What AI Gets Wrong About Strategic Planning

Episode Transcript

Rest in peace planning. Haven't you heard the news? AI, uh, it's killed strategic planning.

It' is official. A headline says your plan is already obsolete. The logic, it's hard to argue with. AI, it can build the competitive analysis before lunch.

So why spend six weeks on a plan that's gonna be stale before February? Just make it up on the spot. Let anarchy rule the world of work.

Ugh. Okay, let's slow it down. It is Q4, right? It is planning. season, but somewhere in your building, I bet there's a leader or two asking, "Does any of this matter?

Do we need a plan?" Well, you need an answer, right? And, and here's mine, in a bit of a metaphor. You know, nobody looks at' the weather or the forecast around it and decides, "Ugh, it's unreliable, so forget it. I'm not even gonna go outside because

I am not certain whether I need sunglasses or an umbrella, if I can put on a pair of shorts, or I need to put on a coat." Think about now people whose jobs are directly influenced by the weather. Let's say a groundskeeper, you know, they don't skip preparing the field and painting those lines because, uh, well, that looks like a storm cell that might roll in. The weather, it's, it's gonna do what it's gonna do.

The lines are needed. They're what let you play the game. If conditions call for an audible, well, that's exactly what the crew is trained for, and they're equipped to handle it.

So AI hasn't killed planning. It's possibly killing the false precision that's been around it, right? The polished, locked in 12-month plan.

That version that we pretend works. However, the dynamic one with the right? level of fidelity to it, the just in time type, well, that type of planning is alive and, well.

And us Agilists, uh, we've been practicing this type of planning since way before AI was cool. So today we'll be covering several things.

Why a protected plan is a brittle one, why AI without direction may get you more output, but not necessarily outcomes, and how to decide things at the last responsible moment. Plus, we'll give you a 90-day experiment you can start now. Welcome to the Agile Marketing Edge, the first podcast dedicated to turning agile theory into real world marketing breakthroughs. I'm Ross Libby, Chief Value Officer at 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. This week it's planning season, and today we're working in the what and how of the Agile marketing operating system, what work you choose and how often you revisit that choice.

I also spent last month at the SAFe and AI Summit in San Diego put on by the team behind the scaled Agile framework, and the same questions kept coming up there. How we plan and what AI changes about it. So you'll hear a few of those moments as we go, alongside what the data says about how Agile marketing teams plan, what the AI debate gets wrong, and how to build a plan that bends instead of breaks.

So lace up and let's go. I wanna start by giving the plans are dead crowd their best argument first because it's a pretty good one. Uh, about a year ago, Soren Kaplan, uh, in his Inc. article talked about this. Most planning frameworks, well, they were built for a slower time. Uh, you gather the information, you set the direction, and you assume it holds.

Well, as he says, you know, competition nowadays, they can launch AI powered products in days, and customer needs, well, they can shift overnight. And the annual plan, well, it's stale before anyone has a chance to present it. So his fix is to think in 30 to 90-day cycles.

Review monthly or. even weekly and let AI watch the data, while humans decide what matters. He's really right on with his diagnosis.

Um, and there's a word for it, or maybe better put, acronyms. Uh, so for years leaders described the world, uh, it was VUCA, volatile, uncertain, complex, ambiguous.

It came out of the military in the late '80s. Well, fast-forward to today and a futurist named Yamez Casio argued it is no longer fit for what we're experiencing.

BANI is a better acronym, brittle, anxious, nonlinear, incomprehensible. Two of those matter most for our conversation today.

Brittle means it looks strong until it's not, right? It doesn't bend, it shatters. Nonlinear means small inputs can make huge differences. One model release or one budget cut and your whole quarter, well, it changes shape entirely. And here's the part, uh, no one really likes hearing or talking about.

Well, these annual plans, they're brittle because we make them that way. We polish them, protect them, build them at an offsite and hand them out in a deck and guard them like they might crack because they probably will.

You, you've seen a coach draw up a play, right, that looks unstoppable on a whiteboard. Well, nobody's guarding it. Uh, that's the protected plan, right? It looks great for the same reason. It's never been tested on the court, and once it gets outside of that protective bubble, well, it pops and hardly stands a chance. At the summit, uh, I went to last month, Mick Kirsten showed two photos of the same stretch of Fifth Avenue in New York over the course of several years. Uh, the first one in 1900, couple of cars sitting in a sea of horses.

13 years later, one horse in a sea of cars. As I took the point, uh, people who got left behind were the ones who kept the old plan and operating model and attempting to bolt on something new to it. Sound familiar? Well, I think that's what's happening right now, with a lot of AI rollouts.

Um, even the folks you'd expect to defend the big plan are changing it. Scale to Agile's new AI native guidance trims that two-day big planning event down to one, and adds a short sense and response session every iteration. Smaller plan with more loops.

Uh, but don't overcorrect, right? We, we have to be mindful. 30-day horizons and weekly replans have a cost. Several at the summit, they put it bluntly.

Changing priorities doesn't create capacity. Swinging from rigid to constant, and you trade a brittle plan for a whiplash team.

So the question isn't how often you plan, it's whether your plan has a feedback loop. When's the last time something outside the room changed it? If the honest answer is that next planning cycle, you're brittle. Find where the loop is.

If you can't, put one on the calendar this month. Here's where the plans are dead argument, well, it quietly falls apart. Start with what AI is actually doing for organizations.

McKinsey's 2026 State of AI survey took about 1,700 respondents, and here's what they found. 80% say  AI has improved personal productivity. Well, great, but only 37 say it contributed to their organization's earnings, and that number didn't budge from last year. McKinsey itself calls the gap between individual gains and enterprise impact the striking finding. But it's always good to get a second opinion.

So BCG surveyed nearly 12,000 people this year. Among frontline employees who use AI regularly, 42% say they are saving about a workday a week. A day a week. But yet two-thirds of these frontline employees report no or little guidance on what to do with the AI time they've saved. BCG's own conclusion, that time leaks out of the organization unless somebody deliberately reinvests it. So output, it goes up.

Outcomes, not so much. Same pattern in two different surveys. BCG's answer is strategic clarity. Employees who own where the organization is heading outperform employees who simply have a better AI tool. Tools without direction, well, it loses to direction without tools. And what's direction? It's a plan.

Not a 60 slide one. One people can actually use. We've seen this movie, right? The Agile Manifesto says responding to change over following a plan. Over, not instead of. Plenty of people heard instead of. You've probably worked with a few.

Declaring AI the end of planning is the same misread, running at AI speed. Kate O'Neill, the tech humanist, was on the main stage at the summit and made a related point. "The real challenge," she argues, "it isn't speed, it's judgment." and she asked whether the future is uncertain or just undecided. Those aren't the same. Uncertain happens to you. Undecided means someone gets to choose. Planning is how you choose. Then it got practical. Later in the summit, Kate recorded a podcast live on the stage with Lindsay Wirthling, who's... leads transformation at the American Red Cross.

That's an organization that's backed by hundreds of thousands of volunteers, so the question of what the technology and service to, uh, it's never abstract. And the conversation kept circling a prime idea. When all anyone wants is an AI, don't leave the problem or the plan behind. Lindsay's team took a design thinking workshop they'd run for years and made it the front door for every AI request.

Problem first, with the data scientists in the room early. Tool second. Kate's phrase for the whole idea, "Build the scaffolding to earn the height." Well, the plan, that's the scaffolding.

So here's your move this week. Before anyone prompts anything, write down the problem you're solving and where the time you save is going. Let's look at what planning looks like for the team who do it differently already.

In our ninth annual State of Agile Marketing report, we asked how often a team updates its plan.

Well, among agile marketers, 89% said at least monthly or more frequently.

Among everyone else, that number drops to 66% who did so as frequently as a month or less. Hold that up against the plans are dead crowd. The teams that plan the agile way aren't planning less, they're planning more often, and they say it's working. Asked whether their planning habits keep them focused on the most important work- 82% of agile marketers said they were extremely or very confident. For the non-agile crowd, only 59%.

When priorities change, 82% of the agile crowd were confident they'd stay organized versus 58% with those that don't work in an agile way. The usual fear is that that constant plan updates mean chaos.

People doing it report the opposite, and look at what they're updating their plans from. 87% of agile marketers say they often use customer insights or data to guide their work versus those on the other side, only 72%.

They're not changing the plan on a whim. They're changing it on evidence. That's the feedback loop from earlier on a schedule.

But here's what keeps frequent updating from turning into whiplash. Every level of the plan gets its own level of detail and its own rhythm.

Think about a cross-country road trip. Nobody plans it at the street level. You start zoomed out, point A to point B, where you're starting and where you're headed. You get more granular as you get closer, and when construction or a closure forces a change, you reroute. The destination doesn't change, though. Your plan, well, it works in a very similar way. The year has a point A to a point B.

Some bets, what you're trying to change. Low detail revisited every quarter. The quarter is the route. The commitments you're making revisited monthly. The next couple of weeks, well, that's where you get more of the turn-by-turn.

Who's doing what? By when? Revisited weekly. Each level gets just enough detail for its horizon and a rhythm to match. That's what makes the updates useful instead of exhausting.

Uh, the PMBOK calls this rolling wave planning. Near-term work in detail, further out work at a higher level.

Product teams, they call it now, next, later. In our operating system, it's the what and the how working together. And if your plan has task-level detail nine months out, that's not the right level of rigor. It's creating friction.

At the summit, I had the opportunity to present with a client. My co-presenter, Petra Kalik, and

I built a lightweight version of our marketing operating system diagnostic right into the session. Attendees rated their organization across the who, what, and how layers in real time. Then we asked each person to commit to an action they'd take forward that following Monday.

Uh, looking at what came back, almost none of it was a new tool or a bigger plan. It was cadence and evidence, bringing last year's cycle times to the next planning conversation.

Run a customer interview in every sprint. Set aside one-on-ones with leaders to get aligned, right? That alignment on who owns what.

Small, specific, calendar-ready. Uh, I did my own, and it was related to these feedback loops, getting more of a defined strategic cadence on the calendar more regularly. So here's the move. Find the parts of your plan sitting at street level that really should be zoomed out.

Then put a revisit date next to every level, a daily stand-up for the next two weeks, monthly check-ins for the quarterly plan, and quarterly planning to revisit the year outlook.

No level of the plan should be the Ronco Showtime Rotisserie, right? You remember the catchphrase from the infomercial?

I sure do. "Set it and forget it." Forget it. You need to revisit the plan.

All right. I'm getting a little cheeky here. Uh, and fair warning, I'm gonna take it a little further because I think here's where I'm allowed to be a little smug.

Um, hold on and, and I'll show you why. So first, we're taking it back. In 2003, that's 23 years plus before anyone said AI era, AI planning, Mary and Tom Poppendieck wrote a book called Lean Software Development. One of the framings from it is the idea of the last responsible moment. The idea, hold off on a commitment until that moment when not deciding would take an important option off the table.

Decide as late as possible and no later. And they're blunt about what happens when you blow past it. The decision gets made for you by default.

Uh, you know, look at it this way. Great players use this. They let the play develop before they commit to the action, the cut, the pass.

Uh, but they're very aware of the surroundings. They have an eye on the clock, right, the whole time. It's the part people can miss. The, the idea of the last responsible moment isn't procrastination.

It's a deadline. Say your deposit on your conference booth is due in March. Well, that's your deadline. So decide by then.

Uh, what the booth says, that can wait until, you know, closer to launching. Most teams don't know which of these decisions have deadlines, so they just make them all in January or none until it's too late. And AI, well, it sharpens this in both directions.

It shortens how long it takes to learn something, so, you know, in essence, we can wait for better evidence closer in.

But it also makes deciding by default almost frictionless, right? 20 campaign concepts by lunch, and it's easy for the first one the tool spits out. Well, that's the plan. The constraints, well, they're here to help.

James Taylor, who emceed the summit, made the case that constraints, well, they actually make us better by limiting choice. A deadline is a constraint. It's a forced ranking. Here's how it played us, for us here at AgileSherpas. Uh, a few weeks back, we planned Q4, uh, and we made three calls. We stopped an offering experiment that never cleared the marketing sales hurdles.

We deprioritized reworking another offering for a virtual format because other priorities and capacity came into play.

And in turn, we prioritized exploratory work on some universal marketing operations and workflow problems we keep seeing that, along with some certified course material, is worth a deeper look. Notice what the plan did here.

It said no, it said not now, and it said let's explore. A plan that can only say yes, well, that's not a plan at all. It's just a, a list of wants and wishes.

So yeah, pardon me if I'm a bit smug that us Agilists, well, we've been planning in this Bonnie way way before AI was a thing.

Uh, if any of this is new to you though, here are a few things to ask for every big decision in your plan. When's the last responsible moment? What do we need to know by then?

And who decides? Just start that way, that simple. All right. I know that here's the objection you'll hear in the planning meeting. "You know, that's great, but leaders wanna know what's shipping and when. You know, I need the information."

Fair. But what they're really asking for is likely predictability, and there's more than one way to give it.

You can promise a list of deliverables stamped with dates from here until end of year. Well, it's gonna be the brittle version. Or you can promise a destination and the rhythm used to get there. That's what OKRs are built for.

The objective is the destination. The key results are how you'll know you're getting there. John Doerr picked up the approach from Andy Grove at Intel and brought it over to Google in 1999.

Here's what it looks like in practice. A marketing organization runs a planning cycle every quarter. They start with a look back. What did they learn? They reconfirm their goals.

Every request into one place, and it gets checked against those goals. They sit down with each other. They sit down with shared services team.

They talk priorities, capacity, risk, map out the quarter, look at the holidays, the PTO. Every risk gets sorted into is it resolved, owned, accepted, or mitigated.

And one question lands on the agenda every time. If work is being deprioritized, who's telling the people that care, the stakeholders, partners?

You know, they usually, these stakeholder partners, they can live with something about not right now, maybe not this quarter. They can't live with finding it out too late or finding it out without a reason. So you have to say, "Not this quarter.

Here's why, and here's what's taking its place." That's dynamic planning. Every no is a real trade, and the trade is visible, and it's the same move behind our own Q4 calls, what we stopped, why, and what took its place.

At the summit, uh, Mick here again urged that every function needs to be on this cadence. Kate O'Neill, uh, pointed to secret private loops running outside of it. My take, when the official cadence can't keep up, the work goes underground. And when Petra and I asked our session attendees which conversations they had been avoiding, the most common answer was the who one, leadership behaviors.

One person's action, get leaders into the prioritization sessions. There's an idea. So sell your stakeholder the cadence, not a calendar.

Tell them that destination, when you'll revisit the route, and who tells them when something changes. So what do you do with all this?

Um, I'm gonna give you four moves in order, and the order, as usual, it matters because if you set the rhythm before you know the destination, well, you just might meet more often about nothing.

That may be fine for a sitcom. Seinfeld ran nine seasons on it. But nobody wants a Kramer from the next cube burst in unannounced through their planning door.

We wanna focus our plan, not just on anything or something. Focus it on the right thing. We do that by, one, name the destination. Write down three to five outcomes you're after this year, the problem each one solves, and how you know you'll get there. No deliverables, no dates. While you're at it, write down where the savings is gonna go when the team gets it and those individuals on it.

Two, zoom each level appropriately. Year gets direction. Quarter gets commitments. The next couple of weeks, well, that gets the detail. Put the revisit dates on the calendar now, quarterly, monthly, weekly, and cut any street-level detail that sits further out than the quarter.

Three, tag your big decisions. For each one, when there's the last responsible moment for it, and what do you need to know by then, and who decides? At every monthly review, bring in one fresh signal, like a customer insight or a piece of data.

Let AI, AI do the scanning. Keep humans on the deciding end. And four, tell your stakeholders the rhythm, when you'll revisit the plan, when they'll hear about the changes. And for every not this quarter why, and what's taking its place.

And then, this is important, run it for 90 days. This isn't asking for a reorg, not a transformation, an experiment.

Track three things, how often the plan changed, how often a stakeholder was surprised, and whether the outcomes from step one moved. Changes are fine.

Surprises, they tend to be the problem. Planning, it's a practice, not an event, and practice takes reps. If you want a place to get them, it's what we built The Ropes for. So let us show you 'em. It's our free community, more than 100 hours of agile marketing leading and agile marketing learning, including self-placed versions of our Agile Marketing Fundamentals and Leadership course, tons of carabiners. Those are the microlearnings. You can join for free at theropes.agilesherpas.com.

And remember the weather. It's gonna do what it's gonna do. Paint the lines anyway. Train the crew, and be ready for the audible.

Until next time, I'm Ross Libby. And remember, the struggle is real, but so is agile marketing.