Turning the judgment your best people carry into capability the institution keeps
Knowledge that decides outcomes rarely lives where it is needed. It sits in a four-hundred-page manual, an unread wiki, or one veteran's head: authoritative, and inert at the moment of the call. KIANI took that failure at its hardest edge, the US Marine Corps call-for-fire decision, where the wrong asset or a slow description is measured in lives, and treated it as a strategy problem rather than a content problem.
Every institution claims its critical knowledge is captured. Almost none can show it is retained. The call-for-fire decision exposes the gap with no mercy: under pressure, an operator must choose which asset, who to call, how to describe the target, and what effect to ask for, correctly, in seconds. The doctrine that governs that choice is complete, current, and authoritative. It is also a document, and a document cannot rehearse anyone.
This is the universal failure wearing its most demanding uniform. The cost is not that people lack the reference. It is that reference, however good, has never once produced fluency under pressure, while the leaders who own training keep funding the reference and wondering why the decision still degrades.
KIANI worked the problem as a strategic chain. First, it identified how authoritative geospatial data from G-EGD could be mapped into realistic, current-terrain scenarios, grounding practice in the actual world rather than abstract training aids, with restricted identifiers handled appropriately. Then it identified where AI earns a place a static manual never can: generating adaptive scenarios, scoring each call against the standard, and debriefing the decision, turning reference into rehearsal with feedback.
On that foundation KIANI drafted the concept end to end: restructure four hundred pages of doctrine around the decision sequence itself, the five transmissions of the call-for-fire, then deliver it as AI-guided drills running on real terrain. Finally it positioned the work for what it is to the people who own the training problem: a substantial improvement in knowledge retention, not a content refresh. The firm's standing in the domain is the authority behind that case, not a footnote to it.
The thesis is simple and it compounds. Take the knowledge that matters most, restructure it around the decision it serves, ground it in real data, identify where AI does what paper cannot, and convert reference into practice with feedback. A well-structured reference already beats a field manual by an order of magnitude; rehearsal that scores and debriefs every decision is not an improvement on that. It is a different ceiling entirely.
The discipline is in the seam. Format held strict where it cannot bend: the five transmissions are exact or they are wrong. Judgment kept flexible where it must, since terrain, target, and effect change every scenario. AI is the medium that makes this work at scale; the insight is what decides where it belongs.
Strip away the uniform and the problem is everywhere: onboarding that fades by week three, an internal wiki no one opens, a runbook that lives in one person's head and walks out the door with them. Same failure, quieter stakes, identical cure.
The move does not change. Map the data and the decisions that actually get made. Identify where AI earns its place and where it does not. Design the practice that makes the knowledge stick. KIANI proved it in the domain that allows the least margin for error. That is the strongest evidence it holds in the ones that allow more.