The Scramble to Transform: From BPO to DefenseTech
- stephentloynd
- 5 hours ago
- 10 min read
Contractors that are willing to change with us will prosper and grow.
Those who don’t, and resist it, will be gone.
―Stephen Feinberg, U.S. Deputy Secretary of Defense
Hear that, BPOs? Transform or disappear.
I recently attended Govini's 4th Annual 2026 Defense Software & Data Summit as well as McAleese’s 17th Annual Defense Programs Conference here in Arlington, Virginia in order to cross-fertilize insights from the increasingly prominent defense sector with what’s happening in the business process outsourcing (BPO) industry.
As I wrote recently in “War Room: The Complex Realities of AI & the Contact Center,” transformation in contact center BPO “is not only determining the future of customer contact… but helping us understand the future of an entire economy.” Here in Washington, D.C., the defense sector is the supreme example. BPO taught us how to manage distributed work, and now the latest wave of defensetech startups is showing us what happens when a somewhat similar model is attempted in national security.
In essence, I’d propose that BPO is becoming more like defense (complex, high-stakes, interdependent, AI-driven systems) as defense reinvents itself to look more like BPO (distributed, ecosystem-based delivery). Unlike BPO, however, defense doesn’t have decades to figure out how to manage complexity through standardization and globalization. And it can’t afford to fail.
Operational Coordination
As I wrote in The Widening Turn, process transformation is accelerating and spreading throughout the American economy. And one could argue that both BPO and defense are operational stories swallowed up by tech hype. In that sense, both realms don’t face a technology challenge so much as integration and coordination challenges within operations.
First, a level set. In the case of both BPO and defense, the client defines mission needs (for example, a big bank or telecom firm in the BPO space and the Department of Defense in the defense sector… a monopsony market with only one customer). The prime contractors (for example, Lockheed Martin or Northrup Grumman and TP or Concentrix) are accountable for delivering on the contract and orchestrating any subcontractors, specialized firms, and tech startups. There’s also a technology platform layer (from AWS or Palantir in the defense sector to Genesys or Avaya in the BPO space). That in turn, is tightly coupled with a data and AI layer where insights enable decisions. Talented people are the distributed human layer that drive the machine.
Despite reforms in defense acquisition, persistent challenges remain—prolonged development cycles, escalating costs, and structural inefficiencies. As such, those challenges “require new iterative approaches to delivering capability with speed.” For the U.S. Defense Department, ecosystem orchestration is the desired endgame—essentially outsourcing at scale. As Jim Taiclet, chairman, president and CEO of Lockheed Martin, put it, “To accomplish this, we need not just the defense industry but the full participation of commercial technology companies, telecommunications companies, experts in artificial intelligence and other advanced technologies—the entire industrial fabric of the U.S.”
In other words, in 2026, modern defense capabilities are assembling across a network that behaves a lot like an outsourced operating model featuring specialists, startups, and software platforms. It’s a system that aims to behave like a global services model, only under a particular set of constraints. So how will complex systems get implemented and understood? As the BPO “War Room” demonstrates, complexities can complicate the functioning of the machine.
Complexities & Inherent Tensions
BPO is all about things like response time and resolution rate, while defense is all about things like system uptime in combat, the speed of intelligence flow, and the accuracy of AI decisions. While Quality Assurance (QA) in BPO means customer satisfaction, in defense, QA means mission success or failure. Whereas BPO accesses global talent pools, defense confronts the realities of geopolitics and localizes or “friend shores” work (consider the U.S.’s new arrangement with the Philippines “to accelerate U.S. manufacturing in defense and other key industries while lessening China’s chokehold on critical minerals and other key components for electronics”).
Also consider that while BPO is about things like access to talent, cost efficiency, and scalability, national security is about things like trust, control, and resilience. Fundamental concepts are similar, but the stakes are much higher in defense. As the U.S. government turns more toward the private sector, Silicon Valley, and a more complex outsourcing model, the logic of outsourcing and national security don’t necessarily align. And as change in the defense sector accelerates, that inherent tension will intensify. After all, associated risks in the BPO space are things like service quality or cost overruns, but associated risks in defense concern national vulnerability.
Moreover, commercial BPO features roles that are relatively clear. It also features mature governance models that operate on a layered, designed service delivery stack—an interdependent network of people, platforms, and processes. But the defense sector stack faces more fragmentation, less visibility, blurred roles, and higher stakes as change hits the system and governance scrambles to keep up. Vendor proliferation in the form of more specialized AI providers produces more integration points, increasing system complexity.
As I pointed out a year ago in “Our Multi-nodal Hall of Mirrors,” we’re living in a kaleidoscopic multi-nodal world order, “an increasingly networked world” that “makes partnerships and alliances more essential than ever.” In such a world, one should expect issues related to integration, performance, and trust to manifest.
Indeed, while McKinsey is right to note that by “disaggregating capabilities into networks of smaller nodes, force planners can reduce points of failure and increase the likelihood of successful missions connecting air, land, sea, and space assets,” new complexities and attendant complications will also arise. Disaggregated capabilities will require effective and resilient communications networks that enable smooth, responsive, real-time operations that “enable the movement of responsive decision-making to the tactical edge” (ie, network-enabling technologies such as 5G, phased-array antennas, AI, and high-density computing). It’s a big ask. Will integration in defense become the bottleneck it has sometimes been in BPO?
In fact, today the U.S. appears to lack the coordinated ecosystem that Russia in particular now appears to feature through a pragmatic approach to military innovation. As Kateryna Bondar of the Center for Strategic and International Studies explains, “Russia avoids abstraction. Rather than pursuing all-encompassing architectures like the U.S. military’s combined joint all-domain command and control concept—intended to connect branches of the military on one network, but still largely unrealized after years of development—Russia builds software that solves immediate battlefield problems.” Alas—
In drone warfare, Russia’s military innovation is increasingly distributed and adaptive, while the United States remains constrained by centralized requirements, slow acquisition and limited integration. The country long assumed to be the bureaucratic giant has become a hub of entrepreneurial vigor.
Washington is investing billions of dollars in drones and AI. But to be effective, new technologies have to be integrated into military units, connected through software, and continuously adapted through training and doctrine. America needs to spend less time focusing on how to buy new technology and more time thinking about how to fight with it.
Enter Private Equity: Reorg Nation
One wonders what U.S. Deputy Defense Secretary Stephen Feinberg thinks of that critique. Feinberg has gone from leading Cerberus Capital Management to flying a desk at the Pentagon. In 21st century America, it seems few stories are complete without introducing private equity (PE) into the narrative. He and his team of finance experts—dubbed “Deal Team 6”—oversee a portfolio of defense companies, negotiating contracts and structuring deals in unprecedented ways, “often ordering firms to shift resources toward military priorities before any contracts exist to pay for the moves.” In the words of Steve Blank, an adjunct Stanford professor of entrepreneurship and national security, “This is a Cerberus takeover. Private equity has just acquired its largest organization.”
Remarkable. Financializaton churns forward. The ongoing restructuring in defense is massive. But as The Wall Street Journal also notes, “It is too soon to tell whether this organizational shake-up—a near-term disruption—will improve programs expected to take years or decades to prove their worth.”
Now consider BPO, long ideal for the classic PE strategy of “buy, improve, combine, and sell.” Whereas defense capabilities are being assembled across a fragmented ecosystem of contractors, vendors, and platforms, in BPO the strategy creates scale by assembling fragmented providers into a coordinate system. PE roll-ups in BPO might feature challenges around integrations, tech stack fragmentation, and vendor management, whereas challenges in defense include data silos, interoperability, and ecosystem orchestration.
It seems that PE turns BPO into a managed ecosystem by bolting together service providers into something akin to a platform, while PE in defense aspires to something similar but with less room for error due to the stakes involved. But anyone who’s spent time inside a PE-backed BPO understands that what often results resembles a carefully managed ecosystem of layered processes, tech stacks, and cultures held together by contracts, governance models, and constant coordination. It’s less a resilient machine than a complex environment.
Enter AI: Continuous Change
Now consider that AI doesn’t simplify these ecosystems so much as it re-wires them, which can increase operational complexity before delivering efficiency. As the system reorganizes, new complexities emerge. In the BPO world, AI turns a labor model into a system model. AI agents handle a percentage of interactions, humans handle exceptions, data pipelines feed models, and model training and tuning is continuous, powered by the likes of Databricks or Snowflake. As a result, things change operationally and new dependencies are established. In this case, you’re no longer managing people plus process, but rather things like models, data quality, prompt logic, and integration layers.
This is an environment of continuous change. Models evolve, outputs drift, and performance shifts as operations become dynamic rather than stable. But when something goes wrong, there’s blurred accountability—what was the cause of the glitch? The model? The data? Workflow design? The human fallback? Suddenly understanding what’s happening is much more difficult than examining issues around agent performance. In fact, as companies embed AI into customer service operations, governance frameworks are already lagging, from automation to agent coaching. Enterprises are struggling to monitor and validate AI-driven interactions at scale, raising concerns about oversight, model behavior, and service quality. “As AI moves from experimentation into day‑to‑day execution across contact centers, the challenge isn’t innovation; it’s ensuring trust, security, and consistent outcomes at scale.”
In defense, similar issues are multiplied. Data becomes mission critical infrastructure such that it’s not just about automation, it’s about decision infrastructure. Companies like Palantir and Anduril are building systems where AI synthesizes data, recommends actions, and sometimes acts autonomously. There are more layers like sensors, data pipelines, models, interfaces, and operators, often owned by different vendors. Whereas in BPO, bad data means a bad customer experience, in defense, bad data produces bad decisions with potentially significant consequences.
Indeed, AI systems are only as good as the data they receive and the systems they connect to. Interoperability becomes a strategic issue. Integration becomes existential. As such, trust becomes a potential bottleneck. Do operators trust the model? Do commanders act on AI recommendations? Can outcomes be explained? It’s not just technical, it’s human plus institutional. Before AI, complexity meant managing distributed labor. After AI, complexity means managing distributed intelligence.
This all has me wondering whether we’re underestimating the operational complexities impacting both BPO and defense. Across both spaces, AI doesn’t eliminate outsourcing-style complexity—it amplifies it and redistributes it. And whereas we used to outsource work, now we’re starting to outsource judgment—and that’s a much harder system to manage. It’s as if the real challenge with AI isn’t building the models, it’s operating the system those models reside within.
New Models Emerging
In essence, today’s U.S. Department of Defense isn’t managing contractors so much as it’s trying to manage platforms, data flows, interoperability, and continuous software updates. As such, defense capabilities aren’t just engineered, they’re assembled in a way that behaves like a delivery system. That’s more akin to a global services model than a traditional procurement model. The defense sector is quietly reinventing itself as a complex, interdependent delivery ecosystem. But unlike BPO, it doesn’t have decades to figure out how to manage complexity through standardization and globalization. And it can’t afford to fail.
In the case of both BPO and defense, operating models are now less about managing people and more about managing systems that make decisions. Both spaces are increasingly relying on systems that interpret data, recommend actions, and sometimes act autonomously. New delivery models aren’t just distributing work, they’re distributing decision-making. Unlike traditional outsourcing, the system is never static. As such, the job is no longer to just deliver a service as it is to continuously stabilize a moving system.
While BPO offers lessons in managing complexity at scale, the defense industry faces another order of complexity altogether. Because it’s not so much a supply chain as it is a distributed, interdependent, multi-vendor stack. Defense isn’t so much about what’s being built as how the system works. Again, think of the contact center in BPO. As the way decisions are made shifts, where do judgment and responsibility reside?
Finally, as contact centers accelerate AI adoption, many lack the controls needed to manage new risks tied to generative and agentic systems. So too in defense. AI-related challenges like invisible failure modes are very real (think data bias, model drift, or edge case breakdowns). Moments that require interpretation or hesitation feel harder to locate amidst the blur of the ongoing intelligence revolution. How do we successfully manage distributed intelligence? Less visibility means more systemic risk.
Converging on the Logic of BPO
In a sense, the logic of BPO is hustling into defense even as the AI revolution storms the party. Each sector should observe how the other manages chaos. BPO’s distributed workforce, platform-driven delivery, and AI-enabled operations sees its reflection in the defense sector’s distributed contractor ecosystem, an increasing reliance on external vendors, and platform-based systems featuring data and AI.
And if BPO can teach defense anything, it’s that distributed systems don’t manage themselves. They require structure, governance, and constant coordination. And that may end up being the hardest issue for defense to solve. Meantime, as AI and data platforms are integrated into BPO and risks grow, defense can teach BPO to design for degradation as well as performance. In defense, redundancy is survival. Complex systems require intentional structure and end-to-end visibility and observability.
Over time, one can imagine defense adopting something akin to BPO-like governance (ie, tighter vendor orchestration, SLA-style thinking, performance metrics across ecosystems). Intriguingly, while defense doesn’t outsource directly to BPOs due to constraints such as security, data sensitivity, and offshore geopolitical risk, defense is being rebuilt as a system that looks a lot more like outsourcing.
Final Thoughts
Ultimately, one wonders what happens when national security depends on a model that behaves like outsourcing, but that’s not yet fully understood. How will that model be managed under pressure? Are we thinking enough about the delivery model behind defense innovation as we are the financing of it? Are we underestimating the operational complexities behind defense innovation?
Beyond that, as more defense startups pursue “dual-use” strategies to build scale and accelerate growth in commercial markets, seeking non-military uses for their technologies while buying the time needed to secure long-term defense contracts, one wonders how American techno-consumerism might grow ever more immersive and fantastical. What will accelerating innovation of 21st century technologies between the defense industrial base and the commercial industrial base mean for the U.S. in light of intensifying use of civil-military fusion strategies?
Still bigger questions loom for both the BPO industry and the defense sector. With a new energy crisis upon us, how will offshore outsourcing adapt to a world of rising social tensions? When is war just? As exponential flywheels fly forward and ambiguity spreads—from the possibilities of an AI induced employment crisis to an eyebrow raising manifesto from Palantir called “The Technological Republic”—what ethical constraints might test the system? How will the U.S. defense enterprise respond? What will BPO eventually look like in an increasingly automated world?
Image credit: Norbert Biedrzycki blog (https://norbertbiedrzycki.pl/en/transformational-leadership-bringing-order-to-chaos/)
Comments