Leading Adaptation at the Point of Friction
By: Maj Jessica AichPosted on August 14,2026
Article Date 01/09/2026
What the Expeditionary Warfare School learned by integrating artificial intelligence
Military organizations often assume adaptation requires resources, expertise, or institutional direction. In reality, adaptation rarely begins under ideal conditions. New technologies emerge faster than doctrine; operational requirements evolve faster than policy; adversaries innovate while organizations debate. The challenge for leaders is not simply identifying what needs to change but creating conditions that enable adaptation before certainty arrives. The Expeditionary Warfare School’s (EWS) integration of artificial intelligence (AI) into planning for academic year (AY) 2026 provides a useful case study in how leaders can enable adaptation in the operating forces.1
In the early summer of 2025, EWS had planned to introduce only a single 90-minute lecture on AI and had no plans to develop hands-on AI opportunities for students. However, as increasingly capable commercial AI models became available throughout 2025, the gap between the EWS curriculum and what students needed to succeed in the operational forces post-graduation widened rapidly. The EWS needed to adapt to meet the needs of the students and their gaining units. By February 2026, students and faculty were using AI-enabled analysis, applications, tools, and agents during the school’s culminating amphibious planning exercise. Most notably, EWS students built the tools themselves with no additional staff and no budget.
Despite EWS’s success in integrating AI into planning, the most important lesson from the experiment was not technological. The EWS’s experience reinforced a familiar historical reality: Military advantage belongs to organizations capable of adapting responsibly before formal doctrine, policy, or certainty arrives. Professional military education institutions such as EWS remain essential because they develop leaders capable of adapting responsibly when no clear roadmap exists, and because they provide a proving ground for ideas without the same risks of life and treasure faced by operational units. The EWS’s example offers lessons for any leader facing an adaptive problem, particularly on how to set conditions that enable rapid adaptation. Leaders must articulate a clear purpose, establish responsible risk parameters, unleash and inspire initiative, ensure mastery of the basics that underpin the adaptation, and harden the unit’s resolve to endure through adaptation despite uncertainty and setbacks.
Purpose (Vision/Intent)
When the faculty and staff were developing the curriculum for the 2026 academic year, the plan was simple: Introduce students to large language models, discuss how artificial intelligence might influence the character of war, and move on with the curriculum. As with many PME discussions surrounding AI at the time, the school intended to expose students to the concepts rather than to make them practitioners. However, over the summer of 2025, agentic warfare and the rapid development of commercially available AI models changed the environment. It became clear that students needed hands-on experience in planning with AI-enabled tools before returning to the FMF, their Services, or their home countries, so as not to be left behind. After all, EWS is a “doing” school—theory about AI would not be enough to prepare students for their next assignments.
As a result, by August 2025, the Director of the Expeditionary Warfare School, Col Christopher Steele, directed that AY 26 EWS students plan with AI tools by the spring planning exercise in February 2026. He identified a clear end-state that initiated EWS’s effective adaptation to incorporate AI into the curriculum. While the method of adaptation remained unclear, having a key leader provide a vision for the future was a critical first step.
The second critical step in defining the purpose of incorporating AI into the EWS curriculum was to determine exactly what problem AI could solve for students. At EWS, students execute several demanding planning exercises designed to develop MAGTF officers proficient in the Marine Corps Planning Process. These exercises require significant coordination, information management, staff integration, and product development. In operational units, many of these responsibilities are distributed across larger staffs with dedicated support sections. At EWS, small student groups must accomplish the same intellectual work with far fewer personnel.
Across the curriculum, the director and faculty determined that the enduring problem at EWS was that time spent planning was increasingly focused on production rather than on thinking, communicating, and developing shared situational awareness. Students often spent disproportionate amounts of time formatting products, consolidating and organizing information, building diagrams, creating planning outputs, or recreating routine staff processes. These tasks mattered, but they consumed time otherwise available for analysis, wargaming, and decision-making.
Artificial intelligence—specifically generative AI and large language models—appeared to offer potential advantages to buy back a precious resource: time. These advantages accrue not because AI could automate commanders’ decisions or negate the responsibility of planners but because it could reduce production burdens that were crowding out critical thinking. Once the problem AI could solve was clear, faculty could more easily understand how to integrate AI into the curriculum and guide students in building AI-enabled tools that supported planning. Leaders must provide and clarify the purpose to enable effective adaptation.
Risk (Policy/Boundaries)
The next hurdle in adapting the EWS curriculum to incorporate AI concerned the level of acceptable risk the faculty and students could assume. No EWS faculty members possessed deep technical expertise in AI development. There was no budget item to rapidly onboard qualified instructors. There was no increase in infrastructure to enable student AI experimentation. The longest government shutdown to date temporarily stunted formal AI-agent development initiatives at Marine Corps University. However, EWS could not wait for solutions to emerge and lacked the budget to purchase a commercial solution, even if one had existed at the time. Instead, the faculty had to manage and communicate the risks of the adaptation and effectively leverage their on-hand resources.
Once it became clear that EWS needed to develop AI-enabled tools, apps, and agents without external support to achieve the EWS director’s intent, the faculty designed an all-volunteer, credit-producing elective for students to build AI agents or tools.2 The faculty designing the elective faced a challenge familiar to many leaders when integrating emerging technologies. Excessive restrictions would stifle experimentation, while too few restrictions risked creating unsafe or irrelevant solutions. The EWS faculty needed to clearly articulate boundaries for tool development to ensure sensitive data was protected and that the tools created would improve the learning outcomes for all students.
Faculty guidance to students on the type of tool desired was intentionally broad. Students were to use AI to solve real planning problems relevant to military planners in an amphibious operation. The faculty did not restrict what commercial tools could be used to create the AI agent or tool. The constraints of the elective were equally clear. The tools needed to use only unclassified data and be accessible with a free account to maximize students’ opportunities to use them. Most importantly, the tool had to remain operationally relevant to planning in the FMF.3
While no AI expertise existed among the faculty yet, EWS faculty deeply understand the planning process, the educational objectives for the culminating exercise, and joint all-domain operations. Together, the staff had hundreds of years of combined experience planning complex military operations. The staff relied on their strengths in planning, rather than expertise in AI, to mentor students in developing AI tools relevant to planners. The faculty, with their expertise, could provide mentorship and oversight on the applicability of tools to the planning process, but not train students how to build AI tools. To overcome this lack of technical instruction, students had to rely on their own initiative, teaching themselves the necessary skills to use commercially available AI models to generate customized tools.
These boundaries and the articulation of acceptable risk during adaptation helped ensure that students’ work remained applicable and reduced confusion or hesitation among students. The EWS students adapted rapidly because faculty created enough freedom to experiment while maintaining clear limits on classification, operational relevance, and responsible use. If policy is unclear, leaders must provide supervision, guidance, and clear left-right limits for adaptation.
Initiative
Despite broad guidance and the lack of available instruction on AI tools, 23 students volunteered for the AI elective, approximately 10 percent of the AY 26 EWS class. All the students committed significant personal time beyond normal academic requirements for the elective. Thirteen of them completed the elective and produced working AI-enabled planning tools. Their personal investment of between 10 and 80 hours beyond the already demanding course load demonstrates initiative above and beyond expectations.
The students’ AI-enabled planning solutions far exceeded the EWS staff’s expectations. In one month, the students developed fifteen concepts ranging from custom generative pre-trained transformers to planning support applications and information management tools. With feedback from the faculty and their peers, the students who completed the elective produced ten tools that were operationally viable enough to integrate into general planning exercises, and particularly the EWS capstone amphibious planning exercise.4 One student-developed tool reduced the hours required to produce command-and-control diagrams to between 20 and 30 minutes. Critically, the reduction in production time fundamentally changed the planning discussion itself. Instead of presenting a single command-and-control option to the commander because time constraints limited further refinement, planning teams could now generate multiple complete options and compare their strengths and weaknesses.
Another tool dramatically improved the speed and detail of logistics planning by consolidating sustainment calculations and identifying friction points planners might otherwise overlook during compressed timelines. A separate agent rapidly generated firing solutions based on available munitions, target parameters, and probability-of-kill calculations, allowing fires planners to evaluate multiple engagement options faster than manual calculations permitted.
The EWS’s most valuable adaptations came from students willing to experiment, iterate, and solve problems without waiting for perfect guidance. But what they produced was not just automation or delegating tasks. The value was the restoration of time to think within the planning space and an opportunity for fellow students to consider AI-enabled inputs for the first time. The tools created by the elective forced faculty and students to evaluate how and when to use AI to enhance their critical thinking, expose bias, or offer ideas they might not have considered otherwise. Once unlocked, the students’ initiative produced outstanding results. Leaders must inspire subordinates, foster a climate of curiosity, and recognize and encourage their people to unleash their creativity and innovation.
Mastery (Expertise)
As students innovated approaches to AI-enabled planning, one lesson became apparent: Adaptation without expertise risks creating polished dysfunction. The EWS students who deeply understood planning produced better AI-enabled tools because they could identify flawed assumptions, doctrinal inconsistencies, and operationally dangerous outputs quickly and efficiently. The inclusion of AI in planning did not reduce the importance of expertise at EWS. Instead, it revealed that using AI, coupled with expertise, is essential.
One of the clearest observations from the elective was that AI-assisted planning was only as effective as the student’s expertise. Students with deeper planning proficiency consistently achieved better results because they could rapidly identify hallucinated information, weak assumptions, and doctrinal inconsistencies. Experienced planners understood what “right” looked like before engaging the technology. Those planners recognized when tools were useful and when they were confidently wrong.
Students with weaker planning fundamentals often struggled to recognize degraded outputs. In some cases, AI-generated products appeared polished and tactically plausible while containing subtle but operationally significant flaws. This created a dangerous temptation: confusing polished outputs with sound analysis.
The most successful teams in the elective combined students who were technically curious, tactically proficient in their military occupational specialty, and possessed a strong foundation in planning process expertise. By the start of the elective, students had already learned the Marine Corps Planning Process with “iron sights” before attempting to enhance planning with AI tools. Because of their baseline mastery of the planning process, students could rapidly identify errors in the tools’ outputs and then modify prompts, inputs, or tool functions to improve performance. What emerged was not simply technological experimentation. It became a case study in the importance of expertise as a baseline for effective adaptation. It is a leader’s responsibility to know what expertise exists within the organization, to seek staff input, and to ensure the right expertise is applied to the problem.
Endurance
The inaugural AI elective at EWS did not demonstrate that AI will replace military expertise or automate planning. It demonstrated something more important and more uncomfortable. It demonstrated that future military advantage belongs to organizations capable of adapting rapidly and responsibly before institutional certainty arrives. The EWS AI initiative survived because students and faculty continued adapting despite limited expertise, no dedicated funding, and repeated setbacks. The lesson for leaders is to maintain the course while allowing in-stride shifts in execution to maximize organizational strengths.
The challenges of incorporating AI education into EWS increasingly resemble the broader operational environment facing the Joint Force: incomplete guidance, rapidly evolving technology, compressed timelines, institutional ambiguity, and adversaries adapting in real time. No perfect solution existed. No mature doctrine explained exactly how to proceed. Adaptation occurred anyway because the problem’s operational relevance was obvious. That lesson extends well beyond AI and PME.
The challenge facing the Marine Corps is not simply learning how to use AI. It is developing leaders capable of exercising judgment and leading adaptation while technology changes faster than doctrine, policy, or institutions can comfortably absorb. The EWS experience demonstrated that AI increased the premium placed on experienced planners capable of evaluating outputs, recognizing flawed assumptions, and exercising military judgment. Professional military education remains essential because it develops the expertise, judgment, and leadership required to adapt responsibly before those lessons must be learned in combat. The organizations that adapt fastest will not necessarily be those with the most resources or the most advanced technology. They will be those led by leaders who understand where friction exists and create conditions that allow their people to overcome it.
ABOUT THE AUTHOR
>Maj Aich is an Intelligence Officer currently serving as a faculty member at Expeditionary Warfare School. She previously served as the Military Adaptation and Innovation Course Director and as a Faculty Advisor at Expeditionary Warfare School.
Notes
1. The EWS is the Marine Corps’ 41-week resident career-level PME school for up to 250 Marine, joint service, and international company-grade officers.
2. Maj Jessica Aich, Course Director for Military Adaptation and Innovation, and Dr. Kirklin Bateman, EWS Dean of Academics, led the elective.
3. Dr. Kirklin Bateman and Jessica Aich, AY26 Expeditionary Warfare School Elective: AI Agents Syllabus, (December 2025).
4. The EWS executes a six-week amphibious planning exercise titled Pacific Guard in the spring semester. Teams of up to 32 students form MEU, Marine Littoral Regiment, and Amphibious Ready Group staffs and plan against a Chinese-based adversary in the Pacific Command area of responsibility. The exercise includes a non-combatant evacuation, an amphibious assault, and amphibious defense, and a rapid planning exercise for foreign humanitarian assistance and a small-scale amphibious raid.
Author’s Note: The views expressed are the author’s own and do not reflect official policy or positions of the Marine Corps, DOD or U.S. Government. During the preparation of this work, the author used AI to improve the readability, clarity, and structure of the article.