m joined by two StackAdapt leaders who sit on opposite sides of the same challenge: Chris Bentley, Director of Marketing Operations and Analytics, and Wayne Coburn, Director of Product.
You’ll hear them together across the episode, because the point isn’t “ops vs product.” The point is: automation breaks when teams run AI in silos. And it gets powerful when AI becomes a shared layer underneath the journey.
Podcast Intro (00:01:21)
Welcome to the AI Advertising Podcast, brought to you by StackAdapt. I’m your host, Diego Pineda. Get ready to dive into AI, Ads, and Aha moments.
Diego Pineda (00:01:35)
Let’s start at ground level. What are marketers missing right now? Wayne says it’s not that the ideas aren’t there—it’s that tool abundance and complexity is getting in the way of effectiveness.
Wayne Coburn (00:01:49)
There’s a lot of stuff people are missing. It’s just because the tools are kind of hard to use or there’s a lot of tools out there, but trying to to get them to be effective can be kind of hard. And where I see the most benefit is helping marketers figure out who to target. So the audiences they want to go after, and then coupled with the messages trying to go out to them. There is some stuff out there right now in the market that will allow you to do some of those things. But I think the really cool innovations are coming. And so it’s less about missing things because they’re not there and more about we’re sort of on the cusp of some really cool innovations that allow marketers to do their job, just be a lot more effective at their job, have much more efficient use of their budgets.
Diego Pineda (00:02:40)
Chris comes at it from the ops side. For most of his career, the word was automation — efficiency, workflows, moving data. But AI changes what automation can mean.
Chris Bentley (00:02:53)
I’ve been in marketing operations for over a decade now. And I’d say, you know, until the last few years, if people asked me about my job, I would have oriented on the word automation, right? That’s really the center of a lot of what’s going on. How do we just get efficient with our processes, you know, move, move data to and fro and and keep things running smoothly. Over the last few years with AI kind of changing the norm.
There’s probably two areas where I broadly see AI making the biggest impact in my work these days. First, you have a greatly improved ability to structure and categorize data. So before AI agents became the norm, when you wanted to enrich your data set, you know, you would have to go to kind of universal data sets that most companies want to use. And so think about company location, industry, employee count, things where there’s like a clear answer and then we can all kind of align on our definition. That’s always been a limiting factor, right? There’s a lot of things that a business cares about in terms of structuring and classifying their data that might be more unique to them and you’re not going to be able to go out and find a data set that already exists. Now, with agentic orchestration businesses can much more easily structure data in those unique ways. Take StackAdapt, for example, we care a lot about our customers’ advertising habits. And that’s something that you know we want to group our audiences by that. That just always isn’t available. With AI agents, you can start to build unique prompts and empower them with the research that allows you to get a pretty good idea of how people fit into your models.
Diego Pineda (00:04:36)
This is the pivot: AI isn’t just an execution engine, but an intelligence engine—that is, if you can structure inputs and connect outputs across the journey.
Diego Pineda (00:04:51)
When adoption moves faster than coordination, you get what I’ll call automation debt. It shows up as: underused capabilities, redundant tools, broken handoffs between paid and owned, and a system that can’t learn because outcomes aren’t connected back to decisions. Chris warns that AI can scale the wrong thing very fast.
Chris Bentley (00:05:16)
I think across the board the risk is bloat. So you have you know you start measuring every single signal that you can imagine, just hoping that you’re going to find something like you know throwing spaghetti at the wall and seeing what sticks. And then what happens is, you know, you start to ask AI to generate messaging at scale, like I mentioned for the sales team, but you’re not putting in the upfront work to actually provide enough context and framing to get a good output. And then what you’re happening, you know, then what you end up with is just a proliferation of data that is not particularly useful and starts to create a lot of noise rather than helping you find the signal.
Diego Pineda (00:05:55)
Wayne describes the human reality of that debt: when acquisition and retention operate separately, the customer experience fractures and performance suffers.
Wayne Coburn (00:06:06)
Ultimately, it’s all about the conversions at the end of the day, right? You have something that you are trying to do with your marketing automation. You have some sort of goal. You want people to buy something you want to generate a lead. And I would definitely be looking at those. I mean, those are lagging indicators, but that’s ultimately what you’re, what you’re trying to drive. And so I’d always keep an eye on that. I’d also try to figure out what some leading indicators are, like visits to websites and things like that, that you can say, okay, this is showing that things are healthy even before we have a conversion, monitor those. And if you see drop-offs, if you see issues, then you know that you have a problem. um If you find that you’ve made major automation more complex, but none of these numbers are changing, and then obviously you’ve done something that is not helping at all. Or if the numbers, like you make a change and the numbers go down, like that’s even worse. So always keep an eye on that. um I’d also try to make sure that my advertising and my retention marketing, my email marketing are connected in some way because you’re acquiring customers and what a lot of brands do, well, they’ll they’ll have an acquisition team, pay for advertising, acquire a bunch of customers, throw those customers over a wall to the retention team to go and try to retain them. And, you know, the retention team really has very little control over where these people are coming from. They don’t, they aren’t necessarily the people they want to be marketing to. And so you need to have really good connection between the acquisition and the retention to make sure that you’re retaining your your most valuable people and you’re growing people with the most potential growing the LTV and you know, whatever you’re trying to do is you’re getting, you’re doing it in a way that is intelligent. You’re pulling the right people through the funnel. And I think that’s where the biggest disconnect usually is, if you have this wall of, you know, so I’m advertising, someone shows up on a website, they fill out a form, great, I have an email address and now I’m starting to email them. I don’t necessarily know what ad took them to the website in the first place. Maybe I know, maybe I don’t, maybe I’m making some assumptions. And so now I’m hitting them with these emails that maybe have a different call to action and the whole thing falls apart and they don’t end up actually so you know signing up for a demo or or buying something.
Diego Pineda (00:08:28)
This is exactly why feature accumulation doesn’t create advantage anymore. If your stack is fragmented, you don’t get compounding learning. You get compounding complexity.
Diego Pineda (00:08:43)
In this episode, when we say marketing automation, we’re not talking about one channel. We’re talking about AI-driven orchestration across paid media and adjacent owned touchpoints. Wayne explains what orchestration looks like in real life.
Wayne Coburn (00:09:00)
So the orchestration piece is, you know, you have a customer journey or multiple customer journeys that you’re trying to get someone to walk down. And the most common example of this is a welcome series. So someone signs up for your service, you send them a welcome email, and then maybe a day later you send them another email. And then you see if they’ve logged in, um And if they’ve logged in, you send them email A and if they haven’t logged in, you send them email B to try to bring them back, et cetera. And so you have decision forks and trying to, you know, delays and forks and things like that where you’re trying to figure out what to do. I think some of the coolest things that are out there are going to help you route people intelligently through these orchestrations and say, okay, well, intelligently, instead of me having to say, if then else to determine whether or not I’m going to send someone Ad A or Ad B, let’s have the AI figure out what’s the propensity of someone going clicking on Ad A versus Ad B, and then showing them that one and taking them down that fork versus another fork.
And so I think there’s some really cool stuff out there that will help help us that decisioning and help tune these orchestrations to the individual person that’s moving through them as you layer in both programmatic and an email if you have it, and make sure that you’re giving everybody an experience that is tuned directly to them as opposed to something that’s based exclusively on some sort of logic and actions or inactions that people have taken.
Diego Pineda (00:10:32)
Chris brings it to the sales handoff: AI can package the context — research, touchpoints, product relevance — so sellers act with speed and precision.
Chris Bentley (00:10:44)
A large B2B marketing team is generating a lot of qualified leads for sales. And we have a vested interest in making sure that those leads get followed up with well and that our sellers are successful in reaching out. Providing context is how we boost that success rate. This has always been true, but I think with with AI agent steps, you you have a lot more power at your fingertips to be able to package up all the information you have, any company research that has been conducted, um you know all the summary of marketing touch points, as well as any product data for this particular customer segment, and provide all that to the seller so that they get all the prep work they need right at the moment that you’re asking them to work a lead. And in many cases, we can even provide them with sample messaging to make that first outreach easy.
Diego Pineda (00:11:29)
That’s orchestration: one journey, one intelligence layer, fewer gaps.
Diego Pineda (00:11:38)
Now here’s a subtle but critical distinction. The best AI automation today isn’t fully autonomous. It’s AI as an intelligence layer beneath the workflow—identifying what’s working, surfacing what isn’t, and guiding better decisions. Wayne puts it plainly.
Wayne Coburn (00:11:57)
AI can really help you identify problems that you need to solve. AI can help you with your visual data visualizations and show you what things are working, what things are not working. So it can help with the execution piece of it, but yeah, it can really help you in understanding what’s going on and give you hints on what you need to do to improve it. That’s a really good use of AI right now.
Diego Pineda (00:12:24)
Chris adds a key point: AI can reduce human bias in qualification, especially when marketers over-index on engagement and miss fit.
Chris Bentley (00:12:34)
When evaluating the quality of a lead, we all have bias, especially marketing, because we have a you know vested interest in qualifying lots of leads for a sales team. And I think typically when you’re scoring the lead, you are looking for some combination of the fit and their engagement. you know So how much do I believe this person is actually who we want to sell to And you know what’s the likelihood that they’re going to respond to us? Are they actually paying attention to us?
And I think marketers can over-index on engagement. We are, you know, excited to see leads downloading a white paper or registering for a webinar. And we want to say, hey, this person’s qualified, you know, get them get them moving in the funnel. When in reality, maybe they’re just learners, you know, maybe they’re just kind of interested in a more abstract way in what we’re doing. And, you know, a seller will tell you right away, this person’s not going to buy. So I think the benefit that AI can add in analyzing these signals is really reducing bias in qualifying for readiness.
Diego Pineda (00:13:44)
Let’s talk about consolidation. Because if AI becomes a shared intelligence layer, your stack can’t be twenty separate brains giving you twenty different answers. Chris talks about evaluating tools with discipline, looking for capabilities that truly move the needle, and being wary of big “one-stop AI” promises.
Chris Bentley (00:14:08)
I think for me and my personal evaluations of tools, I try to think small. We’re at a moment of great AI brainstorming where there are so many ideas of how AI can be applied in a tool. We’re not at a point where you’re getting really rock solid universal solutions. So I try and look, does this AI feature set unlock a new capability that’s really gonna move the needle for my work today? And if it is, and if I can justify that, then I’m comfortable you know layering it on. But I do try and be leery of you know complete overhauls with ah with a tool that promises it’s AI everything across the board and your one-stop solution.
Wayne Coburn (00:14:51)
I’ve talked to a couple different people from different companies and they talk about how in the race to AI-fi everything, they end up with multiple processes, multiple AI stacks inside their companies. And even with the knowledge bases being exactly the same, because the stacks are different, depending on where the employees go, they’re getting different answers. And that is a problem and they need to consolidate. And so I guess this is another place where going faster isn’t necessarily better if you have a lot of duplication. LMS and AI are inherently non-deterministic. And that’s part of it’s part of the beauty and part of what the power is. But because it’s non-deterministic, if you have multiple systems internally, all trying to do the same thing, giving you different answers, you know your employees are going to be like, I give up.
Diego Pineda (00:15:42)
Let’s make this actionable. Here are practical steps you can take in the next 30 to 90 days to reduce automation debt and turn AI into a shared intelligence layer.
Step 1. Map one end-to-end journey—paid to owned—and call out the handoffs. Pick one motion: programmatic ad, then a site visit, then a form fill, then nurture, then sales outreach. And just follow the baton. Where does it get dropped? Where does the next system have no idea what the previous one already learned?
Step 2. Pick a few signals that actually mean someone’s ready to buy. Don’t score every click. Look for a combo: it’s an ICP account, they’ve hit your pricing or demo page, and maybe you’re seeing competitor research. That’s buying. A single whitepaper download? That’s learning.”
Step 3. Since you can’t consolidate the whole stack in 90 days, consolidate decisions instead. If two systems are scoring or labeling leads, pick one owner and one source of truth. Decide: where does the official lead stage live? Where does the official score live? Then make every other tool consume that definition, not invent its own. You can keep multiple tools for now — just stop letting them disagree.”.
Step 4. Put AI where it can guide decisions, not just generate output. Use AI to highlight drop-offs, diagnose underperformance, and recommend focus areas—not just produce more creative.
Step 5. Do a weekly lead review with sales. What converted, what didn’t, and why? Then update the scoring, prompts, or routing rules so next week’s handoffs get sharper.
The main takeaway of this episode is this: AI in marketing automation no longer delivers advantage through features alone. Performance now comes from consolidation, integration, and activating AI as a shared intelligence layer across fewer, more powerful systems.
Podcast Outro (00:18:13)
Thanks for listening to this episode of The AI Advertising Podcast. This podcast is produced by StackAdapt. Visit us at stackadpat.com for more information about using AI in your advertising campaigns. If you liked what you heard, remember to subscribe, and we’ll see you next time.
You May Also Like
Report
GLOBAL
Jan 29, 2026
B2B Advertising Trends: What’s Powering Growth in 2026