Twelve days of synchronization, one dead radio
The wargame ends at 2140 on a Thursday. The planning cell has been at it for twelve days. Every phase is synchronized against every other phase. The fires plan talks to the movement plan. The comms plan names a primary net and two alternates for every element. Decision point three has a trigger, a named authority and a branch plan behind it. All of it lives on one shared surface where a change made by the logistics planner shows up on the operations planner's screen before he finishes his coffee. The colonel calls it the best plan the staff has produced.
The team steps off at 0400 on Saturday. The plan travels with them as a PDF on the team leader's laptop and a stack of printed sheets in a waterproof bag. At 0447 the primary net dies in a valley the propagation study had flagged as marginal. Decision point three depends on a report from the flank element that now cannot be sent. The team leader waits eleven minutes and makes the call from memory. The branch plan that covered this exact case is on page 31 and nobody has time to find page 31.
The staff learns what happened at the after action brief four days later. The slides get built from memory and whatever notes survived the ride back. The eleven minutes at decision point three come out as one bullet. Twelve days of synchronization produced a plan that stopped being synchronized four minutes after the first radio died, and the machinery that built the plan never found out.
Planning got fast
Credit where it is due. The last decade produced excellent planning software, and OneBrief is the standout. It gives a staff one live surface for operational planning and wargaming instead of forty briefing decks passed over email. Combatant commands adopted it and kept using it. Investors kept funding it because the product earns its seat in the planning cell. Work that used to take a staff days now takes hours. That is a real revolution in a domain that resists revolutions.
The revolution stops at the moment of execution. The plan improves at software speed right up until someone hits export. Then it becomes a document, and documents do not hear radios die. Feedback from the mission comes back as slides at the after action brief, days or weeks later, filtered through memory. So the staff plans in hours and learns in weeks. Half the loop runs at software speed and the other half runs at PowerPoint speed.
Ukraine shows what closing that distance is worth. CSIS documented how Ukraine's military went digital under fire. Battlefield data flows into shared systems and the picture updates while the fight is still on. Chatham House found the same pattern in its study of Ukrainian wartime innovation. The units that adapt fastest are the ones where the gap between doing and learning collapsed to days. A force that plans fast but learns slow gets half the benefit of its planning tools.
A product handed to a process
The plan is a product. It has a version number and a distribution list. The mission is a process. It runs and branches and degrades in real time. The failure at 0447 was not a bad plan. The plan had the answer on page 31. The failure was handing a product to a process and expecting the product to keep up.
Every planning artifact points at execution. The loadout says what each element carries. The task organization says who holds which authority. The decision points say what gets decided and on what trigger. On the planning surface these are live objects that update when anything around them changes. In the PDF they are ink. The moment of export is the moment the plan loses the ability to answer questions, and the mission is nothing except questions arriving faster than anyone briefed.
The plan becomes the workspace
EdgeLance is the execution side of that loop. The plan's loadouts, roles and decision points import as live objects in a mission workspace running on hardware the team already carries. Decision point three exists as a tracked item with its trigger and its branch attached. When the primary net dies the workspace keeps running, because every node carries the mission locally and elements resynchronize whenever any link returns. The team leader who made the call from memory instead taps the decision point and reads the branch. Page 31 finds him.
Execution writes the record as it happens. Reports and decisions land in the mission ledger with a timestamp and an author. So does the dead net at 0447. Nobody reconstructs the eleven minutes from memory because the eleven minutes are on the ledger. The after action review assembles from the record instead of from whoever remembered to take notes, and it exists the moment the team steps back inside the wire.
None of this competes with the planning surface. A staff that plans on OneBrief should keep planning on OneBrief. EdgeLance picks up where the export used to be. It carries the plan through contact and hands back a record the planning tool can actually use.
The record is the next plan's ground truth
The mission record changes what the next planning cycle starts from. Today a staff wargames against assumptions. The propagation study said the valley was marginal. The mission record says the primary net died there at 0447 and stayed dead for eleven minutes while a decision waited. That is ground truth with a timestamp, and it flows back to the planners as data instead of as a bullet on slide fourteen.
Planning software and mission software are two halves of one loop that has never been closed. The planning half got very good over the last decade. The mission half is what EdgeLance builds, and the loop is what a force gets when both halves run. Plan on the shared surface. Execute in the live workspace. Feed the record back and plan again from what actually happened.
The next time that team steps off, decision point three carries the eleven minutes with it. The trigger accounts for the dead valley and the branch sits one tap away instead of thirty pages deep. The colonel gets his best plan back with the mission attached.