Skip to content
YC Root AccessYC Root Access

How Onebrief Is Fixing Military Planning With Software

Grant Demaree was a junior US Army officer watching military planners pass around Word docs and PowerPoints during the 2014 Ebola crisis when he realized something: the military’s most critical decisions were being made with outdated tools and slow, legacy workflows. That insight led him to build Onebrief—a platform now used by military headquarters worldwide to dramatically accelerate their planning. But to get there, he had to figure out how to get the military to actually buy software. Grant sat down with YC's Brad Flora to share how they did things that don’t scale to get their first users, closed their first deal with a government credit card, and turned an offline, Navy-approved laptop into a $650M company helping military teams make faster, smarter decisions. Learn more about Onebrief: https://www.onebrief.com Apply to Y Combinator's Summer 2025 batch: https://www.ycombinator.com/apply

Brad FlorahostGrant Demareeguest
May 12, 202524mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Onebrief modernizes military planning, navigating procurement, security, and adoption barriers

  1. Onebrief turns operational military planning from email-and-slide-deck chaos into a shared, structured system where plan components update everywhere automatically.
  2. The company achieved rapid adoption (2.75x QoQ usage growth) by selling command-by-command rather than trying to sell to the Department of Defense as a single monolith.
  3. Early traction came from unconventional validation tactics—paid Saturday planning exercises and tiny government purchase-card buys—to get close to real planners fast.
  4. Scaling required overcoming three persistent defense-specific barriers: contracting, Authority to Operate (ATO), and facilities clearance, with ATO and network access (e.g., JWICS) being pivotal to real usage.
  5. Onebrief is deploying AI first for “mundane utility” automations and aiming longer-term at AI-generated courses of action that outperform human planning speed and quality.

IDEAS WORTH REMEMBERING

5 ideas

Military planning is a high-leverage bottleneck, not a paperwork problem.

Demaree argues decisions—driven by planning—often constrain military performance more than hardware, so speeding planning can create outsized strategic impact.

Replace documents with shared primitives that generate every required output.

Onebrief’s “cards” represent facts, units, and tasks, then produce maps, sync matrices, and ops orders from the same data so changes propagate instantly across artifacts.

Defense isn’t one customer; it’s ~1,200 budget-holding commands.

Instead of selling to “the DoD,” Onebrief targets the command that feels the pain most, enabling faster closes and tighter feedback loops with end users.

Start with procurement shortcuts to earn access, then expand legitimacy.

Their first sale was a single buyer using a government purchase card (often up to ~$25k), letting the team get inside headquarters before mastering full contracting.

If you can’t access users, manufacture realistic usage environments.

They cold-messaged planners to attend paid all-day Saturday exercises, running parallel teams (Word/PowerPoint vs proto-Onebrief) to observe workflows and iterate.

WORDS WORTH SAVING

5 quotes

Uh, well, almost all Word and PowerPoint.

— Grant Demaree

In three hours, they had changed the entire thing—new map overlays, new sync matrix, new written base plan, new slides… commander signs it.

— Grant Demaree

We were able to get the initial sale on a government purchase card.

— Grant Demaree

There is no product market fit before ATO in the military.

— Grant Demaree

Ultimately, where we wanna get people is… AI-generated course of action that are faster and smarter than any human could come up with.

— Grant Demaree

Operational military planning workflows (O-plans, CONPLANs, ops orders)Card-based, reusable data blocks and auto-updating outputsDefense go-to-market: selling to individual commandsPurchase-card “hack” and embedding with usersAuthority to Operate (ATO), JWICS/SIPRNet, SCIF constraintsThings-that-don’t-scale user research (200 interviews, exercises)AI in planning: automation vs course-of-action generation

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome