How To Earn Strategic Influence in Marketing

Episode Transcript

If you had to guess, how much time has your team spent in the last six months talking about AI? The tools, the use cases, the strategy, the governance, the panic — all of it. Probably, probably, a lot, right?

So follow-up question: how much of that conversation was actually about what work matters most, in what order, and why? Now, this is not just a semantic question. It's actually really important, because the teams getting the most out of AI are not the ones with the biggest budgets or the most sophisticated tech stacks. They're the ones who already knew what they were trying to do before AI even showed up.

Now, that might sound obvious, but it is not apparently obvious, because the number of marketing teams trying to bolt AI onto chaotic, reactive, backlog-free operations is nothing short of alarming. And I'm saying this with complete empathy, because the pressure to just do something with AI is real, and it is loud, and it is coming from leadership and competitors, and it is probably coming from someone in your Slack instance right now.

But here's what I wanna talk about today: why Agile marketing is not just compatible with AI, but it is the actual operating system that makes AI in marketing work. And more importantly, why your AI advantage has a whole lot less to do with which tools you're using and a whole lot more to do with whether your team has a functioning backlog. So let's get going.

Welcome to the Agile Marketing Edge, the first podcast dedicated to turning Agile theory into real world marketing breakthroughs. I'm Andrea Fryrear, CEO of Agile Sherpas and your guide on this climb to smarter, faster, outcome-driven marketing. Every week, we dig into the what, who, and how behind Agile marketing — from building high velocity workflows and cutting waste, to measuring what actually matters and scaling success across teams. So lace up those virtual hiking boots, limit your WIP, and let's get going.

Today, we're digging into one of the most significant findings from the ninth annual State of Agile Marketing Report, and it has almost nothing to do with LLMs or prompts or agents, and almost everything to do with how teams are structured to use them. Buckle up everybody, it's gonna be a fun one.

So let's start with the headline finding, because, yeah, it is a biggun. In this year's report, Agile marketers are three times more likely than their non-Agile counterparts to say that AI is now fully integrated into their marketing processes. Three times. Three times more likely. Not a little bit more likely, not a modest but manageable gap. Three x. Three times more likely.

And now, before you jump to any conclusions that Agile teams somehow magically have bigger tools or better budgets or more dedicated AI resources — that is not what our data is showing. The gap is not about access or money or technological sophistication. It's really just about readiness.

And now here's what I mean by readiness, because full I— AI integration doesn't happen because you pick the right software at the right time. It happens when your team has clear enough priorities to know where to apply AI, a defined enough workflow to know how it fits, and a tight enough feedback loop to know whether it's working.

And now, if you are yourself an Agile practitioner, you are probably realizing that these are not strictly AI capabilities. They are, in fact, Agile capabilities, and the teams that built them first are the ones getting three times the AI integration — not because they're psychic and they planned it that way, but because a functioning operating system is what makes any new technology actually work.

And the teams that are still struggling with AI integration — a lot of those are also the ones with murky prioritization, reactive workflows, and a backlog situation that can only be described as everybody's inboxes. And no amount of AI spend is gonna fix that.

So let me make this as concrete as possible, because I think prioritization is one of those words that sounds important but actually means nothing if we are not really specific about it. Here's what good AI integration actually requires.

Your team needs to be able to answer three questions at any given moment. One: what are we working on? Two: why is this the most important thing to be working on right now? And three: what does done look like? So we all know when to move on.

If those three questions have clear, shared, documented answers — congratulations, you have the infrastructure for effective AI use. You can identify where AI is likely to compress your cycle time, where it can draft a first pass so humans can do their human thing and focus on judgment, where it can process data faster than any person can. You can test all of those applications deliberately and measure their impact and iterate. Claps. Mad claps for you.

But if those three questions do not have clear answers — or if the answer to "what are we working on" is "it depends on what was in my inbox this morning," or "it depends on who you ask" — then AI is not your friend. It is just going to accelerate your chaos, not fix it. Add more fuel to your already raging dumpster fire.

And this is where the planning data from our State of Agile Marketing report gets really interesting when we layer it in, because eighty-two percent of Agile marketers say they are extremely or very confident that their planning habits keep them focused on the most important work. For non-Agile marketers, that number drops to a depressing fifty-nine percent. And eighty-nine percent of Agile marketers are updating and adjusting their plans at least monthly, compared to just sixty-six percent of non-Agile teams.

So Agile teams are not just more confident about their priorities — although that is awesome, and they are winning just for that — they're also adjusting those priorities more frequently based on new information. Who has lots of new information in their life? Yeah, we do. All of us have new information every freaking day, and so if you're running off a six-month plan and you're not making adjustments, I cannot even tell you how far behind you are right now. I just can't even see you from here.

So this combination — high confidence in planning plus high adaptability to those plans — is exactly what effective AI deployment requires. You have to know what matters so you can point the tool in the right direction, and you need to be able to adjust that aim when what matters shifts. Again: not an AI skill. That is an Agile skill that makes AI work.

Okay, it's about time for me to pump the brakes here for just a second, because there is a pattern that we are seeing with clients that I really gotta call out verbatim here. We have worked with teams who came to us after they tried to pass over their operational issues and jump straight to, "Hey, let's do a bunch with AI." They had the mandate from leadership, they had the budget, they had genuine enthusiasm, and they hit the wall almost immediately.

Not because they bought the wrong tools, and not because they weren't smart or motivated to make AI work for them. It's because their operations couldn't handle it. Their workflows were too tangled up, their priorities were unclear and undocumented. The teams were too reactive to set up any kind of consistent test and learn practice around this new, extremely fast-moving and untested technology.

So they had no choice. They had to stop, back it up, and start working on the operational fundamentals that they thought they could skip over. They had to deal with things like: how does work move through the team? How do decisions get made? And how do priorities get set and socialized? Essentially — spoiler alert — they had to adopt Agile.

But they now had to do it six months later, and with a team that was also exhausted and skeptical from the failed AI push. That is the hard version of this story, and I share it not to scare anybody off of trying things out with AI. Please go use AI if you're not already. It is powerful, and getting more powerful every single day.

But I share this to make you be honest with yourself about the proper order of operations, because a change fatigued team is a hard team to lead. If you ask people to do the hard work of shifting how they operate twice in a short window, that second ask lands very differently than the first. The smarter play is to fix your operating system deliberately before you try to install the new software. Shift to Agile, then install AI.

And this is not because Agile is some kind of prerequisite on, like, a philosophical level, and it's just better, and we should all do it. It's because Agile, practically and mechanically, is the way that the teams who are getting their AI integration right three times more often — they are using Agile, because they have the foundation first. I mean, if you think about it, the cars that are going three times as fast as everybody else are not the ones with the prettiest paint jobs — they're the ones with the best maintained engines. Fix your engine before you try to make the paint job pretty.

Okay. The next thing that we gotta put a real, solid label on is something that I think is getting lost in the whole, like, frothy AI conversation, especially for marketing teams. AI: very good at speed, at scale, at synthesis. It's gonna produce a first draft faster than any human, hands down. It is gonna be able to process data sets that would take you weeks to analyze manually. It can generate opinions and variations and summaries and headlines at a volume that is simply not even within our realm of understanding a year or two ago.

But what it cannot do — at least not yet, and definitely not with any, uh, high level of reliability — is to tell you what is worth doing. It cannot set strategy. It can't look at your organization's specific situation and tell you which of those 17 number one priorities should get your sprint's capacity. It definitely cannot build trust with your stakeholders or earn your marketing team a seat in strategic planning conversations. These things require judgment and really solid relationships, and they require credibility — and that is exactly what Agile marketing is designed to develop.

Here's the way I think it's helpful to think about it. AI is leverage, and leverage amplifies. It amplifies whatever system you put it into. A clear, well-structured, Agile workflow with AI in it gets leverage on the right work done faster with better feedback loops. A reactive, chaotic, "let's just figure this out as we go" kind of workflow with AI inside it leverages the chaos. You're just getting more chaos faster.

And our SOM data — State of Agile Marketing report data — shows Agile marketers are three times more likely to have AI fully integrated right now. That is a very large jump from our twenty twenty-five data. And so I'm gonna argue that we are not at the ceiling here. This is not the biggest discrepancy that we're gonna see between Agile and non-Agile marketing teams and their level of success with AI. This is just the current read on the teams who happen to have already built the right system and put it in place before AI came along.

This gap between Agile and non-Agile teams on AI effectiveness is just gonna widen as AI gets more capable and more broadly applicable. Because the teams who have clear priorities and tight workflows — the Agile teams — are gonna keep finding new ways to deploy AI well. They're gonna keep finding new benefits, and it's going to keep compounding their advantage over time. Meanwhile, the teams who are still trying to figure out if AI is gonna work for them, and get permission, and sort out where it even fits in their workflows, and dealing with all of the operational fallout of trying to put it in when it's not ready — they're gonna keep finding more and more ways to just trip over their own feet and get in their own way. So we're seeing a really big gap right now: three x gap. We are not at the ceiling here. This is just gonna get worse.

Okay. Let's get into the what, who, and how of this, because, as always, I don't wanna leave you with "Agile as the operating system for AI" as an abstract claim without giving you some really concrete evidence here.

So when we think about the what part of the Agile operating system, we have strategy, we have prioritization, we have goals. Agile is gonna give us a backlog: a single, ruthlessly prioritized, visible list of what the team is gonna work on, and in what order. That backlog is what allows us to make a real decision about where — exactly where — AI can add value. Not a vague, "Hey, let's go use AI," as a nebulous directive. But here are our top three priorities this sprint. Which one benefits most from AI-accelerated drafting, or AI-powered analysis, or AI-generated variations for testing? That's a real question with a real answer, and you can only engage with that question well if you have a real backlog in place.

Now, if we think in the h-how part of the operating system — when we're thinking about day-to-day practices and iterations and the test and learn cycle — Agile gives us the experimentation infrastructure that we need to figure out how AI is really gonna work for us. And this is the place where Agile marketer— Agile marketers are already using to pull ahead on AI, because they already run experiments. They already have an established cadence of trying something, measuring what happened, and then adjusting based on data. So integrating AI into that loop is not a brand-new skill or a new habit that they have to build — it's just a variation on an existing practice.

But for a non-Agile team that is addicted to perf-perfection, wants to have all the information up front, and they have these very anti-Agile habits — integrating AI into this faster, iterative feedback loop requires just even starting the habit from scratch, while also applying a brand-new tool simultaneously. So it's just a much harder ask.

And when we get to the who part of the operating system — where we have things like team structure and stakeholder relationships, psychological safety and trust — Agile gives marketers the credibility that we need to make AI adoption sustainable. Because when your team is predictable, when stakeholders can see what you're working on and trust that you will deliver it when you say you will, the conversation about, "Hey, we'd like to spend some of our capacity experimenting with AI," is a very different conversation than it is on a team where nobody quite knows what marketing is up to at any given moment, and they don't really trust that marketing is gonna be able to deliver on their commitments.

At the end of the day, AI is a what. It is a new thing that informs what we do, what our activities are. But we need Agile as the foundational how, and the thing that informs who we work with, and who we collaborate with, and who we deliver value for. Those things together are what make AI actually work.

All right. Time for the super practical closing segment. Here's what you need to really do with this.

First and foremost: before you go and pay for another LLM subscription or buy another piece of AI-powered martech that you have to try to shoehorn into your existing workflow — please stop. Map your current prioritization system first. And by system, I of course mean something that is written down, documented, agreed upon within and outside of the marketing function, and then is actively maintained, so that everybody always knows what you're currently working on and why. Prioritization is an ongoing activity. It is not a one-time thing that we do and then walk away from.

So if you don't know what your prioritization system is, or if your answer includes any kinds of, "Uh, no, not really," or, "Well, yeah, but sort of," or, "Yeah, we have priorities, but they're in everybody's heads or notebooks or inboxes," that is your first project: documented and adhered to prioritization system, not AI. We're not doing AI until we have priorities. That is first.

Second: once you have that prioritization system, and it's turned into your functional backlog, then it is time to identify the one bottleneck that is costing you the most. We have a whole episode about this, so if you haven't listened to it, go back and do that after this one. But the bottleneck is the place where work slows down, gets stuck, or generates the most waste and rework. That is the place where your first AI experiment goes, in a very pinpointed, scalpel-like motion — not a big sledgehammer.

We do not need a broad, vague, "Hey, let's all try AI" initiative right now. We need a specific, targeted test. Can AI widen this bottleneck? That's our question. Run the test for thirty days and measure your before and after outcomes. This is where we go second, after we have our prioritized backlog.

Third: we need to build the continuing support of that experiment into your existing ways of working. Of course, if you're already running sprints or iterations, this is easy. The AI experiment goes into the sprint just like any other work item. If you are not using sprints or iterations, this is your great opportunity to start. Even something like a loose two-week work cycle with a retrospective at the end is gonna get you a lot closer to having the feedback loop that you do absolutely need in place to know if these AI experiments are working. Remember that study that ninety-five percent of AI initiatives are not delivering ROI? These kinds of fast feedback loops are how we make sure that you are not in that ninety-five percent bucket, and instead are getting real business value from the time and money and energy that you invest in experimenting with AI.

And fourth — now, this is the one that is very tempting to skip, because you just wanna start. It's so cool, and AI is so sexy, and we're gonna get so many amazing benefits from it. We just wanna start. But you have to make sure you're measuring the right thing around your AI experiments. Not, "Hey, did we use AI? Great. We win." No. That's an output measure. We wanna measure outcomes — things like shorter cycle times, the percentage of unplanned work that are coming into our sprints, how many rounds of revision, the speed with which we obtain stakeholder approval — something that tells you whether your application of AI is actually improving the flow of work through your system.

If it is improving it, great. Do more of it, or do more like it. If it's not, great. We learned, and we learned quickly, in one or two sprint cycles, and we adj-adjust the experiment. That's it.

That is the whole playbook. AI is complicated enough; we do not need to add more complication to the system. Super straightforward. It's not a transformation. It's not a highfalutin AI task force. It's a backlog, a bottleneck, an experiment, and a measure of success. That's it.

Go do that right now, and you will be able to quickly join those ranks of Agile marketers who are three times more likely to have fully implemented AI into their processes. Until next time, I am Andrea Fryrear, and remember: the struggle is real, but so is Agile marketing.