Share with your CIO
Agile Defense is betting that AI adoption fails not because of bad software, but because of bad integration. The company won a potential $100 million OTA contract to embed co-located engineering and training teams directly inside NORAD and U.S. Northern Command, supporting the DoD’s Maven Smart System and War Data Platform. The model, called Force Deployed Engineering and Training, puts subject matter experts on the floor alongside operators, building and tuning AI workflows in real mission context rather than in a classroom or a vendor demo environment.
What this means for your business
The DoD is paying $100 million to solve a problem most enterprises already have and refuse to name: they bought the platform, deployed the model, and nothing changed operationally. The recurring failure mode looks like a successful proof of concept followed by a deployment that operators quietly route around. Agile Defense’s thesis is that you can’t separate instruction from engineering support, because the gaps only surface during actual use.
This is what might be called embedded adoption, where specialist teams don’t hand off a tool but sit inside the workflow until the workflow actually changes. That’s a fundamentally different cost structure than a SaaS seat plus a training module. It scales poorly in the traditional vendor sense, which is exactly why it works. The friction of getting humans in the room is also the mechanism that produces real workflow change. The DoD recognizing this pattern in a $100 million prototype suggests enterprise IT leadership should be asking whether their current AI deployment model is optimized for deployment metrics or for adoption outcomes.
The signal worth watching: if this prototype converts to a full operational program across Combatant Commands, it will establish a government-validated playbook for AI adoption that defense contractors and enterprise system integrators will immediately replicate for commercial clients. CIOs who wait for that playbook to filter down are already a cycle behind. The question to pressure-test now is whether your organization’s AI integration vendors are teaching your operators or building alongside them.
Concept deep-dive: Other Transaction Authority (OTA)
OTA agreements let federal agencies bypass traditional procurement rules to prototype commercial technologies faster. They exist because the standard acquisition process, which can take years, is incompatible with the pace of AI development. Think of OTA as a structured pilot contract: the government tests an approach before committing to a full program of record. For enterprise leaders, the business connection is direct. The DoD’s willingness to use OTA for an AI adoption model, not just an AI product, signals that the integration methodology itself is now treated as a prototypable, investable capability.
Based on reporting from Agile Defense to prototype embedded AI engineering model for NORAD, USNORTHCOM, originally published 2026-07-16 03:08:00.
