Enterprise platforms solved the headquarters problem
Enterprise C2 and data platforms have done what a decade of programs could not. Lattice connects sensors and effectors into one real-time picture. Maven Smart System fuses imagery, feeds, and intelligence data for analysts and commanders. Headquarters see more, and decide faster, than they did five years ago.
Those platforms are built for the echelons that have bandwidth, compute, and staff. That is the right design for the job they do. It leaves a separate question open: what does the team at the far end of the link run, and why would they choose to use it?
Adoption at the edge is a trust question
At the team level, software gets used when it helps the person carrying it and when that person understands where the data goes. Every sensor feed, position, and wearable reading is useful to the team in the moment. The same data, sent upward with no context and no rule about who reviews it, changes how people behave around the device.
This is not an argument against sending data up. Command needs the picture and the enterprise needs the record. It is an argument about defaults: what stays with the team, what is promoted, and who set the rule.
The trust problem is real and it is costing capability
A study by the Netherlands Armed Forces on personal health monitoring found that operators accepted wearable data collection when it was framed as a tool for their own safety and managed by their medic. Acceptance dropped sharply when the same data was accessible to command or external review. Same capability, different trust architecture, different outcome.
This pattern repeats across every domain. Operators use ATAK because they control it. They are slower to adopt tools when they do not know where the data goes or who reviews it. The result is that units carry devices they do not use and make decisions with less information than the technology could provide.
The DoD AI Strategy memo calls for AI that operates 'on-board, in real time, and often without any sort of connectivity.' The operational argument for edge AI is bandwidth and latency. The trust argument is stronger: if the AI runs on the operator's device and the data stays with the team, the operator will actually use it.
What operators actually want
Operators want mission technology that works when everything else is broken. No comms, no cloud, no SATCOM, no cell signal. If the tool dies when the link drops, it is dead weight they carried for nothing. It also has to help them do their job without creating a liability: every frame of video and GPS breadcrumb is a potential exhibit in a review they did not ask for. And they want to control what gets shared, when, and with whom.
That is not paranoia. It is learned behavior from twenty years of operations where helmet camera footage, radio transcripts, and engagement data were reviewed by people who were not in the fight and did not understand the context. Decisions made in seconds got second-guessed by people who have never been in those conditions.
The technology that earns operator trust belongs to the team. ATAK works because the operator controls it. Radios work because comms are ephemeral. A paper map does not record who looked at it. The software layer on top of all this needs to follow the same principle: useful to the team, controlled by the team, disposable by the team.
A team layer that feeds the enterprise
The premise is simple: the mission starts with the team running it. AI, data, models, detections, and decisions live on the team's hardware by default. The team keeps working with no cloud, and nothing persists after the mission unless policy says it should.
Each mission is its own loadout. Models, reference data, and prompts are packaged for the specific operation. When the operation ends, the unit keeps what its policy requires, such as the summary, selected evidence, and the after-action record, and destroys the rest with proof.
Exchange with enterprise platforms and with command is set by the unit and its chain of command. EdgeLance publishes tracks, reports, and selected intelligence upward over open standards when that policy allows, and every promotion is recorded. Command gets a cleaner picture with provenance attached. The team gets a tool it has reason to trust. The two layers work together.
Trust comes from architecture
Enterprise platforms serve the enterprise layer and are improving quickly. The open ground is at the team level, where operators need AI, sensors, mesh, and mission tools they trust enough to use every day.
That is what EdgeLance was designed around. Local AI on the team's hardware, a mesh that works when nothing else does, TAK integration on the map operators already use, and a mission lifecycle with clear rules about what is kept, shared, and destroyed. The better that layer works, the better the data every platform above it receives.