Increasing Decision Capacity Post AI

In this video, Stacy Gordon sits down with John Greisner, a management consultant at LiminalArc, to explore why AI doesn’t remove your software delivery bottleneck so much as move it. Drawing on Goldratt’s theory of constraints from The Goal, John walks through what happens once engineering gets fast and cheap: the real slow point shifts to product management, funding cadence, and how much change your customers can actually absorb.

Stacy and John dig into why flooding your backlog with every new idea only makes that constraint worse, and why deciding better, smaller, and more often works instead. Watch so you can stop chasing engineering speed as the finish line, start treating decision capacity as your real constraint, and finally get the return on AI adoption that an operating model built for a slower era has been quietly blocking.

Video Transcript

John Greisner

AI is definitely an enabler to help build things more efficiently and faster and to automate the things that were often manual. And so the engineering work has gotten fast enough and more efficient in reducing costs that it’s kind of prompting a whole new set of questions. Where is the constraint in the system? Where is the weakest link? We are now looking at other areas that are slower than the build cycle that we’ve been trying to follow from engineering to the front end. Once you sell the front end, really then the question is how much could our customers even consume and be able to take to make it effective that they’re going to love our products?

Stacy Gordon

I’m Stacy Gordon, and every week I pull up a chair with someone willing to talk about all the unglamorous stuff that is really behind why the thing actually works or doesn’t. My guest today is John Greisner, a management consultant with Luminal Arc. John published an article titled The Bottleneck Moved: Your Operating Model Hasn’t. And it’s one of those articles that honestly I had to read many times. But at the end, John, your takeaways were simple, but extremely powerful. I’m very excited that you’re on the podcast with me today. Welcome.

John Greisner

Thank you, Stacy, for having me. I appreciate the opportunity to share a little bit more. It was kind of a fun write for me for several reasons, but I do love the story from the book The Goal and it’s one of those pieces of work that’s been around for many years. Companies like Amazon and Tesla through Elon Musk and Jeff Bezos are always kind of promoting the reading because it really teaches a lot of fundamentals about building things. And in our world, mostly we’re talking about building software, but it does teach us the very important lessons about where is the constraint in the system, where is the weakest link? And so the story of the hiking is one of those examples that’s in the book. And I can share my own hiking story. I took a whole group of nephews and nieces and friends out a couple summers ago and I had to live through the story myself.

And it was kind of funny because in the back of my mind, I am thinking about the book the goal and the hiking. And so we started off and the younger guys, they’re just sprinting up the hill and we’re all trying to keep pace. And there was a few of us, me included, that probably hadn’t hit the gym enough and we were kind of huffing. And so we ended up getting spread out. And so we all wanted to stay together and be able to get there at the same time. So the plan was is that I would go first. I would set the pace and I was setting the pace based on generally who was the slowest and that’s how we all kind of stayed together. So that’s my little story and I’m sure you all have similar stories that you can relate to.

Stacy Gordon

Oh my gosh. Well, I love it when something in your real life actually ties back to things that you experience in your professional career. It always makes it more real for me. You’re actually making me want to go read the book. So I’ll have to put that on my list. But one of the things, you have a line in your article that says Goldrat’s rule has two halves and the second half is that a constraint you elevate does not disappear. It moves and it rarely announces its new address. I really loved how you phrased that. So basically your argument is that most companies finally fixed their old slow spot. They found Herbie, they fixed building software, but they didn’t realize that a new slow spot was going to take its place. When did you start seeing this?

John Greisner

Yeah. So I mean obviously in delivering on a lot of the AI work that we’ve been doing over the last couple years, AI is definitely an enabler to help build things more efficiently and faster and to automate the things that were often manual. And so what we’ve started seeing in a few of our clients that we’ve been working with through this phase is that the engineering work has gotten fast enough and more efficient in reducing costs that it’s kind of prompting a whole new set of questions. And so it’s really a statement about yeah, it’s a tremendous accomplishment for a lot of the companies that have spent decades to focus in on adding agile practices, DevOps, cloud migrations and all the different capabilities to make that engine DevOps as well to really make it work more efficient and faster and cheaper. Even breaking down work, like I said, between 90 days of strategy, all of that helps get worked through.

But what it does is when it’s done is to the point of the book, the constraint moves. And the question that I’m posing is where does it move and what’s next?

Stacy Gordon

And so do you find, this is relatively a new concept, right? AI is changing on a regular basis. You hear customers constantly trying to figure out how they use it to develop faster. So do people even realize that they’re solving this problem, the engine of building software was always slow. Have they even really grasped the fact that they are kind of moving their bottleneck somewhere else? Do they recognize that that’s what it is that’s happening to them?That’s

John Greisner

A funny question because when I started reflecting on this this spring after seeing it, it actually took me by surprise because like I said, we’ve been working this problem for decades in this very specific area of our overall strategy to deliver customer value. And so it kind of became that epiphany moment to be like, oh my God, we are now looking at other areas that are slower than the build cycle that we’ve been trying to follow. So this is another reason why I wanted to write the articles because I love being aware of what’s coming and it even took me by surprise because the short time it took to really implement the full benefits is why.

Stacy Gordon

Well, I bet it’s a fascinating conversation to have with executives at companies because again, I’m sure they’re having that same aha moment that it’s taking them by surprise now that they’re sitting here kind of faced with different challenges. So you talked a lot about a company’s operating model, the wiring that turns strategy into things customers actually get was built for a problem that no longer exists. In plain terms, what do you mean by that and why is that kind of what I believe is probably the real headline here?

John Greisner

Yeah. So I mean the operating model, all the components kind of stay the same. So what we’re essentially doing when we’re implementing AI is we’re kind of adding almost like a race car engine to the process. And the way that I kind of think about that a little bit and the effects of that is that if you have your operating model and it’s not quite kind of wired to some of the points that I said earlier, like you don’t really have your DevOps kind of squared away, your practices and synchronizations are a little bit off and not really flowing, you add that bigger engine and it’s like putting it in an old car with a bad front end and you’re going to go off the road a whole lot faster. And so that is some of the things we’re hearing too. They’re like things are getting crazy and we’re discovering a lot more things.

So it amplifies all the operating model components. And so there are different pieces of it that start to need to be tailored a little bit more for that faster engine.

Stacy Gordon

Okay. I loved that description of the old engine with the bad front end and running off the road because I’m sure that that’s what people are feeling like right now. They’re like going, “We’ve solved a problem that we’ve been trying to solve. We’re now going super fast. We’re going to win the race. We’re going to win the competitive race,” but yet they’re not starting to see or they’re not feeling like they’re accomplishing what they though they would by being able to develop and do things faster. If you look at all the wiring that’s in someone’s organization, so all the different pieces of the operating model, help me understand how do decisions get made? How often do you plan? Who owns what? Which roles matter? Which one of those kind of gets out of date the fastest, right? So we’re turning our lens to look at the different parts of the business.

So which one of those do companies need to start looking at?

John Greisner

It’s really who’s been impacted the most in this change is the product manager. One that is getting the feedback from the engineers and the one that’s responsible for providing the highest value of work. They’re being asked more frequently. So the change in which if you think about a team that maybe might do two week sprints, the product manager is trying to build up that backlog. We tend to like to have a little healthy backlog of like two to three sprints out so there’s clarity on the team and what they’re going to be doing. That gets consumed really fast if you have a high speed engine in there. And so that puts a lot of burden on the product manager to keep that backlog healthy as well as validating did they build the right thing? Is it fit for purpose? And being able to. So it’s really putting a lot of strain and stress on that one person and that’s kind of the biggest impact.

How do they then navigate this new world that they’re living in?

Stacy Gordon

Yeah. I think that’s interesting because I think one of the concepts that you had in my plain language is that if ideas are cheap or free, why not build them all? Why does throwing more ideas at the problem actually make things worse? And so this is really, I think, what you’re describing for that product owner of going, he’s got people coming in from every direction giving him all these new ideas and he’s got this really fast engine of being able to build them, but that doesn’t necessarily mean that you should just because you could.

John Greisner

That’s right. And I think that’s the hazard or the impulse is just fill that backlog as fast as you can for them. And that puts a lot of stress on the back end of it, which is the customer receiving things they really don’t need and the change impact to that. So it really takes that discipline, which we have been talking about for many years too, to understand what’s the highest priority and being really spot on. And so it’s that decision that is first and foremost on the spotlight now. It’s like make clear decisions on what the customer needs and being able to prioritize that.

Stacy Gordon

So you believe that a company can actually get better at that. That is something that if they have the right guidance or the right framework or the right research, what is it that a company can do to kind of make more informed decisions about which ideas are worth investing in and which ideas aren’t?

John Greisner

Well, and AI, just like with engineering, is a great companion to help those discussions, but it still, at the end of the day, takes a human to be able to make that judgment call on what’s the most important business problem to solve, what capabilities need improving and to be able to do that. But they do have a helper too to kind of go through and vet ideas, but the cost of delay scenario that I mentioned in the article is a little bit different than the way it’s traditioned because we’ve often looked at how much effort someone takes over value. And now with effort being low, we’re really focused on highest value and what that means to our customers. So the disciplines are there. I mean, different companies excel in that space a little bit more than others, I would say, to the point of if your capabilities and skills aren’t there, this is going to amplify and make it clearer.

So if companies are still struggling with product management and how it should work and connect to strategy, you’re going to feel this pain much more and more faster essentially.

Stacy Gordon

Well, one of the things I thought was so interesting or the claims from your article is that it talked about the real slow point actually may be outside your company entirely. Customers can only, and I love this word, absorb so much change. So how did you actually land on that?

John Greisner

Well, so it’s really the full circle. So in just thinking about if we trace this constraint, if you will, from engineering to the front end, once you sell the front end, really then the question is how much could our customers even consume? And we’ve all been victims to too much change and the change fatigue we feel in some of the products that we use and how rapidly. So it is a real problem and a challenge that companies are now going to have to really wait and take a look at. We didn’t have this problem so much because we often couldn’t keep up with what the customer needed.

Stacy Gordon

So

John Greisner

That wasn’t the problem space, but now we do have to really weigh how much can the customer consume and be able to take to make it effective that they’re going to love our products. So we are now dealing with really two different areas that are highlighted now because of the faster engine.

Stacy Gordon

Yeah. I do wonder, again, I think the pattern within your article talks about, again, having to look at different areas of the organization under different lenses. And so I can only imagine the heightened effort that an organization will have to put on customer research of being able to go, okay, we’re going to really have to measure the changes that we’re making to start getting that real powerful feedback fast so that we are kind of meeting the pace that they’re interested in. Not going too fast, but also not going too slow.

John Greisner

Yeah. I love what you’re saying about measurements because we focus so much on engineering measurements to make sure that engine is running smoothly and is efficient. And we haven’t so much on the front end and back end because that wasn’t our constraint, that wasn’t our problem. But now that those are becoming first class rights to focus on, we have to think about what are the metrics that we want to kind of measure to make sure that we’re not causing or slowing the whole system down. So the question on ideation becomes really a powerful area that we’re going to have to look at and see how long does it take to iterate through our ideas and connect the strategy? Because if it takes us the same amount of time, which was very forgiving in the because we were waiting always on the engineering side to finish the work, God bless them for all the hard work that they have to go through.

So we had slack in the system to be able to deal with this problem. Now the slack is being taken away, so we have to really manage and measure that side as well as like you said, the customer side too, how much could they consume?

Stacy Gordon

Well, and I think it’s probably something that they didn’t even realize that it’s like, I’m sure nobody really had a foundational measurement for how long it took to go from ideas to getting them vetted and put them in the backlog, right? They were just constantly complaining that their ideas never made it out fast enough. So I think the change that organizations are going through is in such real time evaluation of all of their different parts of their organization that they probably, when we talk about that bottleneck or that constraint moves, is it always the product person that they go to first or is it even if the product person gets something, is there more of a revenue officer that might have it sitting on their desk? I mean, I’m assuming an organization may have to look across the board to see who’s slowing the process down.

John Greisner

Well, it’s all interconnected on the front end and even starting to signal, and I mentioned it briefly in this article and there’s another article that’s going to come out in talking about the portfolio side of it with funding. And so when we talk about how often do we fund ideas in this new world of going from ideation to delivery much faster, that feedback loop into how we fund and how we allocate across all the capabilities is also going to connect to that. So the front end is all interconnected. And so what the product manager will need is ability to flex much more faster than the system was willing to do before. Well,

Stacy Gordon

On the finance end, I do think that’s interesting because I think traditionally you get a bucket of money, you build out a roadmap, you align budget items to each one of those things. So maybe that’s where you’re getting that flexibility on the product owner to go, from finance and them managing their P&L, you can’t really look at it the way that you used to do on your roadmap. Okay, I’ve got $100. I’m going to $25 in first quarter for this feature and $25 in second quarter for this feature. I mean, you have to be able to change and move faster about even your overall thinking about budgeting.

John Greisner

Yeah. And like I said, some companies have narrowed the focus to look at budgeting more frequently. Some do it on quarterly. Some are still locked into the annual. And so for those that are still locked in, then there’s going to be definitely a system constraint or a barrier to be able to move faster because it’s tied to something that isn’t, I’ll call it cadenced or synchronized. And that’s something also that I mentioned about one of the changes is that you have to reevaluate your cadences based on your placement of that and the frequency you need to make that decision. And funding is one of those that it does get impacted if you’re able to produce more quickly.

Stacy Gordon

So ultimately, if you kind of step back and you start to look at the entire organization, I mean, ultimately this could mean full organizational transformation unless you just want to tackle one small little thing and then figure out where the bottleneck moves. It potentially says, “Hey, for the best of the best companies out there who recognize they’ve solved this 20 year problem of software development that they’ve had, they really should probably look at the organization over as a whole and start really using those different lenses to kind of identify perhaps for them specifically what their problems are.” But organizational management change and transformation has always been slow and painful and can cost a lot of money. So in your mind, how long is this really going to take for a company to be successful and where do they maybe start or where do they maybe get stuck?

John Greisner

Well, so I think looking at the fundamental question is where is their slowest part is where they should start. And so different companies are in different places in adopting AI, but the cautionary tale is a little bit like if your system isn’t set up to handle a faster engine, you probably have some work to do still in the area of the build to be able to apply it. And so like I said, it’s going to amplify a lot of the outer boundaries and even the things within delivery, it’s going to amplify. So back to the car metaphor, it’s going to steer you off the road very quickly. So I hate to say it depends like the consulting answer, but it depends on where you’re at, where your problems are. But as people are kind of trying to gravel with adopting AI, they’re running their pilots, they’re trying this out, and there are definitely some challenges they’re starting to see.

And I think the challenge is really putting the spotlight on their operating model. And it’s a good thing to try to advance and companies are looking to use it, it is going to highlight deficiencies or other areas that need to be fixed. So transformation is a loose word, but just implementing AI like, “Hey, if we just do these couple things and we set up the AI environment, we’re good.” It definitely impacts a lot more than just the capability itself. So to your point, it’s a wider change that you need to think about and strategize around.

Stacy Gordon

Well, that’s what I thought was interesting about your article too, your example leader, Alex. He doesn’t fix everything. He just starts because sometimes analysis by paralysis. And so if someone were to take a first step, what would you recommend they do?

John Greisner

Yeah. So I think a great question first and kind of lead into when we think about transformations, limital arc, we do like to think about it as a slice, one particular area. And so you can continue to develop the capabilities and understanding and competencies around AI, running some pilots, looking at it, but really look at one slice for your organization and see how you would implement it and the changes you would make and then be able to kind of scale that out. So take more of a pointed view on, yes, fixing the one problem, but in more of an isolated area until you understand the impact and change to the overall area.

Stacy Gordon

I loved the takeaways. I mentioned this up at the top of the interview, that I felt like I walked away with three really simple items from your article that made a lot of sense to me. And so I want to make sure people hear these and walk away with them as well. So it basically says your bottleneck moves. When delivery becomes fast and cheap and is no longer the bottleneck, the bottleneck will find a new address and you need to figure where that is before optimizing anything else. Number two, upstream. The scarcity is decision capacity, not ideas. Feeding the funnel faster makes the bottleneck worse. Deciding better, smaller, and more often makes it work. And finally downstream, the customer sets the pace. What they can absorb, not the capacity, is the final constraint. So build the cadence that senses it. Anything else that we’re missing from those three things?

I think they’re simple but powerful in ways that leaders in their companies really need to look at AI, how they’re using it and how they’re going to be competitive in their landscape.

John Greisner

Yeah. I mean, the important thing is start where you’re at, but understand where it’s going too. That’s the important thing. As you do start to resolve and implement some of these benefits, you will need to start looking more at the pace of the customer. You will need to address the front end frequencies, the funding models. And so yeah, but start exactly where you’re at, just keep moving. And if you have more work to do in the build area, that’s the place to spend it. You can definitely start to take advantage and have AI and some of the agents and skills help you through that process. So there’s some great ways to leverage to get you to a full implementation. So

Stacy Gordon

John, one of the things that I always like to ask is if we had a product owner, for example, listening to our conversation and you were going to recommend him one thing to do on Monday when he went into work that he could impact or change based on what we’ve been talking about, what would it be?

John Greisner

Yeah, I think the first fundamental is what we’ve talked about with cadence and placement. And so the one company that we’ve been working with for a while, one of the things that the product manager quickly did was as the team changed from more of a scrum base to a flow base because time boxing in that sense didn’t make as much sense with the synchronization, it was about placement of those decisions. So understanding what are the decisions you’re making upstream and downstream and then repositioning the cadences around that. And so that will help smooth out what they need to do and keep things going to help kind of flow the work through.

Stacy Gordon

Sure. Well, I think that one big takeaway for me is if people aren’t open to change, they’re in big trouble because I think we all have to look about how we change from a very personal level and how we do work and productivity to an organizational level and at a competitive level and things like that. So I definitely think people better be ready for change for sure.

John Greisner

And there’s a little bit more summer left for your summer reading, and if you haven’t had a chance, you can first read the article, but also read the book the goal that this is based on.

Stacy Gordon

I mean, listen, John, we cannot put too much on my reading to-do list. Let’s be clear. I mean, I’m a girl of $2 words. We talk about it all the time, but I felt like it was very interesting. I’m going to put it at the top of the list, which I have two books on it. So I think we’ll definitely be able to get there. So now I’m going to turn it back to you now that I’ve told the whole world about my maximum two book reading list. I always like to ask my guests two questions. I am a big energy drink fan, and so I’d like to know from my guests, what’s your energy drink of choice and why?

John Greisner

I’m going to throw you off a little bit. It’s probably water.

Stacy Gordon

Water. Oh my gosh. And the answer I never thought I would get, water. Okay.

John Greisner

And maybe at best Gatorade. So that’s my-You’re not

Stacy Gordon

Even

John Greisner

A coffee drinker? I drink too much coffee in the morning, so I have to make sure I stay hydrated in the afternoon.

Stacy Gordon

Oh, another healthy guy. I love it. Okay. I will just on the side stick to drinking the bad energy drinks. Okay. Now the real question, this is my favorite. If you were invited to pick a topic to present on at a conference, what’s a topic you would choose that would get you booed at the industry conference?

John Greisner

That AI adoption is easy.

Stacy Gordon

I mean, you definitely get some giggles for sure. Yeah,

John Greisner

Definitely.

Stacy Gordon

Especially since what’s the actual number, less than 20% of enterprise AI efforts are actually in production today. So I’m going to make sure I get a ticket to that conference for sure. Well, that is a wrap everybody on our topic today. The bottleneck moved, your operating model hasn’t. John, if listeners wanted to find this article or get in touch with you, how would they do that?

John Greisner

Yeah, you can go to the liminalarc.co website and under resources, you’ll find a series of blogs, and this is one of them to what Stacy mentioned. We do try to publish pretty regularly industry perspectives and helps as people are trying to gravel with some of this information. So definitely encourage you to sign up for it and take a read.

Stacy Gordon

Awesome. John, again, thank you so much and thanks to all of you who tuned in. If this made you side-eye your own company a little bit, good. That was the plan. Share our conversation with the one person who needs to hear it. Hit follow so our next conversation finds you, and I’ll catch you next time. Same chair, more caffeine and a new topic. Take care.