Transcription
Welcome to our special series on Agentic AI for government, brought to you by Salesforce. Artificial intelligence is reshaping some of the Pentagon's most complex and fast-moving mission challenges. From accelerating logistics and strengthening supply chain resilience to enhancing strategic decision-making and identifying vulnerabilities before they become a crisis. AI, throughout, is becoming, uh, rapidly becoming a foundational capability for national defense.
I'm Wyatt Cash with Scoop News Group, and joining us to explore where its technology is making the most significant impact today, and what mission needs its best position to address, and how it can help defense leaders move from reactive operations to proactive readiness, is retired Major General Allen Day, vice president for logistics and sustainment industry strategy for Salesforce's global public sector business. Uh, General Day previously served as director of operations for the Defense Logistics Agency. Uh, General Day, thank you so much for joining us and welcome to the program.
>> It's great to be here. Thank you, Wyatt.
>> In our last episode in this series, we talked about Salesforce's deepening investment in its support of the defense and national security sector, including its rollout of Mission Force. From your perspective, and from your days at the Defense Logistics Agency, how would you describe the significance of Mission Force?
>> Well, thank you for the question, Wyatt. When I served as a J3 at the Defense Logistics Agency, I was responsible for a 41 billion enterprise supporting 2,300 weapon systems globally with about 89,000 orders coming in a day. So we had to solve a lot of different kinds of problems every day to really ensure the mission execution for the warfighter, and that warfighter was often in remote environments where there were few logistics lines of communication. So what I learned there is you cannot manage that level of complexity with stovepiped industrial aid systems. And so how does that affect us when we think about Mission Force? So that was my past.
Now, as I think about Mission Force, you know, it's a new initiative, really came about as a direct response to a question and a request from a senior military officer who told us that he wanted to bring the same speed and efficiency he saw in the private sector. And so this dedicated business unit was stood up, and it's focusing Salesforce's proven platform into the defense, intelligence, and aerospace communities, like you mentioned. So, at DA, we often struggle to adapt commercial best practices to government constraints. Mission Force helps us solve all of that in a lot of different ways. It allows us to optimize mission readiness, logistics, personnel, and decision-making kind of problem sets by tailoring the commercial innovation specifically for the warfighter. That is the mission of Mission Force. And it represents, like you said, a deepening of our investment and commitment into the defense and national security sectors. So, I think bringing the best of what we do in commercial to the Department of War and the public sector in general is really how we're going to get after using AI in a better way in the Department of War. We need to bring our best there. We can't, we can't just give them the things that might work or could work. The defense of this nation and national security demands our best, and that's something that I'm excited for, and I think it's exactly what we're trying to provide through Mission Force to deliver an agentic enterprise at the speed that we demand for our mission in the Department of War.
And just diving a little deeper, how do you see the evolution of AI, uh, impacting the rapidly, uh, changing challenges, as it were, uh, faced by the War Department?
>> We are moving from the information age, where we built dashboards to admire the problem, to the agentic age. The strategic environment has fundamentally shifted. We're moving toward a Department of War posture where the greatest challenge is really achieving decision advantage or decision dominance. People put it different ways, but that's the ability to understand, decide, and act faster than our adversaries. And the challenges today are too fast for manual processing. But Agentic AI doesn't just summarize, summarize the facts. It actually takes action. That's the difference. It allows us to move from reacting to crisis to orchestrating outcomes.
In the past, we would have problems that we would try to find the right data to answer the questions. Sometimes there were my questions that I would ask people, and they would try to find the right answers. And sometimes that took weeks. It's crazy to think about, but it took weeks to stitch together the con, the, the data and the information that came out of that data to get to the back to the decision-maker, sometimes me, in a timely way. With Agentic AI, we can literally fuse live streams of data, right, and get clarity in seconds. And to me, this is really powerful when you think about supply chain challenges that we deal with, right? So you can identify vulnerabilities before they become crises. That's the key. We can't do that at the speed that we need to with a human staff. A human staff empowered by Agentic AI can do that, right? Agentic AI also gives you kind of a limitless digital workforce. So you're, you're not paying for extra humans, but you are getting the benefit of lots of different perspectives from a, a digital workforce, and they can take care of that stuff that nobody wants to do. The routine, routine drudgery, right? The, the things that the humans don't want to do, and we can allow the humans to have the time to focus on the more complex and high-stakes kind of decisions that they need to make every day. So I think Agentic AI is really bringing decision advantage or dominance from a concept into an actual capability.
>> So you mentioned the, uh, opportunities with data. Can you talk about what fundamental needs that you see AI addressing when it comes to data and data management?
>> There are several, but I would like to highlight three fundamental needs for AI to address: readiness, resilience, and trust. The first is readiness. Right now, our service members spend way too much time acting as the human glue, stitching together data from disconnected legacy systems. And the most fundamental need is really freeing up their mental capacity. We need our people focused on the mission, not on form-filling or inventory counting. And we also need a layer that sits above those legacy systems and connects them to form, like, this intuitive human interface. Because today, when you bring a new Airman in, they have to learn the old legacy system. They put their iPhone 15, 16, 17 down, and they bring up the green screen and, and they start learning a system that was developed in the 80s and 90s, which sounds crazy, but it's actually true. So, we need to get them back to the mission. By having this engagement layer approach, we can actually pull the data from all of the systems they may need to see, from finance, supply, and operations, or whatever those systems might be, and they're able to get a full picture, not just a stovepipe picture. That is allowing them to maximize their availability because they're now focused on them versus on all of the minutia of doing their job and stitching together the data.
The second area is resilience. You know, we operate in a contested logistics environment everywhere, really. Once you leave the fence line of the base and you go off to do the mission at some far away land, or even in some different part of the United States, the supply chains that you rely on are contested. And it could be a digital supply chain, it could be a physical supply chain, it could be a, a supply chain of a different sort. But the bottom line is they are contested. What does that mean? They can be disrupted by either a weather event. They could be disrupted by an adversary trying to disrupt them. There's a lot of different ways that contested logistics comes into play. Resilience has to be built in right from the beginning. And we need to be able to use AI to help us build that resilience and also to execute with resilience.
And the final area is trust. If the warfighter doesn't trust the data or the algorithm that they, that they're given to use, they won't use it. They simply will set it aside and do it the old way because that's what they end up trusting. And I think if we have clean, governed data that doesn't hallucinate, that actually ensures mission success, success, the warfighter is going to use it, right? So readiness, from my perspective, is the what. Resilience is the how, but trust is the why. Because if the warfighter doesn't trust the algorithm, the mission fails before it even starts.
>> Those are some great points. Thank you for that. Um, I'd like to have you kind of translate that now into, so how does that impact logistics and supply chain issues, uh, particularly for the military, and how do you see AI technology enabling, you know, this key mission area?
>> You know, it's, uh, it's interesting. I think back, my, my last two years in the military were during the pandemic. So I was a J3 for DLA during the pandemic, and if nothing else, the pandemic taught us that we cannot take supply chains for granted. Um, we've, we have to trust them to deliver, but sometimes they don't, right? And it's not always the fault of the planner, right? Things can go wrong you never would expect. Additionally, as a chemist and a materials research engineer, it's really gratifying to see, you know, like lithium and nickel and cobalt and manganese, all of these elements that I studied on the periodic table for my, my schooling and in my actual life as a, as a research chemist in the headlines, right? It was kind of cool. It's kind of cool to see that. But that's a direct result again of the pandemic where we didn't care where things came from. Now we care where things come from. And we should care where things come from because, guess what? The precursors and the antecedents of the things that we need are sometimes in the hands of somebody that we're not friends with. Right? So, we got to figure that thing out, and we've got to understand not only where things come from, but what is the, what are the best ways to bring it into our domain, right? Either those of specifically of the United States or of a partner or friendly nation. How do we, how do we bring that all together? That's a key component of the supply chain problem.
The other piece is something I call the physics of logistics. The three immutable constraints. I would say there's probably more, but I talk about three immutable constraints: time, space, and cost. And the physics of logistics talks about how you have to make trades in time to get space and cost, and vice versa. You've got to be able to work this physics problem. And I would say you can't cheat that kind of physics, but AI allows us to balance and manage the tradeoffs much more effectively.
>> And Agentic AI, um, trades information for inventory. So in a lot of ways, it's like if we don't have enough information, we'll pile more stuff to make sure we got enough, right? But if we have more information, we can stop doing the just-in-case kind of inventory to a more predictive, uh, precision inventory. That's where we need to get to because we can't afford the logistics of mass. We need to move to a logistics of precision. That is where AI comes. So I would say, in summary, you cannot cheat the physics of logistics in terms of time, space, and cost. But with AI, you can finally master the equation to put the right part at the right place before the need even arises.
>> That's a great way to, uh, frame that discussion. Appreciate that. So, shifting gears a little, um, from your perspective, how can AI and Agentic AI also help organizations identify vulnerabilities, maybe more effectively, and then avoid the challenges of those vulnerabilities and really shift instead to from reacting to them to being more proactive towards them?
>> Another great question. You know, for me, it comes down to visibility, resilience, and autonomous action. So one of our biggest vulnerabilities lies in our lack of visibility of our supply chain, right? And the lack of resiliency in our defense industrial base. So those two go hand in hand. We can't see, so we can't manage what we can't see. And our defense industrial base is really shrinking. There's less people that want to do business with the government for whatever reason. Some are trying to get in, but a lot of them have kind of lost faith. Or maybe they've just gone out of business because so-and-so and so-and-so got old and decided that they didn't want to have the business anymore, and so-and-so decided that they didn't want to do it either. And so, you know, but there's also adversary actions that are actively targeting our networks, right? So all of these things come together, and we've got to decide if we can actually gain the visibility between our defense industrial base and our suppliers and dig deep. Is that something we want to do? I think the answer is yes.
Because if you think about it, if you have the top level, level one kind of suppliers, that are, you know, your primes, defense primes, you work your way down to level two, level three, and level four and deeper. That level four supplier might provide a key part, but you never see the level four. I mean, you don't pay attention to them because you're paying attention to the one or two or three due to privity of contract issues. And so if that level four goes out of business, they might provide critical parts to the levels above them, and it can have huge unexpected effects on the rest of the supply chain. So how does, how do we, how do we fix this? How do we get beyond it? This is where I think integrating data right from the right kinds of sources allows us to kind of get X-ray vision, and I think AI can help us do that. We've got to connect the data sources, and we can get, like, a partner 360, right? A DIB, a defense industrial-based partner 360. So, not only can we see who's in the market, who's doing some of the things that we need to do, right? They may not be in the defense market, but they might be in a commercial market, but they do the same kinds of things that we need them to do. If we have a broader view like that, we can incentivize to bring them in. But then we can also look down deep into the supply chains using AI tools to see things that a human may not see. Foreign influence, cyber risks, deep in the sub-tiers, right? Where adversaries are sometimes lurking and and disrupting our supply chains before we know that they're doing it, right? So this is the key.
When you ask a human to see something, it's very difficult sometimes for them to spot a 2% deviation in the supply chain flow, right? A human's going to maybe miss that. But AI can see pattern recognition very well. So where a human, right, is going to have to recover, you know, the problem after they react to it, an AI is going to see it coming, potentially, right? And flag those things, v potential vulnerabilities for the human to take a look at. This is the power of Agentic AI. And I think, I know when I was in the DA and other, it doesn't matter, every organization I've been in, to include the one I'm in now, you know, we have dashboards and we always look for the red and try to fix the red, and we're, we're kind of reacting to the dashboard at sometimes, right? I think AI agents actually can prevent the red from ever happening. And that's a superpower that I think we need to put in the hands of our department, Department of War, to be able to see the patterns, see the minor deviations that could become bigger over time, and deliver better effects so they can get ahead of the problems and not react to them. Thanks.
And then lastly, looking, uh, towards the horizon, what, what do you see changing in the AI landscape from your new perspective, uh, and, and how do you see that, uh, likely to impact the defense sector?
>> I think the future horizon is defined by two primary things: trusted action-taking, trusted action, and getting closer to the tactical edge. Our warfighters are on the tactical edge. That's where they operate. Not every day, but we have people on the tactical edge every day. Our AI tools need to be able to operate there as well as every single day in normal operations. We need to be able to do both, and trust is the foundation for both of it, right? So if, if we, we need to get beyond chatbots, we need to get beyond simple LLM kind of actions, which are helpful, but they're not warfighting capabilities. We need to be able to get clean, connected, trusted data, right? And bring it together in such a way that we can trust its answers and start to take action faster than our adversaries. That is trusted action. We also need to be able to operate closer to the tactical edge. And that is an, that is an area we call DDL: disconnected, degraded, intermittent, and low bandwidth environments. DDL is hard. That is hard because if it's disconnected, cloud systems don't necessarily work in that operation. But we have solutions that can operate in the DDL environments, and then re, as they get reconnected, they can, they can synchronize the data. And I think the company that brings this together, the, the solution that comes together, is where we need to be right now because that's where we're going to help the Department of War the most, right? The future belongs to those who can treat context as well, like, can we have the right context around our data to one, know that it's trusted, and also that it is something we can rely on even at the edge? Are we able to push it that far? So I think the future isn't really AI replacing a warfighter, but it's enabling a warfighter. It's empowering the warfighter. It's making the warfighter better because they're using the tools that make them smarter, faster, and more capable of making the right decision in a time of relevance.
>> Well, uh, General Allen Day, thank you very much for joining us today and in particular sharing your insights from your perspective in the military, uh, how AI may likely impact the world of logistics and supply chains and the defense sector at large. So, thank you so much for being with us. Appreciate it.
>> You bet.