
All Quiet on the Second Front · 39 min
Key moments - from our scoring
Substance score
61 / 100
Five dimensions, 20 points each
The partnership between Microsoft, Armada, and DISA tackles the Pentagon's most pressing infrastructure challenge: extending cloud and AI capabilities to forward-deployed units without sacrificing performance or security. Rather than building dozens of bespoke solutions, the three organizations are creating a foundational platform that can be customized for maritime, aviation, and land-based scenarios. The key innovation is enabling low-latency data processing at the edge - critical for life-or-death decisions - while maintaining connectivity to cloud-based model training in Azure. SPEAKER_00 (Armada's founding CTO, formerly of Azure) emphasizes that mission owners shouldn't think about infrastructure; instead, automated deployments and a marketplace approach - similar to Azure Marketplace - will let independent software vendors and application developers deploy securely without managing underlying systems. SPEAKER_03 from DISA frames this as moving from tactical point solutions to enterprise-scale deployment, with a marketplace authorization model where the platform itself is approved once, then individual capabilities get audited separately, removing the burden of each tactical unit running its own accreditation cycle.
By deploying a foundational edge platform (combining Azure infrastructure, Armada's edge orchestration, and DISA's DevSecOps tools) that mirrors the cloud experience stateside, enabling real-time data processing at the edge while maintaining secure connectivity to cloud-based model training in Azure.
Long IL-5 and IL-6 accreditation cycles, fragmented procurement processes, and complex technical requirements that force every vendor to submit multiple artifact revisions - but marketplace authorization models that approve the platform once and audit services separately can dramatically shorten this path.
Develop enterprise capabilities and approved frameworks that most missions can use, while building rapid assessment mechanisms to authorize tactical-only customizations when units have specialized needs - avoiding both rigid one-size-fits-all solutions and proliferation of incompatible systems.
Azure Local provides the same cloud-like experience, security, and consistency at forward locations that developers get in commercial Azure, removing infrastructure management burden from application developers and enabling single-click deployments across any edge location.
In traditional siloed processes, vendors and government reviews work separately, slowing incident response; shared responsibility means both parties understand data provenance, protection, and deployment, enabling faster collaborative resolution when issues occur.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode delivers a mix of valuable operational and technical insights about edge cloud infrastructure, accreditation processes, and AI deployment challenges, but significant portions devolve into abstract discussion without concrete examples or novel claims. The discussion of reciprocal authorizations and agentic AI accreditation timelines is substantive, but much of the infrastructure and marketplace framing recycles existing cloud-to-edge concepts without surprising detail.
how do we take that out to the edge in a complete way, in a way that enables us to address latency challenges, environmental challenges
18 months at the fees are long. The platform and the app could completely change. So almost I think about the, for example, an app platform. like we are releasing so many changes on a daily basis, right?
While the discussion of reciprocal authorizations and agentic AI accreditation represents somewhat novel framing in defense tech context, the core concepts - cloud migration, edge computing, marketplace models, and accreditation friction - are well-trodden ground. The insight about mission owners deploying faster than enterprise processes can approve is known; the solutions proposed (AI-assisted GRC, machine-readable artifacts) are incremental refinements rather than breakthrough thinking.
They'll go procure it commercially and they're using applications on Azure commercially to prosecute the mission
you can't wait for you know, accredited novice that to look from one to the other. There's going to be changes happening there. There has to be an agentic AI
Three credible practitioners with genuine operational experience: a Microsoft infrastructure executive who built Azure for 11 years, an Armada CTO/founder with platform expertise, and a DISA authority figure managing accreditation at enterprise scale. All three have hands-on experience with the problems discussed, though none are household names or represent the absolute top tier of their respective domains.
I actually was building Azure for over 11 years. So now it's OK, now take a cloud to the edge, right?
we are now open to cloud on the Oconus side of the house, it provides such massive capability to deliver enterprise grade capabilities
The episode lacks named customer examples, specific deployment metrics, concrete timelines, and dollar figures. References to Ukraine, IL-5/IL-6 compliance levels, and general accreditation processes provide some grounding, but most claims remain abstract. The '36 months' and '18 months' deployment cycles are mentioned but not tied to specific programs. No revenue, adoption, or performance numbers are cited.
I had done an 18 month deployment. Come home, reset, recycle, and gone back and was on my tip before that piece of software would be accredited
Ukraine was a watershed event in the world of technology and the application and technology on the battlefield. And we watched how drones were used and some of the counter-drone tactics
The host (Tyler Sweat) asks thoughtful follow-up questions and genuinely probes tensions - speed vs. security, tactical vs. enterprise, reciprocity challenges - but rarely challenges guest assertions directly. Questions are open-ended and exploratory rather than sharp. The host does redirect conversation productively but doesn't push back on vague claims or demand specificity. The final segment asking 'what are you optimistic about' is softball positioning rather than rigorous examination.
How do you make that more accessible to those who need it most?
So how do we enable that broader, sort of, like GRC community? That AOism sort of Scott community to help them scale.
Computed from the transcript - who did the talking, and the words that came up most.
Today, Armada is opening its first Galleon Experience Center in Washington, DC. A live environment showcasing what it actually takes to run AI at the edge - bringing together ruggedized infrastructure, cloud, and secure software deployment. With 2F Frontier as part of the stack, extending DevSecOps to disconnected, real-world environments. And with Offset 2026 right around the corner, it felt like the right time to revisit a timely conversation from last year. On the Offset stage, Tyler Sweatt, Pradeep Nair (Armada), Leigh Madden (Microsoft), and Korie Seville (DISA) got into: The gap between data and decision at the edge What breaks between connectivity, compute, and deployment Why speed to mission still comes down to authorization And if you want to be part of where this is going next, register for Offset 2026 . ______ From oil platforms to special‑ops posts, cloud infrastructures are now being deployed in environments they weren’t initially designed for. This panel explores how customized edge architectures and commercial integrations orchestrate distributed services and securely manage intermittent connectivity.
Transcribed and scored by The B2B Podcast Index.
[SPEAKER_02]: as we've started to understand how to use cloud computing, how to inject AI into the mission, how do we take that out to the edge, and how do we take it out to the edge in a complete way, in a way that enables us to address latency challenges, environmental challenges and number of things, and put a capability forward that is going to look just like the capability that we have back in the United States. [SPEAKER_01]: This is all quiet on the second front. [SPEAKER_01]: Podcasts were boring conversations around defense tech and national security come to die.
[SPEAKER_01]: Join me, Tyler Sweat, and my second front homies. [SPEAKER_01]: As we dismantle the mundane and cut through the bureaucratic bullshit to demystify the world of defense tech. [SPEAKER_01]: But be warned, this is not a typical government podcast. [SPEAKER_01]: Are you ready to get [SPEAKER_01]: Thanks everybody again for being here.
[SPEAKER_01]: Jan, so appreciate you guys coming on. [SPEAKER_01]: I'm excited to chat about not just some of the news that came out this week on the broader partnership, but also I think some of the opportunity that that unlocks for capability for mission and sort of where we look to the futures we think about the department's relationship with software. [SPEAKER_01]: Um, so start sort of first question to lob to Lee, just on the broader sort of partnership and you guys work with Armada and us coming in now and what that means for Mike Soft and sort of as you look at into the future, how you think about the opportunity that unlocks.
[SPEAKER_02]: Well, Tyler, I think that thinking about this partnership in particular, we have brought together, I think, three best-to-breed companies to address the biggest challenge that our DOD is facing, and that is, as we've started to understand how to use cloud computing, how to inject AI into the mission, how do we take that out to the edge, and how do we take it out to the edge in a complete way, in a way that enables us to address [SPEAKER_02]: Latency challenges, environmental challenges, and number of things, and put a capability forward that is going to look just like the capability that we have back in the United States.
[SPEAKER_02]: And I think that's a game changer for us, especially as we look to the Pacific in the coming years. [SPEAKER_01]: Yeah, and I would have dovetail on the latency side of it and sort of the performance and predict and is the founding CTO sort of the brains behind it. [SPEAKER_01]: How do you look at that and think about the performance for requirements and some of the environmental limitations that have impeded that previously? [SPEAKER_01]: What does this start to unlock for some of those low latency requirements?
[SPEAKER_00]: Yeah, I think, at least point, I actually was building Azure for over 11 years. [SPEAKER_00]: So now it's OK, now take a cloud to the edge, right? [SPEAKER_00]: How do you take a cloud to the edge and why is it? [SPEAKER_00]: Because I'm seeing this obviously across multiple industries as well, this huge amount of data being generated, right?
[SPEAKER_00]: at the edge. [SPEAKER_00]: Now, some of them are latency sensitive, so it means that if you are pushing that data to the cloud, then hey, the data is still valuable, but you are not getting the immediate output of that, because you are not processing it on the edge. [SPEAKER_00]: That is where you know, Armada came in and said, hey, we got to make the same cloud like experience to all the warfighters, etc. [SPEAKER_00]: on the edge, so that hey, what are the data they are getting, the signals they are getting, they should immediately be able to get the output and respond [SPEAKER_00]: critically, right?
[SPEAKER_00]: You know, in many cases, it's a life and debt situation. [SPEAKER_00]: So you need that data to be processed immediately. [SPEAKER_00]: While still having connectivity to the cloud is important because you need to, like, still we are running models on the cloud. [SPEAKER_00]: We use actually Azure for example, to run models on the cloud.
[SPEAKER_00]: Train that, and again, inference is super important. [SPEAKER_00]: Hate the train models. [SPEAKER_00]: It has to run on the edge. [SPEAKER_00]: That's, that's where it is.
[SPEAKER_00]: And I think Super Accelerator, the second front because it's one hay reliability being able to process faster. [SPEAKER_00]: At the same time, security compliance is super important as well, right? [SPEAKER_00]: Just because you went from start processing stuff on the edge, it doesn't mean that you lower your security bar. [SPEAKER_00]: And that's where the right partnership is super important to make sure that people can deploy and run applications on the edge, securely.
[SPEAKER_01]: You sit in this really interesting position where, you know, we joke a lot about having to help the department sort of just better understand what we all equated to Cloud 10 years ago. [SPEAKER_01]: And so you've got sort of a migration to the Cloud, but juxtaposed with a mission requirement that demands some of the edge low latency, real forward deployed capability. [SPEAKER_01]: How do you see the department sort of thinking about that and then how do you work to balance both of those?
[SPEAKER_01]: Because it's to somewhat competing tensions there. [SPEAKER_03]: Yeah, it's the typical thing in the department, right? [SPEAKER_03]: Other newer technologies, a lot of the things that kind of are pioneered in the beginning. [SPEAKER_03]: They start out in the tactical use case, you know, they start out in.
[SPEAKER_03]: Well, one organization is doing it this way, they kind of kick it off and then, you know, word amount of spreads and you have about 20 of these things sitting around, they don't interoperate, they don't undercommunicate and then it becomes this, well, this is a really good thing but we need it everywhere and we need to just talk to everything. [SPEAKER_03]: And so I think we are, I think we are starting to, [SPEAKER_03]: come over the curve from tactical beginning of development to okay we know this is a good capability now we need to get cloud deployed universally and we've got it good on the cone aside we've got it good in in the states but it's now extending that outward um and [SPEAKER_03]: When we develop these, you know, mission says we develop these use cases, a lot of times it's thought of in, you know, the vein of my particular mission, my particular use case my particular set of data, my particular data types things of that nature.
[SPEAKER_03]: And so I think where where I sit is having to take those requirements from the different organizations play and understand hey your requirement and your requirement. [SPEAKER_03]: aren't that different from a, you know, technological standpoint, and then figure out how I can strike that type of capability to meet what I'll call the general use cases as far out toward the edges I can, but realizing that at some point, in at some level, emissions that will become tactical and tactical only.
[SPEAKER_03]: So it's being able to balance those two things and deploy a capability that fits most of the mission sets, but then can integrate those tactical sets [SPEAKER_03]: when they're able to do so. [SPEAKER_01]: Yeah, and I think that plays a lot to what we talk about on the requirements out at the tactical edge, but also the need to mirror that performance that either sits state side or sits at home station. [SPEAKER_01]: And there's a little bit of design in there on that user experience thinking through.
[SPEAKER_01]: There's also an aspect where if we can produce some consistency, you can bring adoption, you can bring a little bit more [SPEAKER_01]: I'll go lead and then just I'm curious of how you guys think about that because there's an infrastructure part of this for sure, which is a queries part like the quiz point. [SPEAKER_01]: There's then an application of workload like an impact what's going on there and there's a different sort of users and consumers across that. [SPEAKER_01]: How do you bring that down and look at that and think about where we're going to make different decisions or build relationships in there to accelerate that?
[SPEAKER_02]: So I think Cory brought up a great point and it enables us to at least build that foundational solution that we can then build the offsuits off of rather than building a dozen bespoke solutions from the ground up. [SPEAKER_02]: So this is I think it's really a model that allows us to scale rapidly. [SPEAKER_02]: We've got a foundational solution and then we can customize to the specific mission. [SPEAKER_02]: is it, are we adapting it to a maritime environment and, you know, an at-see mission versus troops on the ground in a purely land-based scenario and how is that adapting to aviation and how do we bring all those together?
[SPEAKER_02]: So I think that's critical, but that foundational approach and the approach that this is taking is what's enabling us to do this at scale and do it affordably. [SPEAKER_00]: I think to add to this point, I think about this a lot, what has happened is what cloud enabled was the app developers or business owners to be able to use infrastructure anywhere around the world. [SPEAKER_00]: Its consistency, their security, their availability, so they could use it across.
[SPEAKER_00]: So, what we have been focused on is, hey, we want to extend the same experience to the address well, hey, be it is in corners, or corners, you have the same experience, you know, the mission owner should not think about the infrastructure, should not have to worry about that. [SPEAKER_00]: Rather, at a single click automated deployments is what we are looking at. [SPEAKER_00]: The other big thing I see is even before we start at a lot of, I think about SaaS providers or applications who are using the latest AI tech to actually develop and help the mission owners, how do they deploy?
[SPEAKER_00]: You don't want all of these folks, I think it's about building on each everybody's strength. [SPEAKER_00]: What we look at is that we want to provide that secure, ruggedized infrastructure, that not just the mission owners use, but we also enable a marketplace that we can enable all of these newer, in a generation AI companies to be available for those mission owners. [SPEAKER_00]: That's something which we have been focused on. [SPEAKER_00]: I mean, we had done similar in Azure Marketplace, just a lot of that for, like in the cloud, we want to extend the same thing here, and hey, they don't need to worry about the infrastructure.
[SPEAKER_00]: That's what we've been focused on. [SPEAKER_01]: You know, Corrin, talking about sort of the diversity of sort of users and consumers, I think one of the more complex the reason I'm not jealous of your job in the least bit, right? [SPEAKER_01]: Is, I think the very, we'll call it the varied understanding of like consumption, how to contract and sort of buy software in the department. [SPEAKER_01]: how to think about integration and deployment.
[SPEAKER_01]: So when you hear and you're supporting, hey, we're able to mirror capability, bring it out to the edge, think about sort of extensibility and tactical use cases. [SPEAKER_01]: How are you guys thinking about maybe removing some of the barriers where, hey, in order for me at the tactical edge to go make a bite and get a system like this, I've got to go all the way up through this other crazy process [SPEAKER_01]: Maybe talk to folks who don't have the proximity to the mission or the understanding.
[SPEAKER_01]: How do you make that more accessible to those who need it most? [SPEAKER_03]: I think the key to bringing a lot of those disparate systems together is deploying and building and operating those systems at the enterprise scale. [SPEAKER_03]: What we run into typically is and I know everybody's favorite topic is the risk management process, the authority to operate process. [SPEAKER_03]: I haven't heard of it.
[SPEAKER_03]: Yeah, everybody has had their experiences and all of that. [SPEAKER_03]: And I know that we continually try to make that process more streamlined, try to make that process cleaner as we move from technology to technology. [SPEAKER_03]: And I think the cloud environment in the environment as the whole has really opened the department's eyes to the concept of sort of that marketplace idea, right? [SPEAKER_03]: Where, you know, I'm, I'm approving a framework, you know, I'm approving an enterprise capability and then as you deliver capabilities and solutions, those can get audited separately and be able to then be approved to move forward from the enterprise.
[SPEAKER_03]: What it saves is every tactical organization having to submit their own [SPEAKER_03]: authorizations, right, run their own accreditation, their own capabilities, a lot of what, you know, we get from the discipline perspective is organizations coming to us saying, I don't want to run it. [SPEAKER_03]: I want you to deliver the service to me. [SPEAKER_03]: And so I think by kind of enterprising a lot of this and then enterprising again to kind of the earlier point I made about that separation between sort of the tactical mission set and the larger enterprise use case [SPEAKER_03]: deliver an enterprise capability that mission users can then adapt to their sort of mission set, right?
[SPEAKER_03]: In our view, they know their mission better than we do. [SPEAKER_03]: We're not there to dictate how they should do their mission, what tools and capabilities they should use to do their mission. [SPEAKER_03]: What we want to do is walk in the door with, hey, we have a capability for you, or we have a set of capabilities that's approved, that we know is secure, that we can move your mission forward with. [SPEAKER_03]: And then if you have something that, hey, I actually have to do it this way, then we want to be there to kind of rapidly assess, rapidly give the thumbs up for that kind of design paradigm, [SPEAKER_01]: You talk about sort of the marketplace and the community and sort of democratizing access to a bunch of capability.
[SPEAKER_01]: You know, Microsoft, you get clearly have been doing this for decades, bringing capability, looking at some of the delivery mechanisms you guys have in place. [SPEAKER_01]: Where do you guys see opportunity? [SPEAKER_01]: As we talk about, you know, we've got countless ISVs in here, we've got a bunch of tremendous apps getting demoed in the back there. [SPEAKER_01]: You know, when we think about, hey, how do we get all of this software and power it for good and really bring it to those?
[SPEAKER_01]: Where do you see opportunity there? [SPEAKER_02]: Well, I think that's where the partnership with the second front really comes in. [SPEAKER_02]: I think there are a myriad of ISVs out there. [SPEAKER_02]: Customers, I think, have a need, not always a familiarity with the companies are there, but even the ones that they know need an easy path, you know, through a marketplace, to get on top of the infrastructure.
[SPEAKER_02]: In this case, [SPEAKER_02]: We want them on the infrastructure, on the cloud, here in the US, and then we want that same experience pushed forward for the operator. [SPEAKER_02]: At the end of the day, if we can help our customers use the technology and understand it here in an easy environment, and we have to go to the hard environment to go fight the next war, we've got to make sure that they have [SPEAKER_02]: the same experience or almost exactly the same experience and are applying the power that's awkward.
[SPEAKER_01]: When you think about, you know, not just the, hey, I want parody of experience, you start to think about, you know, the broader futures a little bit more than Ramada specific question. [SPEAKER_01]: But you're seeing increasingly sort of like a re-balkingization of a lot of infrastructure where things are sort of getting vertically integrated. [SPEAKER_01]: there are even at a corporate level you're seeing way more data provenance and sovereignty and you're seeing just a bunch of changes I think happen.
[SPEAKER_01]: What does that mean for you know if I'm an ISV out here and I say hey I want to be a guy in tomorrow right. [SPEAKER_01]: How are you thinking about sort of the evolution of what that experience is and what that edge is and what that point of consumption is? [SPEAKER_00]: So, while I was at Microsoft, it took us a long time to go through IL-5 and IL-6, right? [SPEAKER_00]: You can't expect every ISV to go through that long journey.
[SPEAKER_00]: So, we talk about the concept of a platform, right? [SPEAKER_00]: This is where the, you know, like just going back, this partnership actually is creating a platform, like for example, we bring in the realized, you know, [SPEAKER_00]: Containers, orchestration, Azure, with Azure local, you get hardware and with second-front you get actually a DevSecOps platform, right? [SPEAKER_00]: So what we want to enable for the ISVs and the App Builders is, hey, abstract lot of these things, right?
[SPEAKER_00]: You need to be focused on what you are good at. [SPEAKER_00]: It's primarily developing the application and being able to deploy easily. [SPEAKER_00]: And we take care of all the, you know, platform responsibilities. [SPEAKER_00]: And that's, that's where we look at that as the vision where, hey, we want to be the single platform, any app provider, any ISVs can come and deploy it and say and cater to the market.
[SPEAKER_00]: Because like when I look at commercial and the go market, the biggest thing is, like when I look at, I mean, I, I live in the Silicon Valley so many startups there. [SPEAKER_00]: Like, [SPEAKER_00]: Every day they are creating new things, every time you talk to them about, you know, bringing to the go, they said, oh, it's a long path, right, how do you, so what we wanted to create that platform where they can go from commercial to government really fast, right, that's really how we think about it.
[SPEAKER_01]: And then you hear us on the, and I recognize Dyssa doesn't own, you know, the broader accreditation for everything. [SPEAKER_01]: So, but it's you look and think about, hey, the challenge of thinking about, like, Kona's an O-Kona sort of accreditation cycles, thinking about, sort of, either nested or inherited sort of compliance and making sure that all sits. [SPEAKER_01]: Do you see either opportunity for sort of policy changes or moving forward or have you guys found an approach on how you think about that?
[SPEAKER_01]: Because that complexity gets really interesting as you start to compound there. [SPEAKER_03]: Yeah, and you know. [SPEAKER_03]: personal opinion on a lot of this, but you know, we, the arm of process as it sits day does have some mechanisms, you know, it's got, it's got changed in heritons and other things. [SPEAKER_03]: But what we find is that the prerequisites to that, the package is being set up a certain way and all that, they're not there.
[SPEAKER_03]: So when you try to build those, they're not clean. [SPEAKER_03]: You know, I think, [SPEAKER_03]: where we are with the, I know that the process is long for doing things like the professional authorizations and things of that nature, but I will say that that really was our kind of first foray into doing a sort of large scale, you know, marketplace style authorization, where we say hey, the platform is good, you bring us services, okay, that service is good, that service is good, that service is good, and we can kind of then.
[SPEAKER_03]: make it easier, it's not necessarily easier in the start-up, but what it is easier is now when somebody comes forward and says, I want to use capability A. [SPEAKER_03]: Well, it's already provisionally authorized. [SPEAKER_03]: All you do is write in your dialysis and move forward. [SPEAKER_03]: But is there opportunity to change 100%.
[SPEAKER_03]: I mean, this is where, [SPEAKER_03]: I think we have some faults is the relationship that is built when we bring an organization into the government to deploy a capability, and that relationship very much is, okay, you stand there, I stand here, you give me your artifacts, I review them, I tell you what I don't like, you make adjustments, you send me revisions, I look at it, I say go ahead and move from there. [SPEAKER_03]: And then you go to the next organization and they say, please send me your artifacts and we'll do this thing, right?
[SPEAKER_03]: Ignolage that is an absolute mind-bender for a lot of folks. [SPEAKER_03]: And it causes a lot of confusion. [SPEAKER_03]: And I've been in this world, you know, where I've lived through that confusion. [SPEAKER_03]: So, [SPEAKER_03]: I think there's two fold.
[SPEAKER_03]: One fold is the technology around the policy. [SPEAKER_03]: And these are things where if we can not have people submitting three versions of the same document or not having to revise and rev four, rev five in an email chain, right? [SPEAKER_03]: But actually having a system where we can put these things and easily adjust them, adjudicate them and look at them. [SPEAKER_03]: And then, you know, from the policy perspective, I really think it's more about getting involved early in sort of the development and release of a capability, right?
[SPEAKER_03]: And treating it not so much as that sure side of the room, that's my side of the room. [SPEAKER_03]: You pass your documents through some, you know. [SPEAKER_03]: hidden process and I will look at them behind the curtain and come back with a response, right? [SPEAKER_03]: It really is a partnership.
[SPEAKER_03]: When we use the use of platform and software as the service capabilities have really taught us that this is no longer a my app, your app world. [SPEAKER_03]: This is very much a we use these capabilities to further our mission we use these capabilities to to go forward and fight and and you know and secure our information and it creates a shared responsibility model between us and the platforms that we use to share not only understanding of how it's built but pay if something happens you need to you know we need to work together to resolve it we need to understand what happened where our data lives how it's being protected and that is not a well that's your job that's my job.
[SPEAKER_03]: It's got to be an awesome job. [SPEAKER_01]: I'll ask this to the group, take it in whatever sort of order. [SPEAKER_01]: You know, I think there's the pace of some of the pace at which technology is sort of evolving. [SPEAKER_01]: Sort of the speed with which you're seeing new trends, new capabilities come out.
[SPEAKER_01]: I think there's usually sort of an education gap or an awareness gap, especially for folks that maybe are living in a mission and not, not focused on hey, what's the next thing? [SPEAKER_01]: Let me pull a documentation or either their non-technical. [SPEAKER_01]: Where do we see either barriers that need to be overcome or some opportunity to, I think, help educate, communicate, and get folks more comfortable with some of the capabilities that are coming down the pike.
[SPEAKER_02]: I'll take that to start. [SPEAKER_02]: I think that interestingly, the folks in the mission, the closest to the battlefield. [SPEAKER_02]: I find or the ones who tend to be most educated and have a dot desire for using the latest greatest technologies. [SPEAKER_02]: I mean, if you look at the drive for agentic AI solutions amongst the operators at the edge, leading the mission, that's where the knowledge seems to live and our challenge then is how do we take that mission owner at the edge who [SPEAKER_02]: either knows what they want or thinks they know what they want and they've got some pretty good ideas and then how do we connect the dots with DISA and with the providers like Microsoft, and Armada and second front to shorten that chain.
[SPEAKER_02]: And so that I think that's where the challenge lies. [SPEAKER_02]: We talk about the A2 process and other things. [SPEAKER_02]: We all acknowledge that we've got problems there that we have to solve to shorten that chain, but I definitely don't think it's, [SPEAKER_02]: The issue is not that the mission owners don't get it. [SPEAKER_02]: I think they really do, and we've got to figure out if the mission owners get it, how do we complete the change so everybody is thinking the same way.
[SPEAKER_03]: Yeah, I would agree and I'll tell you I think a good example of where this did start to work for us was during what we call our cloud infrastructure code project that we had done in cooperation with some of the major major CSPs and what we ended up doing was creating a [SPEAKER_03]: basically an approved reference architecture combined with a basically feedback from the security systems within those cloud providers to then feedback to the accreditors to say, hey, if you're seeing green across the board on this and you're following the approved methodology, then you're in good shape.
[SPEAKER_03]: Prior to that time, the assessors really weren't living in the world of software defined environments and basically being able to read code definitions and say, [SPEAKER_03]: Well, that's what your environment looks like. [SPEAKER_03]: And I know that because it's pipeline controlled, it's not, you know, it's in a femoral environment. [SPEAKER_03]: So I know when you spit out this line of this piece of code, or I take an inventory of your environment, I know exactly what it looks like.
[SPEAKER_03]: I don't have to, you know, send people to look over your shoulder and open check boxes and that kind of thing. [SPEAKER_03]: So it was a really good example of changing the way that the assessors look at review and ultimately authorized capability. [SPEAKER_03]: To your point about the mission owners, the mission owners don't understand range, they do know this technology because that's where it seems to start. [SPEAKER_03]: A lot of times, we push for enterprise, next generation capabilities, and a lot of it gets mired in sort of everybody's kind of got a different idea of where things should go.
[SPEAKER_03]: And so we take pilots and we kind of point them in this direction, point them in that direction. [SPEAKER_03]: The mission owners [SPEAKER_03]: They're not messing around with that. [SPEAKER_03]: I need it for this particular thing for my use case and I'm going to deploy it. [SPEAKER_03]: I'm going to use it.
[SPEAKER_03]: I'm going to move forward. [SPEAKER_03]: So they get it rapidly. [SPEAKER_03]: They're able to move with it. [SPEAKER_03]: It's when they want to then bring it back and make it, you know, divest themselves of it or, you know, not have to run it, not have to run it themselves.
[SPEAKER_03]: that it turns into, well, we're not super familiar with how to do this at scale. [SPEAKER_03]: So I think it really is just an education of us looking at things more at the tactical envelope, understanding the use case, understanding how it operated there, and then just being able to extrapolate that into the larger enterprise. [SPEAKER_00]: I think to add to both of their points, mission owners love seeing new tech, they know it, and they say, I love this, okay, I want to test it.
[SPEAKER_00]: Once I tested, how do I deploy it and scale it across hundreds of these things? [SPEAKER_00]: So that's where the problem happens. [SPEAKER_00]: So you say that, oh, now that's going to go like 36 months because hey, in a year for deployment, then year for accreditation, that's going to take a long time. [SPEAKER_00]: That's where I think, you know, joining together, even these are so important to say [SPEAKER_00]: how do we lean forward and say that I like to kind of pull back some of these things when you show a tech from the mission analysis I love this, I tested it like how can I shorten the time to scale that out right in in months not like years.
[SPEAKER_01]: I mean you think about 36 months just to put like a vivid particular [SPEAKER_01]: I had done an 18 month deployment. [SPEAKER_01]: Come home, reset, recycle, and gone back and was on my tip before that piece of software would be accredited and available there. [SPEAKER_01]: I mean, that's where I think you've got. [SPEAKER_01]: You're seeing speed happen out of that tactical edge because it's the mission demands it.
[SPEAKER_01]: I think it's an interesting, it's almost sort of backwards where you'd think a hub and spoke, you know, where hammer on the hubs, domestically, cone-ass, moving those out and it's giving us a capability. [SPEAKER_01]: You're almost seeing that reverse sometimes, where, you know, that mission owner, that sort of tactical use case. [SPEAKER_01]: is manifesting in a capability, and then to the point, they come home, and they're like, live being used in, you know, beep, beep, boop, boop, why can't I use that here?
[SPEAKER_01]: How do we start to connect those two worlds? [SPEAKER_01]: Because it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's [SPEAKER_01]: There are too, there are many DODs, but there are two very distinct DODs.
[SPEAKER_03]: Yeah, this is where reciprocity becomes massively important. [SPEAKER_03]: And we see this in a lot of different areas. [SPEAKER_03]: You know, authorizing officials have different opinions on how something should be accredited to the enterprise scale. [SPEAKER_03]: We get it right in pieces, right?
[SPEAKER_03]: I mean, the provision authorization process, while, well, maybe not perfect. [SPEAKER_03]: It was a really good example of everybody agrees. [SPEAKER_03]: We're all good. [SPEAKER_03]: We're going to move forward with this.
[SPEAKER_03]: And you can then leverage that body of evidence and that information in order to give it a thumbs up. [SPEAKER_03]: But in the tactical case, what we find is, [SPEAKER_03]: Mission-specific authorizing officials are, they're very quick to look at a thing, go, yep, because they're invested. [SPEAKER_03]: It's their mission. [SPEAKER_03]: You know, we're all invested in mission-going forward, we're all invested in protecting the warfighter, but it's a level of proximity, right?
[SPEAKER_03]: They know the mission, they know the mission risk, they know the risk to the information, so it's very easy for them to go, yep, nope, I'm, I'm go with what you're doing here. [SPEAKER_03]: But the next leg of that is they go to a whole new authorizing official that has no idea what [SPEAKER_03]: Well, I don't know how you're going to protect this, how you're going to protect that. [SPEAKER_03]: Well, this is where a reciprocity should come in, and that mission A.
O. [SPEAKER_03]: can say, I've reviewed it. [SPEAKER_03]: Here's the, you know, litmus test I put it through in order to, to give it the thumbs up. [SPEAKER_03]: And then, all that receiving A.
O. [SPEAKER_03]: should have to do is go, I agree with all of that. [SPEAKER_03]: Hey, there's one concern that I need to mitigate here. [SPEAKER_03]: And once you mitigate that, I'll give it the thumbs up and now we've got it across the board.
[SPEAKER_03]: Versus having to start over at phase one to every organization you go to. [SPEAKER_01]: Yeah, I was like, let's see it nod and go ahead and jump in on that. [SPEAKER_01]: I think it's a problem we all live is, you know, but I've lived in government out of government as we're trying to build as you're getting adoption. [SPEAKER_01]: You talk about it.
[SPEAKER_01]: Hey, I found user adoption. [SPEAKER_01]: I found product mission fit. [SPEAKER_01]: This is great. [SPEAKER_01]: And then all of a sudden you realize that wait a second, like this isn't the path that's going to get me to the scale there.
[SPEAKER_01]: So how do you see that? [SPEAKER_02]: I think the operators at the edge are going to force our hands in getting it right. [SPEAKER_02]: Because what we're seeing is the operators at the edge aren't waiting for the accreditation process. [SPEAKER_02]: They'll go procure it commercially and they're using applications on Azure commercially to prosecute the mission.
[SPEAKER_02]: and which is a great thing, but it highlights that we've got to come together to solve this problem so that we can get a secure accredited solution in their hands fast enough that we don't incur the risk of an unvetted commercial solution which may have some cyber vulnerabilities ultimately impacting a mission. [SPEAKER_02]: But I think [SPEAKER_02]: you know anybody who's been out there at the edge like yeah I don't care I'm not waiting I'm getting it done and I could host to the folks who are getting it done we've got to figure out how to how to close that gap really rapidly because if we find ourselves in the next conflict in the Pacific [SPEAKER_00]: Nobody is waiting.
[SPEAKER_00]: Yeah. [SPEAKER_00]: I think I think to tell the point is, you know, before 36 months was long, now even in 18 months the, you know, the pace at the AI is, you know, making changes, 18 months at the fees are long. [SPEAKER_00]: The platform and the app could completely change. [SPEAKER_00]: So almost I think about the, for example, an app platform.
[SPEAKER_00]: like we are releasing so many changes on a daily basis, right? [SPEAKER_00]: We are looking at the AI use cases, how do you make it? [SPEAKER_00]: I almost think about like I want to use the word of agentic AI accreditation, right? [SPEAKER_00]: You can't wait for you know, accredited novice that to look from one to the other.
[SPEAKER_00]: There's going to be changes happening there. [SPEAKER_00]: There has to be an agentic AI to make sure that OK, we are not missing the mark. [SPEAKER_00]: It is still staying within the same security baseline. [SPEAKER_00]: We don't have time because our enemies are going to be doing this at a faster pace.
[SPEAKER_00]: So how do we enable the latest really fast without going through the long documentation and bureaucratic process? [SPEAKER_01]: I think there's definitely a tension which I get between sort of speed and mission and security. [SPEAKER_01]: I think we've got to recognize that like, you know, regulated industries are regulated for a reason, right, there's carried risk. [SPEAKER_01]: There's a bunch, there are implications to it going wrong and people own sort of consequences of those.
[SPEAKER_01]: And I think more often than not, right, the AOs are just getting kicked from all direction. [SPEAKER_01]: right like you're either going too fast and like it's not secure you're not going fast enough and you don't care about it. [SPEAKER_01]: It's like the least envisage job in the world. [SPEAKER_01]: How do we think about scaling or enabling them because to your point for you?
[SPEAKER_01]: We can't have this be, hey I've uploaded like version 72 to SharePoint, you know please pull down and review what your leisure and sign the PDF. [SPEAKER_01]: But we also can't say, hey, just trust it's going to be fine. [SPEAKER_01]: So there's somewhere in between that we've got to get to. [SPEAKER_01]: So how do we enable that broader, sort of, like GRC community?
[SPEAKER_01]: That AOism sort of Scott community to help them scale. [SPEAKER_01]: And maybe, of course, yeah, the lay up for everybody. [SPEAKER_03]: Hey, that's the question. [SPEAKER_03]: I think that's the million dollar question.
[SPEAKER_03]: I think that's the one that is where we're living, right? [SPEAKER_03]: And I think it starts with, [SPEAKER_03]: removing, he'd say, removing the human from the process, right? [SPEAKER_03]: Making these artifacts, machine readable, machine accessible, right? [SPEAKER_03]: And then being able to, I'll say this way.
[SPEAKER_03]: We talk a lot about what AI can do and the things that that technology can do to help enhance capability. [SPEAKER_03]: And it's not just, you know, AI taking over the function sets, right? [SPEAKER_03]: or I'm mixing metaphors now, but having AI, you know, take the information, collect it, chew on it, look at it, and point out discrepancies and things that now the human can contextualize and say, yeah, that's a problem or, no, no, no, I get that looks bad, but, you know, I'll sign off of it from there.
[SPEAKER_03]: So, I think that's the, you know, everybody, everybody looks at their own business processes and their own process and they go, [SPEAKER_03]: Well, AI can fix all of this. [SPEAKER_03]: It's like, well, it can, and it can definitely enhance it. [SPEAKER_03]: But you've got to do it. [SPEAKER_03]: You have to have the prerequisites to make it work, right?
[SPEAKER_03]: Bad data in is bad data out. [SPEAKER_03]: And so having an ability to take these artifacts that we use as body of evidence to show how, you know, how securely a system is built, making the machine auditable, machine accessible, and providing a pathway for organizations both, [SPEAKER_03]: government organizations and, you know, providers to modify that data at their speed, right, because, you know, industry moves quickly and that the current HEO process is not built for that level of rapid change.
[SPEAKER_03]: It just isn't, right? [SPEAKER_03]: And so, like any good process, it will mature, but I think it starts with having those prerequisites available, being able to get fast data in, [SPEAKER_03]: And then using AI tools or others to help contextualize the data and present it to the AOS. [SPEAKER_03]: So the AOS not looking at a diagram. [SPEAKER_03]: They're looking at an overall system assessment says moderate concern, moderate concern, major concern, minor concern.
[SPEAKER_03]: Now you contextualize it and tell me where we've got problems. [SPEAKER_01]: I want to, we got about five minutes left and I'm going to try to have 90 seconds, maybe each of you, doing the math right with a little buffer in there. [SPEAKER_01]: What are you most optimistic about? [SPEAKER_01]: As we look at, and I'll bound it on, like, let's call it a four year horizon, right?
[SPEAKER_01]: Let's not line it up to administrations. [SPEAKER_01]: Let's not say current year, but let's not say, like, 30 years down the road, right? [SPEAKER_01]: What are you most excited about or optimistic about, and why you'll see there's a trend I try to, I just want to end on a good note. [SPEAKER_01]: We always talk about things that aren't working.
[SPEAKER_01]: Let's talk about what could be working. [SPEAKER_03]: So, it's a very low bar for me, but what I'm excited about is right now, [SPEAKER_03]: In the Oconus world, people are saying, I can do cloud. [SPEAKER_03]: For the longest time it was, I can't touch cloud. [SPEAKER_03]: It's back, Coneus.
[SPEAKER_03]: I can't deal with that. [SPEAKER_03]: My mission can't survive that. [SPEAKER_03]: So realistically, the fact that we are now open to cloud on the Oconus side of the house, it provides such massive capability to deliver enterprise grade capabilities and distributed capabilities much faster than we would have been able to originally. [SPEAKER_02]: I think Ukraine was a watershed event in the world of technology and the application and technology on the battlefield.
[SPEAKER_02]: And we watched how drones were used and some of the counter-drone tactics and how electronic warfare was used on both sides of the conflict. [SPEAKER_02]: And what I find really encouraging is if you think about this, a… [SPEAKER_02]: military that did not have the resource the United States that had to be creative in the way they prosecuted a large military. [SPEAKER_02]: They've done this in such a way that we've taken away thousands of lessons learned from that.
[SPEAKER_02]: And I'm really optimistic that we can apply those lessons learned in the United States military and with our allies to be incredibly effective on the battlefield to integrate AI into integrate all the software capability at the edge and do it in a way that [SPEAKER_02]: The Ukrainians weren't able to do, so look at what they did and then multiply that by 10,000. [SPEAKER_02]: That's what our future holds and I'm really excited about what we can do there. [SPEAKER_00]: Yeah, I think for me, if I look at the journey, there's an on-prem, and when we build the cloud, there's a huge, you know, multi-air process to even, you know, convince the accreditation, the owner, etc.
[SPEAKER_00]: to move stuff to the cloud, right? [SPEAKER_00]: But what cloud as enabled is now, you can do like, you know, huge compute processing. [SPEAKER_00]: You can, you, you, you can play around with tons of data. [SPEAKER_00]: Now, what I see is, hey, now people have started to think, because I've done the cloud transformation, it's easy for me to take that and run it on the edge.
[SPEAKER_00]: Like if I had to go from an on-prem to edge, it was like a very hard problem, right? [SPEAKER_00]: So that transition is help. [SPEAKER_00]: And now what is happening is with the AI stuff where is it? [SPEAKER_00]: You don't need to think about having all the data generated, push to the cloud and then start training.
[SPEAKER_00]: You can actually also do a lot of training like with the help of AI. [SPEAKER_00]: And you don't need people for all of these operations either, right? [SPEAKER_00]: So that's where I'm super optimistic about hey, I went to the cloud, cloud is going to exist. [SPEAKER_00]: It's going to grow.
[SPEAKER_00]: But hey, how do I take all the benefits of the cloud and then apply it or the ads and make it easier and easier for people on the ground? [SPEAKER_01]: Yeah. [SPEAKER_01]: I mean, I like the sense of it there, pretty. [SPEAKER_01]: Like my optimism, especially, I think magnified during events like this is you think about the diversity of folks we have in the audience on allocators, on builders, on operators, users, [SPEAKER_01]: I think I was really worried a year ago that the fat of defense tech is the new crypto and it's maybe got some staying power, right?
[SPEAKER_01]: And so there's some goodness there, but that only works if you enable consumption to happen and you enable growth to happen, right? [SPEAKER_01]: So I could pour a trillion for gun dollars. [SPEAKER_01]: But if every cool capability that gets built gets built, gets into RDT&E, gets into some short room and the deployment process and can't get out of it. [SPEAKER_01]: At some point, rightfully so, the board or the capital allocators are going to take their ball and go home.
[SPEAKER_01]: And the great of capability that happens there is almost undefined. [SPEAKER_01]: So when we look at folks that are building new extensible secure resilient delivery models that modern software can adapt to and it's at the speed of mission, and it fires me up. [SPEAKER_01]: So I want to thank you guys for joining me here. [SPEAKER_01]: I want to thank everybody.
[SPEAKER_01]: I know we're sitting between you and lunch. [SPEAKER_01]: So Sam knows what to do next. [SPEAKER_01]: They're waving at me, but I don't know what they're saying.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.