Mission Adaptation at the Speed of Operations

About

Operational planning is driven by conversation

Commanders and planners refine ideas through working groups, huddles, and planning sessions where decisions, assumptions, risks, and tasks emerge long before they appear in an order.

During Project Convergence Capstone 6 (PCC6), Legion embedded alongside the Future Operations (FUOPS) planning cell to better understand how AI could support real operational workflows. Rather than beginning with predefined software requirements, the team focused on understanding how planners carried out their mission and where AI could reduce friction without disrupting existing processes.

Challenge

Planning context is fragmented before it reaches an order

Planning discussions generate critical operational context, but much of that information remains fragmented across notes, documents, chat messages, and individual memory. Reconstructing conversations after the fact requires planners to manually identify tasks, open questions, key assertions, and other planning details that inform future Courses of Action (COAs).

Traditional software development compounds this challenge. Many operational workflows only become visible during mission execution, making them difficult to identify through requirements gathered before deployment. As a result, valuable opportunities to reduce planner workload often go undiscovered until after the mission is complete.

Solution

A planning capture workflow built overnight

During discussions with the FUOPS planning cell, planners explained that critical planning discussions required extensive synthesis after the session, with a dedicated note taker manually capturing tasks, open questions, and assertions that informed future COAs.

In response, Legion's embedded engineers developed a Planning Discussion Capture & Synthesis workflow overnight. The workflow was designed to:

• Capture planning discussions for post-session review.
• Automatically identify tasks, open questions, and key assertions from planning conversations.
• Reduce the manual effort required to reconstruct planning discussions and support follow-on planning activities.

The workflow demonstrated a different approach to delivering mission software. Rather than relying solely on predefined requirements and scheduled release cycles, Legion adapted the software in response to an operational need identified during the exercise. Today, that adaptation is performed by Legion engineers working alongside the customer. Tomorrow, customer builders will create and evolve these same mission-specific workflows, allowing software to adapt as quickly as operational requirements change.

Delivery Highlights

What the deployment showed

By adapting to an operational need identified during PCC6, Legion demonstrated a new model for delivering mission software.

• Workflow discovered through operations: A new planning workflow emerged directly from operator feedback during the exercise.
• Capability delivered overnight: Legion transformed an operational need into a working workflow in less than 24 hours.
• Software that evolves: Built around operational needs instead of release cycles.
• Customer builders: Today, Legion engineers create mission-specific workflows. Tomorrow, customer builders create and evolve them themselves.

"I think Legion has great potential. The idea of localized AI that doesn't require cloud is super valuable, especially if we don't have the access to cloud internet." — Operations Officer, 4/10 Special Forces Group

‍