Rewiring the Operating Model for AI

Hey everyone, welcome back here to Techstrong TV. This next guest, this gentleman, you know what? He used to be a staple on Techstrong TV.

I used to have him on at least every other month or so it seemed. He's not been on here a while. He's accomplished author, CEO founder, now a sort of a entrepreneur or expert, AI expert in residence.

I want to introduce you to Dr. Mik Kersten. Mik, how are you, man?

It's great to see you. I am very well, Alan. It's great to be here with you as well again.

It has been a while. I've been writing this book in a hole for what feels like a couple years now. It's all good.

Well worth it, I'm sure. So Mik, I want to jump into the book right away, but not everyone may be familiar. I glossed over a lot of what you did there.

Let's spend a quick minute or two, though, give people a little of your background. Oh, sure. Yeah.

So I spent kind of a decade as an open source developer writing a lot of code and building a lot of, I think for me, kind of neat architectures and open source systems. Then founded a company around this new category of value stream management, which was the work that came out of my PhD thesis of kind of how to connect teams with architecture and value streams and really new ways of delivering that connecting value. And then, we built a company around that called Tasktop, sold that to Planview, I'm still supporting Planview as an executive in the residence, but really my focus for the last few years has just been entirely around AI and then writing this new book, "Output to Outcome," on how companies need to completely pivot and adjust and rewire their operating model to make use of any of the benefits at a really at a business and customer level that are possible, so.

I love it. Mik, I feel like we'd be doing an injustice if we didn't mention the first book, though, because it really was, for a lot of people, not a Bible, but a real guide to moving up the, not the food chain, but modernizing how they look at things. So if you wouldn't mind, spend a quick minute on your first book.

Yeah. So "Project to Product," right? It's been I think around eight years now since that book went out, and I think it's- Sure ...

I was kind of really surprised at how much the timing and the impact that book was able to have. I saw it back in number one bestsellers a couple weeks ago on Amazon. I didn't realize, it's still, I think, helping people on the journey of moving away from treating technology and teams as this cost center that you throw work to and actually treating them as a core part of the business.

What I realized with "Project to Product" is that some organizations did this rewiring, right? Where they made technology and teams really central to their entire operating model, not just IT, right? But it became how the companies delivered value.

And then as we started seeing what was going on with AI and the productivity amplification from AI, I realized that the teams that had created these product value streams and learned how to measure flow, the things that "Project to Product" outlined, they were actually doing much better at leveraging the AI amplifications. Whereas companies who had these sort of legacy structures, or just too many structures, right? Some companies had their traditional org chart, then they have Agile deployed as this matrix over the org chart, these two or three sets of ceremonies, and all of that was getting in the way of getting that benefit from AI.

Sort of like when you think back 30 years, 40 years, there was this digital paradox, right? Some companies were wired to make use of digital and internet and software, others weren't. And so I realized it was really time to revisit, what does an operating model look like that's actually going to support all of this productivity amplification with AI?

And that's really what the new book's all about. So I want to make sure people out here get this straight. What's the name of the new book, Mik?

" So the idea is that in the world of, in this age of AI, outputs actually become very cheap, right? They're kind of, in the end, they'll be the cost of Kimi or of electricity, or it's not free, but it no longer takes months for teams to develop a mobile application. It takes hours.

So all of a sudden, organizations who were wired entirely around this main constraint of their delivery capacity, everything being around prioritization and kind of very sequential investing, all of those things no longer matter. It's no longer around, you're no longer competitive because you can deliver more output, because output is becoming cheaper and cheaper and more and more scalable. It's really now around delivering outcomes.

And so imagine if you were a car company CEO and all of a sudden you could deliver an infinite number of cars over the course, well, not infinite, but let's say 1,000 times more, 100 times more cars. That's really what's happening in the organizations today. And I did this kind of deeper analysis.

This is something I think that's kind of near and dear to you as well, Alan, as you've looked over these technological ages. Each technological revolution and age has had some kind of constraint on production, right? Initially, it was kind of manual labor, and then we got steam- Yeah ...

and then we got electricity. The constraints from the last age were really around developer and knowledge worker productivity, right? That was the main constraint on delivering value.

And in the age of AI, that constraint is gone. So I think- Yeah ... all operating models and management systems that are built entirely around managing that constraint are now legacy.

They're no longer valid, and we need to shift from managing outcomes and how many features and initiatives teams deliver, really demanding those outputs, we need to shift to now to managing outcomes. I call this the AI scale issue, Mik. And what we're living in now, because it's early in this AI era that we're entering into, is it's very lumpy where AI is .

So all of a sudden, AI touches down here, and boom, it scales up output more than our system was able to handle it. But theory of constraints, it creates bottlenecks, one after the next, as we seek to sort of bring the rest of the process and organization up to that AI scale. And so it makes the AI almost non-efficient.

Because the scale of where it touched down kind of broke everything around it, because we don't have enough humans to do it. Look, Mythos, security vulnerabilities. Wow, all of a sudden, we can find more vulnerabilities than we ever could find before.

Now what? Now we've got to fix some or remediate them, mediate them. We're not equipped to do that.

We're not equipped to do it. So there's this disruption that, it's growing pains is what I think, but it's really a question of AI scale. We were talking before about electricity.

So when steam engines first came about with the Industrial Revolution, and then we invented electric engines, Mik, it took 40 years for them to say, you just don't drop an electric engine into where a steam engine was, because steam engines, you had one steam engine with a lot of pulleys and belts and stuff. Yeah, and centralized. In the middle of- Yeah.

Right. It was centralized. Electric engines, we put all through the factory where it makes sense.

Yeah. And all of a sudden, boom, that's when you get your 10x, 40x, whatever, bump. We got the same thing here with AI.

That's right. We've got to figure out how to do that. So is that the kind of thing in your book that you're talking about?

That's exactly it. I think I actually have that exact comparison in the book. Because that's what companies are doing.

They're actually putting, it's exactly that. They're taking the new engine, which is electricity, and they're replacing the steam engine with it. And so what's happening is teams are now kind of given AI, they can produce 10 times more, let's say soon they'll be able almost 100 times more with proper agentic engineering.

Yeah. But it's not turning into the business outcomes the way it's expected. It's consuming a fair amount of cost because those tokens are not cheap.

And so I think, Alan, the core premise of "Output to Outcome" is actually what you just said earlier. It is, if we apply the theory of constraints, I think the theory of constraints always applies to any complex system. If we apply the theory of constraints to kind of end-to-end value delivery of the organization, with AI, and agents, we've removed the constraint on team productivity.

That used to be the main constraint. It's now gone. So the question just becomes, what is the new constraint?

There's always a bottleneck. There's always a constraint. And in all the research I did behind the book, it's actually the organizational structure is the main constraint.

Yeah. Just the same way that centralized steam engine, or just the centralized assembly line, was the constraint. Until you actually put electricity all along the line, you weren't getting those benefits of electricity.

So you had this other productivity paradox, just like we have the AI productivity paradox now. " And it is going to be the organizational coordination costs and approvals and budgeting and planning and funding processes. Let's measure and identify the constraint, and let's create a new kind of AI native organizational structure where that constraint is no longer the constraint, where you're able to leverage that massive amount of throughput.

But if you actually pour AI onto, again, kind of the old assembly line architecture, in your organization, you're going to generate a whole lot of outputs. Chances are that those outputs won't be aligned to your strategy, so it'll be kind of slop and not substance. And you'll spend a lot without getting a lot for it.

Whereas we now already see organizations who are structured around getting that AI amplification. I go through several case studies in the book of those as well. So that's really my goal is I think, we're seeing right now some organizations pull ahead so fast because they have these structures.

Back to the electricity example, they're decentralized. Teams have to be empowered at pretty low levels for you to move at the speed of AI. You can't go and centralize and ask permissions to senior leaders in your company to do things like you used to before, because you can now do five times the work or 10 things in parallel.

Whereas before, you'd only deliver a feature and an application at a time. So it kind of is similar to electricity because it has to do with the decentralization of empowerment and ownership in the organization. And really, I think companies and businesses and nonprofits and government institutions who adopt these structures will just move 10 times faster than the ones who won't.

And then, of course, we get that creative disruption. My goal, of course, is to help the incumbents and established organizations move as quickly as they can into these AI native organizational structures. I love it.

I love it, Mik, and I get it. I see it here. I talk to people every day, and I hear these stories.

So I think it's not a stretch of the imagination to see what you're saying here is absolutely right. Who would you recommend is the audience for the book, Mik? Yeah.

The book is really targeted at organizational leaders, so the people who are responsible, but all the way down to team leadership. Where, in the end, if you see your team is moving faster, but the organization's constraining you, my goal is that the book... And the book has these mental models that can, I think, help leaders identify, okay, what is the constraint?

Has this notion of an outcome loop, where we actually apply the theory of constraints to end-to-end outcomes into this fast feedback and learning loop. So it's leaders at every level of the organization, and all the way up to the CEO, because fundamentally, the organizational structure and design is the CEO's responsibility. And so, we can have leaders changing the organization within their portion of it, and it provides guidance and mechanisms for that.

But to really move to what the book proposes, which is this single structure called an outcome tree, where the entire organization's able to move at the speed of AI, that fundamentally rolls all the way up to the top levels of the organization. I love this. Mik, is the book out already?

Yeah. Yeah, the book just came out last week. So it's been great- Okay ...

to see all of the chatter and conversation of-- I forgot that LinkedIn can actually have very meaningful comments threads. So I've been commenting all over the place. It's not easy.

Yeah. It's been great. But yes, every once in a while- Exactly ...

you can do that on LinkedIn. I'm assuming the book's available on Amazon, Barnes & Noble- Yes ... all the places wherever you get books.

Yeah. Everywhere you go. E-edition and soft cover at least.

That's right. There's the e-book. I don't read the e-book.

It's a much more compelling voice actor than I would pretend to be. So yeah, you can get "Output to Outcome" wherever you like to buy and read your books. So...

I love it, man. Hey, Mik. First of all, welcome back to Techstrong TV.

I hope I don't have to wait for you to write another book to come back on here. But best of luck, and I'm gonna go check it out. I'll have to go read this.

It sounds like there's some similarities in stuff I've been working on. So maybe- Yeah. And there are- ...

there's something there to learn- I think some deeper ones ... I can sift through in my edits. Yeah, exactly.

Really. Exactly. We'd love to get your thoughts and everyone's thoughts on it.

I think we're all trying to, again, help organizations adopt this kind of new pace of work. I think it can be a significant lift, and my hope is that the book... The book, by the way, has these seven shifts, and the goal is that you can apply each of those shifts, like slop to substance is a shift.

Matrix to modularity is a shift. So you can just apply those to wherever your main constraints are. And again, it's really significant work that people have to undertake right now.

So I hope it makes it just a little bit easier. So... I love it.

All right. Dr. Mik Kersten here on Techstrong TV.

We're gonna take a break. We'll be back with more. Go check the book out, though, right now.

Mik, the title again is? " Thanks so much, Al. Go check it out.

We're gonna take a break. We'll be back.