Abstract: The Russo-Ukrainian war betrays an approach that overinvests in things rather than people. U.S. reliance on defense contracting solutions can be compared to the concept of deus ex machina. The preferable alternative is to invest in winning teams that put humans at the center of modernization, particularly in the field of AI adoption. The author defines this concept as the “homo post machina,” or “human behind the machine.”
As the Army prepares for the coming era of sustained great-power competition, the defense industry is already churning out countless solutions to solve supposed problems. However, the U.S. military often treats defense industry products like a “deus ex machina” from an ancient Greek theater. To solve an intractable plot situation, Greek playwrights introduced a godlike figure on stage by lowering them from the heavens with a crane; the “god” from the “machine”. The god’s arrival resolved the situation and allowed the playwright to wrap up the play conveniently. However, an ending with a deus ex machina could feel like a rushed solution. The U.S. military is falling into the same trap with many of its current defense solutions, primarily hardware but increasingly with AI-enabled software as well. Rather than doing the hard work to determine the root issue of its problems, the U.S. military simply drops convenient solutions from the proverbial offstage defense machine, promising a quick fix. It is time for the military to look to a different model: the homo post machina. Although fast-moving defense technology companies provide advanced hardware and software capabilities to the joint force, the U.S. military cannot lose sight of an equally consequential requirement: investing in the cognitive development and leadership capacity of the people who employ those systems and act on its information – the human behind the machine.
Addressing the Right Problem with the Wrong Lessons
The Army repeatedly defaults to counting missiles, drones, and rifles as proxies for readiness at the expense of the more complex and more consequential challenge of measuring how effectively its people perform, develop, and lead.
The Army’s preference for the deus ex machina over the homo post machina model is evident in the disparity between what the Army spends on hardware and what it invests in the cognitive development of its force. As a concrete illustration, a recent wartime requirement to procure additional funding to support combat operations in Iran resulted in a significant transfer of funds away from the Army’s Cadet Command. While this decision was logical given the immediate demands of conflict, it exemplifies a pattern that has persisted for decades: emphasizing short- and medium-term hardware procurement over long-term human investment. The Army repeatedly defaults to counting missiles, drones, and rifles as proxies for readiness at the expense of the more complex and more consequential challenge of measuring how effectively its people perform, develop, and lead.
Furthermore, counseling and evaluation reports do not sufficiently address this challenge. For decades, “modernizing” evaluation systems meant little more than digitizing paper forms. As AI proliferates across the force, the next iteration is already visible: the Army using AI to migrate from a digital form that a leader fills out to a digital form that AI fills out. While this creates the appearance of an AI-enabled force, it does nothing to advance the decision-making depth and cognitive tempo that modern warfare demands.
The appetite for drones and AI-enabled software seemed voracious but the appetite for investing in the human dimension of these new platforms was nearly absent.
In 2022, my assignment as a Bilateral Affairs Officer (BAO) at the U.S. Embassy in North Macedonia gave me firsthand experience of the gap between hardware capability and human performance. Russia’s invasion of Ukraine defied the near-universal expectation that Russian armored formations would rapidly overwhelm Ukrainian defenses. Our efforts at the embassy to coordinate with host nation partners and rush legacy equipment to Ukraine were noble but had a negligible impact. As we realized later, this war was not simply a contest of hardware: there was a decisive human element. The mismatch between the sophistication of Russia’s equipment and the competence of its soldiers and leaders resulted in a tactical paralysis that bogged down the Russian offensive, turning what was supposed a swift strategic victory into a years-long slog. Experiences at U.S. Combat Training Centers underscore this observation: formations that practice decentralized decision-making with empowered junior leaders excel and recover from setbacks quickly while formations that centralize control are slower to act, generate more errors, and struggle to recover from mistakes.
As an infantry battalion executive officer, I expected the Army to wrestle seriously with the human capital lessons of Ukraine alongside the hardware ones. Yet, while nearly every article hammers on the need to modernize the force, move quickly with innovation, and adopt new technologies, the institutional focus remains firmly on hardware over humans. The appetite for drones and AI-enabled software seemed voracious but the appetite for investing in the human dimension of these new platforms was nearly absent. There was not just underinvestment but an active divestment from exactly the kinds of pivotal leader development experiences that build the decentralized and adaptive judgment that Ukraine’s defenders routinely demonstrate. Over a year as an executive officer, I fought repeatedly to send soldiers to training such as ranger school only to watch funding fail to materialize or be redirected at the last minute.
Those back-to-back experiences as BAO and executive officer inspired my current professional focus: building an improved modern training system that maximizes every dollar the Army invests in its people by capturing superior performance, transmitting knowledge across the organization, and generating measurable improvement in both individuals and units. In 2024, I embarked on a journey to create a purpose-built solution for the problem I had spent two decades watching go unaddressed.
Towards a Modern Training and Mentorship System
This problem is best exemplified in the shortfalls of the Training and Evaluation Outline (T&EO) system. For decades, the Army’s standard for collective training events has been the T&EO. T&EOs define the conditions under which a unit will be evaluated and the performance measures that evaluators assess. In principle, they are a sound mechanism for ensuring units meet universal standards. In practice, they carry structural limitations that significantly constrain their value.
The first limitation is mechanical: T&EOs are paper-based instruments that typically end up balled up in a sweaty cargo pocket. Even if a T&EO gets digitally loaded into a database, there is no way to aggregate the data across multiple T&EOs in an analytically meaningful way. In more than two decades of service, I have never seen a unit execute that process with any consistency or rigor. The knowledge embedded in completed T&EOs is lost.
The second limitation is analytical: T&EOs operate on a binary go/no-go scoring system. That binary verdict provides no indication of whether a “go” was marginal or decisive and no insight into whether unit performance is improving or degrading across successive training events. Most critically, the system cannot distinguish between an outcome driven by exceptional individual leadership versus genuine collective proficiency. Without that distinction, commanders cannot accurately target their developmental investments on their teams versus their individual leaders.
Over the past year, I have worked alongside talented students from Northeastern University’s Roux Institute to build a digital training system that integrates digital simulations and real-world evaluations into a single, AI-informed analytics engine. This system is designed to enable organizational leaders to rapidly assess performance, capture lessons learned, and apply findings directly to improve human performance. We recently applied elements of this technology in a Squad Situational Training Exercise (STX) for a deploying National Guard unit. We first reorganized the standard T&EO performance measures into three distinct scoring categories: leader performance points, unit collective performance points, and general tactical performance points. We replaced the binary go/no-go scale with a three-tier scoring system familiar to Army evaluators: T (Trained), P (Practiced), and U (Untrained), assigning point values of five, three, and one respectively to each tier to uncover more nuance in performance than a simple GO / NO GO system.
Working alongside a talented AI engineer, we built a mobile application for evaluators and a web-based dashboard for command teams. The mobile app gave evaluators an intuitive interface to record scores and narrative notes in real time. Those inputs flowed automatically into a web dashboard where company and battalion commanders could view developing performance trends across the formation in real time, not hours or days later. Over a week-long STX conducted across three maneuver companies, evaluators used the mobile application to assess squad leaders executing missions including ambush, attack, movement to contact, and reconnaissance. Each squad leader received between two and four complete evaluations, with twenty to forty discrete scored data points per evaluation. Evaluators could also append verbal narrative notes to flag specific areas of outstanding or deficient performance. The resulting dataset was richer, more structured, and more immediately actionable than the traditional T&EO process.
That insight enabled an immediate, targeted response. At the synchronization meeting that evening, battalion and company commanders retooled the training emphasis for the following day. The following day, Charlie company applied those lessons before their squads ever stepped off on a mission, learning from the documented performance of their peers and beginning at a higher baseline.
By the end of the first training day, clear performance trends were visible across alpha and bravo Companies without a single staff officer having to collect or manually analyze paper forms. When the last lane concluded, the battalion command team reviewed the performance profile of every squad leader across both companies in a single session. The trend was clear: troop-leading procedures and terrain model construction were the common weak points across the battalion. That insight enabled an immediate, targeted response. At the synchronization meeting that evening, battalion and company commanders retooled the training emphasis for the following day. The following day, Charlie company applied those lessons before their squads ever stepped off on a mission, learning from the documented performance of their peers and beginning at a higher baseline. This knowledge capture process generated a genuine developmental advantage for Charlie company.
The near-term development roadmap for this system shows promise for decision analytics and leader development. The next step is integrating a system that combines evaluator observations and reveals patterns relevant to tactics and mentorship. Such a system will aggregate information across evaluators and identify common characteristics of patterns that would require hours of manual reading to detect. Whether a particular leader consistently fails to establish a terrain model, hesitates on the objective, or struggles with a specific planning step, the chain of command can observe patterns of behavior as actionable intelligence rather than anecdotal impression.
The end state extends well beyond individual training events. The same data that surfaces performance trends in a Squad STX can drive Sergeant’s Time Training content, shape individual development plans, inform professional development counseling, and ground evaluation reports in documented, verifiable evidence rather than the subjective recollections of a supervisor. This is the transformative potential of AI applied to human performance: not the automated generation of generic evaluation bullets, but the surfacing of previously hidden trends that enable meaningful mentorship, accurate assessment, and targeted development.
Knowledge Capture as Competitive Warfighting Advantage
Two concepts underpin the strategic value of this architecture. The first is the competitive advantage gained by knowledge capture. High-performing organizations in competitive domains (e.g., professional sports franchises, special operations units, elite corporate teams) treat performance data as proprietary intellectual capital: captured, retained, and applied to close capability gaps that hardware and resources alone cannot bridge. In this framework, knowledge capture is not a support function but a warfighting function.
Identifying performance trends, surfacing insights, and cataloguing developmental recommendations in accessible databases removes that friction. The goal of this system is not to automate mentorship, it is to make mentorship frictionless enough that it occurs at the frequency and quality the force actually requires.
The second concept is best described as “frictionless mentorship,” an analogy to frictionless payments. Removing friction does not simply make a transaction easier, it fundamentally changes the volume, frequency, and quality of future transactions. Traditional mentorship is high-friction, demanding that a supervisor recall subjective impressions, carve deliberate time from a compressed schedule, and construct developmental guidance largely from memory. Identifying performance trends, surfacing insights, and cataloguing developmental recommendations in accessible databases removes that friction. The goal of this system is not to automate mentorship, it is to make mentorship frictionless enough that it occurs at the frequency and quality the force actually requires.
The Future: Augmentation over Automation
Rather, the future of artificial intelligence in the U.S. military must follow a different track: the augmentation of human intelligence rather than automating it away. It must transform the Army’s ability to capture knowledge and immediately repurpose it to train its formations and the next generation of leaders.
The Army is making genuine efforts to modernize and prepare for the next generation of warfare. However, we often assume that defense contractors will simply provide “godlike” solutions. Although the defense industry is happy to supply such solutions, this method is unlikely to produce results that hold up against the friction of war. Furthermore, although recent unconventional approaches such as the direct commissioning of technology executives into military service are a step in the right direction, such solutions are tantamount to switching one machine for another. Tech executives do not know what it means to lead soldiers in complex environments. Without this context, their instinct will remain to optimize systems and workflows rather than address the deeper human performance problems that determine outcomes in battle – to invest in the human behind the machine.
The promise that AI will solve all problems needs a critical look as it threatens to outpace overinvestment in hardware as the misguided apple of the defense industry’s eye. The idea that a new deus ex machina, a “god from Machine Learning,” will magically make officers better planners or NCOs better tactical decision makers is dangerous as it encourages us to accelerate divestment away from human cognitive training. Rather, the future of artificial intelligence in the U.S. military must follow a different track: the augmentation of human intelligence rather than automating it away. It must transform the Army’s ability to capture knowledge and immediately repurpose it to train its formations and the next generation of leaders.
During the Global War on Terror, many of the Army’s highest-return innovations originated with creative soldiers who identified real problems at the lowest levels of the organization and scaled solutions upward. We consistently lose these superior junior leaders to organizations that promise more leadership autonomy and opportunity for innovation. As the Army looks to rapidly modernize the force and maintain a decisive edge over its adversaries, the military must not sacrifice its greatest core product and most valuable export to the national ecosystem – capable, decisive, morally grounded leaders. We need a force that learns faster than our adversaries and can solve complex problems when the generators die, undersea cables sever, satellites go offline, and all that is left is a homo post machina, a human behind the machine.
