Why Fabletics Chose Speed Over Perfection in AI

Breanna Lawlor: Your customers are getting better at ignoring you, not because your marketing is bad, but because they’ve seen it all before. The nudge, the reminder, the “We saved this for you,” they know the playbook and they’re well past it. For CMOs, this is not a future problem because the brands that are winning right now aren’t sending more. They’re communicating with a level of relevance that feels personal because it actually is, and the gap between doing this at scale and not is widening fast.

Adrian Rohr is the VP of CRM at Fabletics, where he owns email, SMS, and push across Fabletics, YTTY, and Scrubs. His mandate is simple and unforgiving: Make every message worth opening.

In this conversation, we get into how Fabletics built an AI content engine that generates cart abandonment messaging customized to each person at the moment of send, not from a template but built for them. We talk about the shift from personalization to what Adrian calls personal relevance, why trust is the hardest problem to solve when you hand decisions to AI, and what he means when he says the goal of AI is to become a daily operating system, not a tool people remember to use.

This is the CMO Club Podcast. Let’s get into it.

I’m really excited to chat with you today, and I’d love to hear a little bit about who you are- Yeah … the role you hold, and the company you work for, and then we’ll dig into the conversation. But thank you so much for speaking with me at The CMO Club, as we’re at the Activate summit hosted by Iterable.

Adrian Rohr: Thanks for having me. Yeah, my name is Adrian Rohr. I am the VP of CRM at Fabletics. My team oversees everything around email, SMS, and push notifications for our brands. Fabletics, Legs Scraps, and Yiti.

Breanna Lawlor: Fabulous.

Adrian Rohr: Yeah.

Breanna Lawlor: Fabletics has built quite a loyal following over the years, and I’m sure this has been a very thrilling place to be in.

And I have no doubts that AI has probably been an anchor point for a lot of conversations- … experiments, and also processes that you’ve established. Yes. Do you wanna speak to a little bit about sort of the company’s journey with AI, sort of your relationship with making that happen- … and what you’re finding as a result?

Adrian Rohr: Sure. I mean, AI is on everyone’s mind at the moment, right? Yeah. Like, I think for us it became, like, really, like, everyone felt it probably, like, last year. Right. Around that time, right? Yeah. Like, early in the year. We started really top-down and also bottoms-up from all angles.

We approached it, right?

Our CEO was very clear, like, it’s gonna be a big initiative for us, right? Yes. So there was this, like, perceived pressure of like, “Oh, now AI’s a big thing.”

Breanna Lawlor: It’s a religion.

Adrian Rohr: Yes. Unfor- like, unfortunately, or great. Like, you know, you can see it from both angles. I think it’s actually great because it motivates the entire company- Yes

to be after AI, right? It’s not just, like, one team that tries to keep pushing. It’s like everyone is gonna be on it, right? True. So there was this, like this top-down mandate, but also this bottom-up, like, curiosity and, like, motivation to try things out.

Breanna Lawlor: Great.

Adrian Rohr: Right? People started using, like, a million different tools, and, like, I feel like everyone was overwhelmed of, like, the amount of tools that we had.

Like, there was so much news about, like, all the tools that came out, things that you can do, image generation, content generation, workflow automation, like so many different things-

Breanna Lawlor: Yeah …

Adrian Rohr: that we all tried to get after. And then I think last year we started rolling out ChatGPT for the entire company, right?

That was, like, one of the moments when we really made it accessible to everyone.

Breanna Lawlor: Right.

Adrian Rohr: And people tried it out. People played around with it. People had success with it, but it was very much focused on individual use cases.

Breanna Lawlor: Yeah. Right?

Adrian Rohr: Like, the marketing team wrote some copy lines. The creative team made some image adjustments.

Like, it was not very operationalized throughout the company.

Breanna Lawlor: This is a theme I’ve heard, too.

Adrian Rohr: Yes. I guess it’s good, right? Because everyone just, like, gets motivated, like, tries things out. Yes. I think it’s this discovery, exploration phase that we need, right? Yes. I think it’s hard to just define a use case and go after because no one knows how it works.

There’s no playbook for AI.

Breanna Lawlor: No.

Adrian Rohr: And we just have to figure it out together, right?

Breanna Lawlor: Yes. That’s a really delicate place as we see CRM solutions in this, because you’re tasked with rolling this out. But you don’t exactly know what that will look like. Yeah. You have to have that buy-in secured from your CEO and from the team that’s doing the work.

They’re in the trenches. Yes. How are you navigating that? How are you gaining that trust or keeping it as a focal point?

Adrian Rohr: Yes.

Breanna Lawlor: And then also encouraging people and empowering them, “Try these things out. If you fail, it’s okay. We’ll learn from it.” Yeah. “But if you win, this is gonna be awesome because we’ll take it and run with it.”

Adrian Rohr: Yes.

Breanna Lawlor: How do you… That’s a delicate balance.

Adrian Rohr: It’s such a delicate balance, and it’s also, like, hard to, like, you know, meet the expectations of, like, okay, we’ve got to deliver something. Absolutely. But, like, we all don’t know how it works, right? Yeah. I’m, like, I’m very vulnerable in sharing with the team, like, I have no idea of what we create here.

I’m just, like, trying out myself. Yes. Like, you know, learning with you together, and as we create something, we share what we created.

Breanna Lawlor: Oh, yeah.

Adrian Rohr: Right? There’s a lot of, like, trial and error and, like, big wins and, like, big risks that we take. Yes. But in the end, we all are in the boat together because as I mentioned, there’s no playbook of how we apply it, right?

Yeah. So we as a company, again, we embrace AI a lot. We have, like, a central AI department that, that is managing, like, you know, all of our governance, security, the tools we wanna test.

Breanna Lawlor: That’s huge.

Adrian Rohr: During all company huddles, we celebrate AI wins. We have, like, AI Power Player Awards it’s called.

Nice. Every month we, like, give it to one person that did something great in AI. And then within the team I’m, like, you know, doing show-and-tell meetings where we just sit down, show what we just did in, like, you know, Claude or ChatGPT or, like, workflows that we built, so others can learn from it and how we share that.

Breanna Lawlor: And that is so valuable. Yeah. And I’m sure that also came up yesterday in the customer advisory boards. Yes. You’ve got a little bit of show-and-tell. Yeah. People may be sharing similar challenges- Yeah … but also solutions, and it seems that as we spend our time so much online as marketers, we’re really leaning into the peer-to-peer connection as a source of truth.

And also like, “Hey, you know, did you give this a try, and did it work for you, or did you know, did you have to make some adjustments there?” So what kind of themes and patterns came up in that conversation yesterday?

Adrian Rohr: I think the one thing that is, like, everyone has the same challenges. Yeah. Like, it’s not something uncommon that you try to run into, like, you know, your data set is not-

Breanna Lawlor: Right.

Adrian Rohr: Clean, right? You get generic responses that don’t work. Like, how do you really, like, deeply embed AI, and it’s not just, like, a use case that you run a chat in Claude or ChatGPT- Deep enough … and you get something out of it, right? How do you really deeply operationalize it? Yeah. And I think that’s where everyone is challenged.

Yes. But then there are tools out there that already, like, offer you AI decisioning and support in that journey. But what the common theme there was is, like, people lack the trust into the decisions that were being made.

Breanna Lawlor: Yes.

Adrian Rohr: Right? Like, humans often, especially, like, in CRM, like, we test so many different variations, like- Oh, yeah

subject lines, colors every single day.

Breanna Lawlor: Yes.

Adrian Rohr: But that knowledge is, like, a human knowledge. True. Like, people know what works. If you ask my team about, like, what’s the best performing subject line or, like, the best performing creative, they will tell you.

Breanna Lawlor: Yes.

Adrian Rohr: And they probably have, like, 90, 95% hit rate.

Breanna Lawlor: Wow.

Adrian Rohr: If you run this with AI, you might get the same results, but if you don’t have the visibility and the, like, transparency into, like, what actually drove certain decisions, that is hard.

And I think that was also one common theme that I saw yesterday during the client advisory board is that people, like, need trust in AI and decisions that it’s making.

Breanna Lawlor: It’s almost like reverse engineering the process. Like, sure- Yeah … you gave me the answer, but can you explain how you got there?

Adrian Rohr: Yes, because we only see, like, the result that, you know- Yeah

like, write a prompt, you get something back. Oh, wait, how do you get there?

Breanna Lawlor: How do you get there?

Adrian Rohr: Yeah. Right? And I think that flow is what people need to understand and have trust in.

Breanna Lawlor: Completely, yeah. Trust is coming up a lot, right? Yeah. ‘Cause we, like, humans are very adaptable, but they’re also wary of things they can’t trust.

There’s a fear element here. There’s curiosity. Like, there’s a lot of humanness in this whole adoption of AI. As far as the customer base that you have- … do you feel like you have a certain, there’s a little bit of wiggle room there because you have such a strong base of first-party data and loyal customers.

So it gives you liberty to experiment?

Adrian Rohr: Yes.

Breanna Lawlor: Yes? Okay.

Adrian Rohr: To me, what AI allows us to do- … is I brand it, like internally the, this theme of we move from personalization- Yes … to personal re- relevance.

Breanna Lawlor: Ooh, I like that.

Adrian Rohr: And what is the difference? Personalization is manual. You have like a, “Hi, first name.”

“Check out this product,” right? Yeah. And then the tool just like plugs in the variables that you have and fills it out. Personal relevance to me comes up with tools like a ChatGPT or Claude, where you write a prompt and you get something very personal back.

Breanna Lawlor: Yes.

Adrian Rohr: Right? To me, personal relevance is the context.

Breanna Lawlor: It is.

Adrian Rohr: So tools in the future in AI will allow us to customize messages to be truly one-on-one. It’s not like a customized field. Yeah. It’s a truly one-on-one messaging, right? So you have some customer data, you have some product data. Right. And the message gets created as it gets deployed. And it’s not a pre-filled template, it’s like one message that is truly individual for a customer- as in the message. And I think with the data that we have, that’s the true unlock where we can get to. Like a human can’t deploy a million messages individually.

Breanna Lawlor: Totally, it’s a volume component that-

Adrian Rohr: You gotta have a system that helps you with that. I strongly believe AI will be able to help us with that.

And what it means is we build stronger loyalty with our customers, stronger relations. Yeah. And with that also like a better brand reputation.

Breanna Lawlor: Yeah, absolutely. Yes. Thank you for sharing that, and also like a little bit of a glimpse into the future.

Right? ‘Cause it’s like-

Adrian Rohr: I think we’re already there a little bit.

Breanna Lawlor: Oh we are. It’s every day it’s a little bit further ahead. Yeah. Speaking of, if you had to retain a theme or a philosophy as far as how you’re approaching what you’re doing, what gets done.

Adrian Rohr: Yeah.

Breanna Lawlor: If you had to choose between speed, experimentation, or quality.

Choose one. Oh, that’s a tough one.

What’s gonna be your focal point this year in relation to AI and the company that you are, that you’re working within?

Adrian Rohr: That’s a tough one. I would say speed, because I think we all will learn. Yes. We all will fail. I think the one constraint I have for speed is we gotta have somewhat of control over it.

My, I’m always making fun of that. My risk tolerance is pretty high.

Breanna Lawlor: I bet.

Adrian Rohr: Right? Like, I’m okay to try things out. Yes. But I need to know that we are in control of what the outcome is. But I don’t wanna overthink it, because AI, I feel like it creates so much content, so much assets that it’s like- It’s volume

overwhelming. Overwhelming, right? And when you start getting after that-

…

Adrian Rohr: You lose, like, the actual speed.

Breanna Lawlor: Yes.

Adrian Rohr: Because you just keep thinking, thinking, and you don’t get your product.

Breanna Lawlor: Right.

Adrian Rohr: So I rather do something small and quick-

…

Adrian Rohr: And learn from that-

…

Adrian Rohr: So we can adjust our process to actually get to the great or better outcome.

So I think speed is, like, my theme.

Breanna Lawlor: Yes.

Adrian Rohr: I was debating between speed and experimentation- Sure … but I like speed.

Breanna Lawlor: Yeah. It’s interesting, ’cause this isn’t the time to be perfect.

Adrian Rohr: No, it’s definitely not.

Breanna Lawlor: No. It’s a time to really test things out and see what you get back. Yeah. And then move forward from that.

And you can’t afford to miss that right now. The window is sort of like things are changing under our feet, and- Yeah … if you’re not on this boat, you’re gonna be left behind. And then also that comes up a lot in themes. Like, there’s this- Fear element, so people are paralyzed because they don’t know what to do.

But it’s like you just pick one small thing to experiment with, see what happens as a result. For marketing leaders who are maybe not as far along in their AI adoption journey as you are, do you have any piece of advice, like anything that got shared yesterday that sort of helped them kind of get out of their own head and into the possibility versus the things that are maybe slowing them down or stopping them from trying this out?

Adrian Rohr: Yeah. I think the big thing for me is, and that’s what I’m personally working on- … we all wrote prompt in an LLM and got something very generic back.

And it felt like, well, that might be useful for us, but it’s not really useful. The big topic that we’re working on is just documentation-

Breanna Lawlor: Okay.

Adrian Rohr: And giving context.

Because as we start using those AI tools, the better context you can feed it-

Breanna Lawlor: Yes …

Adrian Rohr: the better the outcome will be for you, right?

Breanna Lawlor: Makes sense. Just like with a person.

Adrian Rohr: And it’s not a generic copy line or a generic content count that you can generate. Like without any context, right?

It’s the perfect product that fits your needs and your brand- … your language, your expectations- … your performance metrics, everything that is, you know, relevant to us as performance marketers-

Breanna Lawlor: Right …

Adrian Rohr: to really meet that outcome. So we work a lot on just documentation. I know there goes a lot of like, you know, how does our data look like?

Breanna Lawlor: Yeah.

Adrian Rohr: How do we even define our data set? What’s this immense layer for that? Right. Like, what’s the context? There’s a lot that goes into it.

And I think this can feel overwhelming, but starting like very simple of like, you know, what’s our brand?

Breanna Lawlor: Yeah.

Adrian Rohr: Who are our customer segments? How do we look at revenue attribution?

Breanna Lawlor: Yeah.

Adrian Rohr: What are our KPIs that are important for us, right? Just like having an understanding of that- … will help you and your teams-

Breanna Lawlor: Yes …

Adrian Rohr: like make AI useful for you in a much better way.

Breanna Lawlor: I wanna dig into one point you said. Yes. Understanding your audience- … and being really clear on what they want. Yes.

Because if you have an idea of the outcome you want, then you can leverage any tool to get to that point. But if you don’t know what your customers need, how are you gaining those insights on what your audience wants? And on a consistent basis so that you know it’s real, it’s provable, it’s data you can do something with.

Adrian Rohr: Like various angles that you can approach it from, and I think none of them are related to AI. Okay. Nice. There’s a lot of like, like we have a VIP membership model- Yes … so we call this like member insights. Like, we run quality trackers, just ask customers- Yeah … like basic market research. Great. We understand who our customers are.

Now, at the same time, with having first-party data, like-

…

Adrian Rohr: Tons of that, right? Yes. We know exactly like when you come to the site, what you shopped, what you clicked- … what you’re interested in.

Breanna Lawlor: Oh, yeah.

Adrian Rohr: Do you cross shop your different product categories? Do you cross shop the different genders? We have all this data.

We have member insights, and I think what- Now AI allows us to just like use all those data points that we have and bring it together in one view, right? The AI allows you to connect the dots-

Breanna Lawlor: Yeah …

Adrian Rohr: and gives you a narrative of that. So the data foundation itself is not AI driven.

Breanna Lawlor: Right.

Adrian Rohr: It’s market research. It’s data that you collect, right? It’s like even CRM data you can use.

Breanna Lawlor: Yes.

Adrian Rohr: When do people open? When do people click? When are people most engaged?

Breanna Lawlor: Totally.

Adrian Rohr: What do people click on? That’s all data points that you can collect without any AI.

Breanna Lawlor: True.

Adrian Rohr: AI will help you though to connect those data points-

Breanna Lawlor: Yes.

Adrian Rohr: At a much larger scale than a human could ever do.

Breanna Lawlor: Yes, exactly.

Adrian Rohr: Right? I think that is where the unlock is coming from, is the volume that AI can help us to connect it-

Breanna Lawlor: Yeah …

Adrian Rohr: and generate insights, and I think that is one important learning for myself. You can use AI as a tool to just write a copy line.

Right. But to me, the true unlock is coming when you use AI as a strategic partner. Connecting all those data points is a tool, but then brainstorming, asking questions back like what does it mean? Can you pull like, you know, this additional data point in there? Yeah. How the narrative changes- Yes … and thinking through that, putting your human expertise on top of that.

That’s where AI really comes in as a great strategic partner.

Breanna Lawlor: Big time, and you’re moving into predictive analysis territory. Yes. So you can predict what your user- … are gonna want based on past behavior. And there’s this amalgamation of mass amounts of data- Yeah … to help you do that faster and arguably better than you ever could before.

For you in your role, given all of these possibilities, what are you really excited about over the next three, six, 12 months? Because things are moving at a rapid pace.

Adrian Rohr: Yeah.

Breanna Lawlor: Two-part question. And what impact or influence have you had over the last little while that you’re also really proud of that’s maybe kind of got you to where you are now- Yeah

and it’s gonna act as the foundation for moving forward?

Adrian Rohr: 12 months is a long time from, I feel like we don’t even know what new- I agree … what new LLM program is coming out next week. Yeah.

Breanna Lawlor: Right? You… “What’s your five-year plan?” Yeah. Like, no, what’s your three-month plan?

Adrian Rohr: Number one, I think it’s internal component is just like the team really adopting AI as- Yeah an operating system.

Breanna Lawlor: Okay.

Adrian Rohr: Right? That’s my goal. I want everyone to be on the forefront of AI and just have like spark ideas of how it can help us.

That’s number one, and that just comes from like trying it out and learning about it, right? True. Making it your daily companion. And it is definitely my goal that we roll this out as a tool that we use on a daily basis- Yes

but also as a tool that is embedded deeply in our workflows. I think one thing I am proud of, not just for myself, but as an entire team, is we built an AI-based content generation engine that feeds into our cart abandonment solution that is truly what I mentioned earlier, like generating this one-to-one message-

Breanna Lawlor: Yes.

Adrian Rohr: As we deploy this cart abandonment. So it’s not a, “Hey, name, you left this product in your cart.” True. True. “Check it out now.” Right? We all know this message. Yeah. Everyone gets it every single time that you browse a site. Like, we truly customize that message almost like as a retail store associate would tell you, “Hey, how did you like this product that you just tried on?”

You know? Yeah. And that is something that unlocked great revenue potential for us.

Breanna Lawlor: Already?

Adrian Rohr: A- already.

Breanna Lawlor: Wow.

Adrian Rohr: Already. Like, we implemented that like a year ago, and we saw an 8 to 10 cents per message incremental revenue

And rolling this out more and more over the next, you know, six to 12 months is definitely a goal.

Breanna Lawlor: Yes.

Adrian Rohr: Because that’s where AI can unlock-

…

Adrian Rohr: That relationship building in a much greater scale for us.

Breanna Lawlor: Lots of good stuff there.

Adrian Rohr: Lots of good stuff, yes.

Breanna Lawlor: Any points of concern? Especially given what’s possible and even what your rivals are doing, like how, what’s gonna be the key differentiator here when everyone has these tools available to do virtually the same thing?

What is going to be what distinguishes you and your brand, in your view?

Adrian Rohr: There’s lots of concerns. I think AI does great things for all of us at the moment, and we all are, like, excited about what it does.

Breanna Lawlor: Totally.

Adrian Rohr: But it’s changing customer behavior, and I don’t think we all are 100% just clear yet what it means for us.

Breanna Lawlor: No.

Adrian Rohr: I feel like this is like everyone’s gonna have agents and they’re gonna shop for us, like, but are we gonna have that already? No. Like, there’s- … sort of signs of it, right? Like, how does it change our customer behavior? I don’t know yet. I have an idea, maybe.

Breanna Lawlor: Yeah.

Adrian Rohr: But we all don’t know what it’s gonna be in the end.

I do think there’s definitely, like, some security concerns as well- … especially at the enterprise level that we are. Yes. You can’t just, like, have AI run fully, like, all anonymous in a way that makes decisions that you can’t control maybe. Also, like, when you use LLMs and tools like ChatGPT or Claude-

There’s a security risk that you just need to be able to control.

Breanna Lawlor: You do.

Adrian Rohr: Right? And then just, like, I think uncertainty and, like, pressure that you feel.

Breanna Lawlor: Yeah.

Adrian Rohr: It is still there, and I feel like we all feel, like, sometimes overwhelmed-

…

Adrian Rohr: By it, and I don’t want this to become, like, a burnout reason- No for people.

Breanna Lawlor: No, exactly.

Adrian Rohr: Just if you, like, really deepens into, like, AI every single day, I think you can feel it. So I’m definitely aware of that, and I don’t want this to become, like, a, you know, a threat to us over the next months.

Breanna Lawlor: I think that’s a really wise standpoint, and one thing that even came up in the keynote this morning is that trust sits at the center.

Like, Iterable’s rolling out all these amazing AI features. However, they’re cognizant of the fact that the people using the product want to have trust. They wanna know they can control the system, and they wanna know who can be responsible for using it and make changes as necessary. Yeah. They’re in kind of like this, like, the line in the sand of how we’re gonna use this, but what we need it to do in order to continue to pursue using it.

There needs to be the control aspect, which is really interesting to hear because it’s so important. Like, it’s gonna be really hard to walk things back if we just let AI run wild and we don’t have these security and, like, data governance in place to protect all of us.

Yeah. There’s so much there. It’s very exciting.

Adrian Rohr: And I think what, like, what I’m seeing is there are other tools out in the market that just- Yeah … like, keep launching AI features every single day, and you’re like- Yeah … “Oh my gosh, like, why are we not having this,” right?

Breanna Lawlor: Well, yeah, the FOMO comes in.

Adrian Rohr: There’s true FOMO. Like, why don’t we have this campaign builder? Why don’t we have, like, this segment builder and, like, all the shiny marketing materials you see around it?

Breanna Lawlor: Sure.

Adrian Rohr: But then when you look at it, you realize, okay, it’s just, like, another shiny object, right? Totally. So you keep chasing after something new. But I do appreciate Iterable’s approach of, you know- Yeah this trust, control, security.

Breanna Lawlor: Yes.

Adrian Rohr: Also, it was mentioned this morning, like, where you build a context layer-

…

Adrian Rohr: For the AI to understand what is your brand, what are your goals- … how do you operate as a company?

Because that’s really the foundation to truly generate something or have a workflow that’s, like, ca- tailored to you and your brand- Yeah

and delivers tangible results, right? So it’s not,

Breanna Lawlor: So many good things. And kudos to you for navigating that pressure, and also, like, the antidote to anxiety is action.

Adrian Rohr: Yes.

Breanna Lawlor: And so if you take one step, doesn’t matter how big it is, and you get by, and you kinda can rally the troops. Yeah. And get everyone to explore their learnings and share it, well, this is a really exciting phase to be in.

Adrian Rohr: 100%.

Breanna Lawlor: You’re always gonna learn something new. It’s just whether or not you have the wherewithal to try it out. Yes. And see what comes back.

Adrian Rohr: 100% agree.

Breanna Lawlor: Yeah. Adrian, is there anything else you wanna share? This has been such an insightful conversation, too, to hear at the enterprise level. Yeah.

You know, what you’re dealing with.

Adrian Rohr: I think nothing around there. I’m just like, you know, it’s always great to connect with, like, growth and CRM leaders here at Activate. This is, like, one of the highlights of the year. I think going to a conference is always a great moment. Yes. Learning from others, seeing what others do.

Breanna Lawlor: Yeah.

Adrian Rohr: Even if it’s just, like, a nod to, like, “Oh, we are doing this,” you know? Yes. It’s like your own reassurance. Like- Yeah. Don’t … It’s not just, like, learning something completely new. No. It’s just, like, reassurance for yourself. I really appreciate that. And having Activate in LA is also great because that’s my home turf, so.

Breanna Lawlor: Hey, right on.

Adrian Rohr: I love it here, yes.

Breanna Lawlor: So that’s awesome. It’s a beautiful city.

Adrian Rohr: Yes.

Breanna Lawlor: Well, thank you so much for speaking with me on The CMO Club. Like, you’ve shared so many insights that I think other leaders, and even people who are really just trying to get a read on this, they’re gonna be able to find this useful, and apply, and also just maybe change their philosophy in how they approach this.

But so much here that you shared. Thank you very much for sitting down with me.

Adrian Rohr: Thank you for having me.

Breanna Lawlor: Yeah. Great pleasure.

Adrian Rohr: It’s been a pleasure. Thank you.

Breanna Lawlor: Awesome.

Adrian Rohr: Cool.