5 Mission Critical Scenarios

00;00;03;15 – 00;00;32;13
Nick Powers
It’s a really interesting time right now in the Air Force. We have a lot of complexity and a lot of modernization in front of us. As the Secretary of Defense has said, he wants to go faster. He wants to get capability in place. We have to be able to compete with our near-peer adversaries. There are a lot of things going on: a lot of legacy systems, a lot of COTS products trying to be brought online, and a lot of integration opportunities.

00;00;32;13 – 00;00;57;27
Nick Powers
I’m really excited today because we have our new team member in our group today, Joe Monaco, retired colonel, U.S. Air Force, who has lived this recently, too. I think it’s really important that we have a conversation around the complexity of solving some of these challenges that the Air Force faces today.

00;00;57;28 – 00;01;21;16
Nick Powers
I think it starts with organizing and identifying the problems. Joe, I think you’ve lived this, like I said. Before we get into the dissection and the solution of a problem, talk a little bit about framing. How is that framed? How do you get a better understanding of what we really need to do to get after something?

00;01;21;17 – 00;01;47;22
Joe Monaco
Absolutely, Nick. Before we talk about how to frame it, let’s just talk about complexity in general. It’s something I’ve been thinking about a lot lately: chaos and complexity. People ask, “What keeps you up at night?” Well, it’s not complexity, which might be a surprise, right? We all deal with complexity every single day: weather patterns, our families, and things in our personal lives.

00;01;47;23 – 00;01;49;21
Joe Monaco
The stock market, right?

00;01;49;22 – 00;01;50;16
Joe Monaco
It’s all around us.

00;01;50;17 – 00;02;26;04
Joe Monaco
It’s all around us. But for some reason, it’s human nature, and even more so across the Department of War, to be surprised or intimidated by complexity. We should embrace it, stop trying to manage it, and just acknowledge that we live inside of multiple complex adaptive systems. One of the things we need to think about as we look at complex problems all around us is a systems-based approach using systems-thinking principles.

00;02;26;06 – 00;02;52;12
Joe Monaco
That means being able to frame the current environment of a complex adaptive system that has internal and external feedback loops, properties of emergence, and a number of actors and agents throughout what is really a system of systems. It means being able to take that environment and then acknowledge, “Here’s where we’re trying to go.” Just by doing that, you’re going to frame the problem.

00;02;52;12 – 00;03;38;13
Joe Monaco
And you don’t just frame the problem one time. You create a model-based system architecture where you can continually reframe the problem to produce optionality. At the end of the day, for the warfighter, we’re trying to harness complexity and create optionality, which equals decision advantage in the operational environment for our warfighter. I know you know this: with our pacing threat, the Indo-Pacific is increasingly becoming more complex across multiple warfighting domains: air, maritime, land, cyber, space, and the human domain, if you will, which surrounds all of those.

00;03;38;15 – 00;04;09;10
Joe Monaco
There is a moral, mental, and physical aspect to this, to quote J.F.C. Fuller, a military strategist, that we really have to acknowledge when we look at how we do digital transformation. Digital transformation is really taking legacy systems and modernizing them, which inherently drives complexity because you have constraints with the legacy systems on what you can and cannot do.

00;04;09;12 – 00;04;30;01
Joe Monaco
You have boundaries you have to operate within, but then you need to layer on new technologies and artificial intelligence and machine learning. I know that’s a very long-winded answer, but those are the things I think about at night. It doesn’t keep me up at night, but I’m not intimidated by complexity.

00;04;30;03 – 00;04;42;16
Joe Monaco
I acknowledge that we live inside of it in our personal lives, and our customers can use us to embrace complexity and use it to their advantage for optionality.

00;04;42;18 – 00;05;07;25
Nick Powers
I like your analogy there with regard to how we live in complexity every day, right? I think we almost undersell ourselves a little bit because of what we deal with. We’re always channel-switching. We’re moving across multiple priorities and things like that. Yet when we come to this question of, “Okay, we have a legacy system that has to get from here to there,” all of a sudden, we’re frozen for years at a time.

00;05;07;27 – 00;05;25;29
Nick Powers
We can’t seem to make the right decisions. We can’t seem to get off our tail ends in order to move things forward. Yet every day, we’re making decisions in the blink of an eye that sometimes have even bigger implications. “Should I let my daughter drive home with that kid from school?” There are a lot of things that could go wrong there, right?

00;05;25;29 – 00;05;51;09
Nick Powers
And yet we can’t get an operational system into the cloud. I love that analogy. I think that’s a great way of talking about it. But even better, I like your reference to systems thinking. I know one of the things we’re really excited about is your background in some of that and some of the things you’ve learned in the Air Force. Can you talk a little bit more about why systems thinking is really important for breaking down complex problems?

00;05;51;10 – 00;06;36;04
Joe Monaco
Yes. Systems thinking is absolutely crucial, not to fully understand a complex adaptive system, but to be able to map it and continually map that system. The environments we are operating within, we’ll call them the digital environments, and the customers we are engaging with are moving targets. In the Air Force and across the joint force, they teach systems thinking as the theory, and then they teach operational design as the methodology to frame the environment, the problem, and the approach to get after that, which is an iterative approach.

00;06;36;04 – 00;07;07;16
Joe Monaco
In the context of digital transformation, using agile empirical processes would be a corollary to that methodology. On the front end of that, systems thinking is the theory that helps us understand an emergent problem, or even a static problem that can become dynamic, right? If we wake a sleeping giant or go to war with a peer adversary, so on and so forth.

00;07;07;19 – 00;07;29;17
Joe Monaco
But our adversaries are also doing this as well. Systems thinking is the theory. Design methodology is putting it into practice as reflective practitioners, which is something the U.S. military, especially through its professional military education, teaches in every course I’ve been to over my 24-year career.

00;07;29;19 – 00;07;50;10
Nick Powers
Can you give us an example, maybe touching on some of those points, related or not related to your time in the Air Force? What I’m curious about is a layman’s walkthrough of, “Okay, I have this problem, and this is what I would do.”

00;07;50;10 – 00;08;06;06
Nick Powers
And the outcomes of that effort would result in this, and then I would move that forward and do some design thinking, and then that moves forward from there. Do you have a quick example of how this applies so people can understand what that means?

00;08;06;07 – 00;08;49;06
Joe Monaco
Sure. In a previous life, I was involved in a project where we were trying to layer on AI tools to make sense of human language, more specifically operator language, in the context of global command and control for logistics. To do that, we had to take a systems-based approach and do operational design to actually figure out the tendencies of the controller-operator system, if you will: how they communicate, what causes them stress, and all the different elements that go into that system.

00;08;49;08 – 00;09;18;06
Joe Monaco
If we were to take deterministic models and try to put that on top of that language, which we tried, it would be complete garbage for the output. What we ended up doing was taking a representative set of data and handing it off to AFRL. They were able to make sense of how we communicate and figure out the best technical solution from the plethora of large language models that are out there now.

00;09;18;09 – 00;09;54;12
Joe Monaco
They are commercial partners, right? They were able to build a transition package to hand off to a software factory. That was our pipeline for DevSecOps to be able to make sense of controller-operator communication. Really, the output of that was an indeterminate model where an AI agent could sense stress, safety-of-flight issues, and all those types of items that a more linear way of approaching the problem would not have produced.

00;09;54;13 – 00;10;23;03
Nick Powers
I want to go back to something you mentioned that I think is very important as we try to conquer a lot of our rudimentary processes with AI and automation, which is that it was really important to model behavior, right? Sit down with the operator, understand what they’re doing, even looking at button clicks, looking at when they pick up the phone, and looking at when they go to chat in order to really understand what they do.

00;10;23;04 – 00;10;34;22
Nick Powers
So you modeled that behavior, and then using that model, you went and selected the appropriate LLM, because not all LLMs are created equal, right?

00;10;34;24 – 00;10;37;03
Joe Monaco
That’s right. Yeah.

00;10;37;04 – 00;11;26;16
Joe Monaco
If we think of a change model from somebody like John Kotter, he’s got seven steps of change, right? Step four or five is where we start going out to solution ways to optimize the problem solution. To your point, that’s exactly what we did. We teamed with the operator, figured out what their pain points were while they were on shift, and where they had scar tissue from real-world operations and large exercises. There are patterns here that occur in these more closed systems of how we execute missions and talk to each other, so we can use closed systems to inform open systems and indeterminate models, essentially.

00;11;26;18 – 00;11;50;26
Joe Monaco
That’s exactly what we did. By framing the problem and looking at the architecture that the data will ride on, with the unstructured and structured data, we’re really creating the fuel for the artificial intelligence solution, because you can’t do AI without data, and you can’t do AI without data that’s relevant to the persona that’s going to be prompting the AI.

00;11;50;28 – 00;12;16;14
Joe Monaco
There is no prescriptive formula for this in terms of the order of how you build the pipeline. What’s important is building the foundation and having that systems-thinking approach to what this customer needs, what their current system is, and what other systems they will interact with for the actual data part of it and the digital part of it.

00;12;16;14 – 00;12;43;21
Joe Monaco
Then there is another layer of complexity we haven’t talked about yet today: the actual organizations between all these entities. In the military context, we’re talking different services, different wings in the Pacific, the major commands, the combatant commands, geographic and functional commands, the air operations centers, theater operations centers, and then the functional ones.

00;12;43;24 – 00;13;02;29
Joe Monaco
All of these organizations themselves are complex adaptive systems. They have their own way of planning, tasking, executing, and assessing. You and I are complex adaptive systems. You’re a better golfer than I am. You’re probably better at trading stocks and options than I am.

00;13;03;00 – 00;13;04;00
Nick Powers
Maybe. Maybe not.

00;13;04;01 – 00;13;22;26
Joe Monaco
Well, we should talk about that sometime. That’s kind of the point. It’s a layered approach to systems. To be able to do digital transformation, you need to at least map that at a foundational level and continually refine those systems to be able to field working capability.

00;13;22;27 – 00;13;45;17
Nick Powers
Yeah, and I think you’re hammering a point forward, which is that not only are you looking at – and we all have to realize this – there are so many things that connect to other things, right? That’s the systems-thinking approach. As you mentioned, even though that operator might be logging into one system, that system is getting its data from 10 other systems.

00;13;45;17 – 00;14;07;10
Nick Powers
As you continually model that and map that, you start seeing all of those interactions. You start seeing those data flows. You start seeing the manipulations happening on that data. You’re seeing what the operator is applying and, by the way, how that flows back to those other 10 systems. If you don’t have a good map of the playing field, it’s really hard to solve a problem.

00;14;07;11 – 00;14;19;29
Nick Powers
What kind of maps can we create of these things in the playing field? Is there a specific methodology that is being adapted that really gives you that layout?

00;14;20;00 – 00;14;47;05
Joe Monaco
For one thing, we need to think in loops and not linear lines, roadmaps, and network maps. Those have value. I call it – well, a lot of people call it – reductionist logic, right? Reductionism. At some point, you’ve got to take this complexity and be able to reduce it down into something that tells a story.

00;14;47;07 – 00;15;20;07
Joe Monaco
This is how we’re digitally transforming. This is how we’re going to move data from CONUS into theater to the operator at the edge. But there is a lot of complex work through systems thinking and design that happens to do that reductionism. Using models from John Boyd, Frans Osinga, or systems thinking, there is a lot of good literature out there about this, and it has been codified in a lot of military doctrine.

00;15;20;07 – 00;15;56;05
Joe Monaco
If we just follow our own doctrine, even when it comes to artificial intelligence, the Air Force has Air Force doctrine that talks about a phased approach to AI: narrow AI versus general AI, things like that. People try to jump to this panacea of a solution set, and that really works against the transformation, because you can’t take a legacy system and immediately put state-of-the-art AI, a generative model agent, on top of that.

00;15;56;06 – 00;16;27;17
Joe Monaco
There is so much work that has to be done. Within that path to get from the current state to that future state, where it’s ready for future AI solutions, there is a lot of iteration that has to happen. In short, become more iterative thinkers, more adaptive thinkers, and really harness multiple approaches to solving complex problems, if they can even be solved, which we haven’t talked about yet.

00;16;27;20 – 00;16;29;20
Joe Monaco
Oh boy. That’s another part.

00;16;29;21 – 00;16;58;28
Nick Powers
It’s funny, and I’m going a little off script here, but you bring up something that I know I’ve started getting a much better appreciation for, which is that our military service has put a lot of time and effort into developing policy, methodologies, and frameworks. I would say only a very small fraction of people doing the work actually go and read, understand, and digest those.

00;16;58;28 – 00;17;24;06
Nick Powers
You brought up a few examples of these documents that are available to help technologists implement solutions and really think differently about how to implement those solutions. There’s a lot of value in gathering and pulling that stuff together. I know you just kind of have a knack for finding this stuff.

00;17;24;06 – 00;17;45;27
Nick Powers
But is there a way that a regular person could get a little bit better integrated with how the Air Force maybe wants you to do AI or software development? How do you go find those documents? How do you pull those in? I really do feel like it’s something that people miss a lot, and it could save them a lot of time.

00;17;45;28 – 00;18;22;08
Joe Monaco
There are a lot of resources out there today, especially with the internet and AI, to be able to quickly link things together and harness all these capabilities from research. Really, what I do is talk to people and do a lot of reading. When it comes to strategic planning, one of my favorite books – I’m going to show the nerd in me here – is The Rise and Fall of Strategic Planning by Henry Mintzberg.

00;18;22;10 – 00;18;51;13
Joe Monaco
What he’s really doing in that book is drawing a very healthy dichotomy between a strategist and a planner, and in the context of whether strategy should drive plans or plans should drive strategy. The answer is yes, right? Because it’s this iterative thing that’s going on. I already talked about Frans Osinga. He was a Dutch Air Force officer who studied John Boyd.

00;18;51;15 – 00;19;16;16
Joe Monaco
What Boyd wrote and lectured about in the 1980s with his OODA loop is really good work because that puts more of a scientific spin on all of this, where you can put your finger on probabilities and concepts of emergence and things like that. I think there’s a theoretical part of this.

00;19;16;19 – 00;20;07;03
Joe Monaco
There is a historical part of this. There is a doctrinal part of this. Then there is the experience and philosophy part, which you only get by talking to humans, customers, clients, and operators that have flown their airplane without a certain capability, or a command-and-control operator who has been in an exercise in the first island chain and can tell you, “It would have been nice if I had this capability.” Even stuff back here in garrison, looking at our command and control, what you can see and not see 2,000 or 3,000 miles away from the fight, that is really where you learn and become a learning organization informed by systems thinking and operational design, where you can

00;20;07;10 – 00;20;10;26
Joe Monaco
manage and solve these complex problems.

0;20;10;28 – 00;20;45;04
Nick Powers
That’s awesome. You talked a lot about factors that go into building up an organization, but as you and I both know, culture is super important here. As we talk about digital modernization, we talk about adapting and thinking differently about how we solve our problems. Modifying organizational behavior probably starts first with modeling it before we try to start tweaking it and moving on it.

00;20;45;04 – 00;20;52;06
Nick Powers
Talk a little bit about how important that is as part of the entire process and solving these complex problems.

00;20;52;06 – 00;21;27;13
Joe Monaco
It’s probably the most – it is the most important part of the equation. It’s why I’m sitting in this chair right now at UNCOMN, right? This is a company that loves its people, loves problems, and has a culture supporting its people and supporting the love of problems. Being able to have a flatter organization where you can do stand-ups, share ideas, and participate in agile ceremonies, not just for software development but for other forms of development across the company’s lines of service, is extremely important.

00;21;27;15 – 00;21;57;12
Joe Monaco
It relates to where I found myself when I was in the service, both as an agile participant in a software development pipeline and as an end user. That brought out a lot of culture. What that taught me is that even with the most sound approach and understanding of the systems at play and the data, there is still a lot of organizational inertia to overcome.

00;21;57;12 – 00;22;27;21
Joe Monaco
Even if you live with the people, you know the people, and you sit right next to them in the same building, there is another layer to doing it from an industry perspective. I wanted to go to the right place where I had the freedom of maneuver and the freedom to digitally transform and innovate in a way that would be closest to what I experienced from the inside of the military organization.

00;22;27;22 – 00;23;02;00
Joe Monaco
That’s right here. I’m extremely blessed, and how that’s going to play out, I’ll let you know because I’m only two weeks into the job. But I can sense already that there is a clear sense of purpose here. There are amazing people: a humble, approachable, credible workforce. It’s easy to fit in and get right after the problem, which, as we’ve described in the previous 30 minutes, is a complex adaptive problem with a number of systems that need to continually be mapped.

00;23;02;00 – 00;23;06;24
Joe Monaco
We have the right DNA and approach to do that here.

00;23;06;26 – 00;23;27;13
Nick Powers
Absolutely. We’ve seen some progress in the last several years: you from the inside, us from the outside a little bit. As we continue to progress the modernization of a lot of Air Force capabilities, there’s a long way to go, obviously. But the good news is it feels like progress is being made.

00;23;27;13 – 00;24;00;03
Nick Powers
There have been some great steps forward, for instance, an architecture-first approach to how we develop our capabilities. As a software developer in my previous days, one thing I never had was a blueprint of what I was getting ready to tackle. What I had was a set of code in front of me, with maybe a requirement that was a paragraph that said, “I’m a user in the AOC, and I have this problem, and I need this button to be changed in order for it to work better.”

00;24;00;05 – 00;24;21;05
Nick Powers
But at the end of the day, all the connective tissue behind that problem or requirement wasn’t really documented. I would go in and do the best I could to understand everything that was going on. But at the end of the day, I overlooked something, I didn’t understand something, or I didn’t ask the right question.

00;24;21;05 – 00;24;51;17
Nick Powers
And I put the project in jeopardy, potentially. This architecture-first approach tries to take a lot of that out of the equation. We have folks go out, get use cases, develop interface diagrams, give overviews of system functionality and business-process flow, and really try to create what I call the blueprint for how to deliver that new change or that new capability.

00;24;51;19 – 00;25;11;25
Nick Powers
Have you seen, or maybe are you just encouraged by, the progress that has been made since you’re recently retired? I know some of this has been brought up recently. How do you think architecture first will impact future capability development?

00;25;11;25 – 00;25;47;21
Joe Monaco
I think what that’s going to do is create a starting point with the operator. Customer and user are a lot of different words to describe the same persona. What that’s going to do, and I’ve seen it happen in my past life, is create opportunities in the solution space that were never an original requirement. I would even posit and hypothesize that some of the best capabilities we’ll be able to field for our customer will be ones they’re not even thinking about, and we’re not even thinking about.

00;25;47;21 – 00;26;18;05
Joe Monaco
By having an architecture-first approach, you’ve created that systems analysis and mapping of the system in a way where you can then use creative thinking with the operator to draw the right path through it, create new paths, add in new data sources – whatever it may be. That’s the magic of this approach to digital transformation and digital innovation, whatever you want to call it.

00;26;18;06 – 00;26;49;21
Joe Monaco
Some people call it modernization. I call it transformation. Each word has its nuances, but I like to talk about transformation because it’s really changing the game. That’s what we have to do to enable the Department of War to go up against peer adversaries: the People’s Republic of China, Russia to a lesser degree, but all are formidable threats to our national security and to our allies and partners.

00;26;49;24 – 00;26;50;23
Nick Powers
Absolutely.

00;26;50;24 – 00;27;16;18
Joe Monaco
When I talk about emergence, I’m just quoting Dr. Daji, who has a systems-thinking book out there. There are other systems-thinking theorists and practitioners who have written books about emergence, but I really like that word. To get back to the beginning of our conversation, we live in a complex world. No shit, right? I can’t say that, but no kidding.

00;27;16;20 – 00;27;43;02
Joe Monaco
It’s all around us. I’ve got eight kids. You have kids. We go home and cannot control a lot of things, but we embrace complexity every day. That’s emergent. Every complex adaptive system will have emergent properties that actors and agents, which also change over time, are going to influence from within and from outside the system.

00;27;43;04 – 00;27;46;26
Joe Monaco
That’s what I’m talking about when I talk about emergence.

00;27;46;27 – 00;28;20;04
Nick Powers
I also like – I know it’s not exactly the definition of the word – but you talked a little bit about how, as we map the architecture and understand the interactions of the systems, in those interacting points there are opportunities, right? We call those opportunity points, where things emerge from that. It’s like, “Oh crap. If we would just fix this thing, all of a sudden all these systems work 10 times better,” or maybe I can save 50% of the cost of data storage because it’s all being stored here already.

00;28;20;05 – 00;28;42;08
Nick Powers
Why am I replicating it 20 times? We’ve had some conversations around putting an antenna on an F-35. Well, guess what? They’re going to put that antenna on an F-22 and an F/A-18 in the next two years. Why reinvent the wheel when probably 60 to 70% of that work on the F-35 is the same exact work we do on those other planes?

00;28;42;14 – 00;28;58;19
Nick Powers
As you look and see that and discover those things, that stuff pops up and you’re like, “Oh crap, that’s a whole other way of thinking about it or solving that problem,” or at least making things more efficient.

00;28;58;20 – 00;29;27;05
Joe Monaco
Yeah, I agree. If you’re doing adaptive planning or adaptive delivery the right way, those things should naturally occur. I’ll give you an example. When I was working on a previous project, we integrated a commercial data source from a company into our data ecosystem. It immediately benefited one application, and then two others that we didn’t even think about originally as use cases.

00;29;27;05 – 00;29;52;00
Joe Monaco
For the second and third applications, it had a bigger impact than it did on the primary one we got it for. That is emergence when you take an adaptive approach through design and systems thinking to mapping out a system architecture that has data sources, application layers, AI/ML engines, and all those types of things.

00;29;52;06 – 00;29;54;18
Nick Powers
Yeah, absolutely. That was good, man.