In our last article we argued that the gap in critical operations sits after detection, in the moment one officer has to read everything, recall the right procedure and act on it. OpCopilot 1.5 closes that gap, and it is the first release we are presenting in full.
OpCopilot is a team of tireless watch agents that have read every manual your organisation owns. They watch the data, look things up, run the calculations and draft the paperwork. The operator directs them, approves what they propose and gives every order. By default, nothing they do runs without the operator's approval.
What's new in OpCopilot 1.5
01 / Doctrine
Your doctrine, encoded
Upload your standards, guidelines and reference material as they are, whether they are written documents, scanned pages or diagrams. OpCopilot turns this living doctrine library into a concept of operations its agents follow.
02 / Insight
Live operational insights
OpCopilot agents subscribe to your live data streams and give a first read on every event as it arrives: what it is likely to be, what it connects to and what your procedures say to do next. A beacon that activates, a buoy light that fails at dusk, a vessel that goes quiet near the coast or a movement at a remote post gets the same attention, whatever the hour.
03 / Collaboration
Incident rooms
Every incident gets its own room, where operators and agents work the case together from the first alert to the final report. Everything said, decided and drafted stays with the incident. Rooms can be open to the whole watch or kept private.
04 / Mapping
Operational maps
The agents use the map the way an officer uses a chart table. They draw zones, stack layers, plot targets and tracks, and calculate drift or an intercept. Everyone in the room sees it live.
05 / Reporting
Reports on demand
Situation reports, handover notes and any report your concept of operations calls for are drafted on a live canvas from the incident room. They follow your templates, whether you provide them or the agents find them in your doctrine library.
06 / Memory
Operational memory
Every incident your team works becomes part of your organisation's operational memory, and OpCopilot's agents bring that expertise to every operator on watch. More on this below.
07 / Sovereignty
100% sovereign
OpCopilot's agents work with several AI models and understand text, images and documents, all running on your own hardware. You get frontier-level intelligence with no AI subscription, no calls to an outside model and no data leaving your infrastructure.
08 / Deployment
Edge and Server
OpCopilot Edge runs agents on compact models close to the sensors, at a surveillance station, an outpost or on board. OpCopilot Server runs larger models with frontier capability and serves a whole organisation from the command center. Both reason the same way, so a station and its command center work from the same procedures.
Turn every incident into expertise your organisation owns
Of everything in this release, memory is the capability we care about most.
Most of what makes an operations center good lives in the heads of its most experienced people, and it leaves with them at handover, rotation or retirement. Documentation never captures all of it. So no AI system is complete on its first day, and choosing the right documents to load is harder than it sounds.
OpCopilot does not depend on getting that right at the start. As operators and agents work incidents together, OpCopilot builds the organisation's operational memory: what was verified, what was decided and what worked. Operators and agents learn together, and every incident leaves the next one better informed.
That memory is an asset your organisation owns. It sits on your own infrastructure, grows with every shift and stays when your people move on. OpCopilot's agents bring it to whoever is on watch, so a newer operator at a remote station works with the benefit of everything the organisation has learned.
Built for the people on watch
For newer operators, OpCopilot is a colleague who knows the procedures by heart and explains them. Every recommendation comes with the rule behind it, so they learn as they work and reach independence sooner.
For experienced operators, it takes on the parts of the job that cost the most attention: switching between cases, holding too much information at once, writing reports and rebuilding the timeline afterwards. Every case keeps a complete record of what was known, recommended and decided, which makes handovers, audits and investigations far easier.
See it working
The video walks through OpCopilot 1.5 on a simulated rescue coordination watch, driven by our scenario simulator. In under five minutes it shows the whole loop:
- The center's procedures and reference material loaded into the doctrine library.
- An agent introducing itself to the duty desk, including the what it can help the operator with.
- A live case: distress alerts, relays and on-scene reports from one vessel arriving as separate events, grouped into a single incident with its own room, and the agents' first read with numbered next steps.
- The incident's memory: the agents recording what the case has established and calling on it as they work.
- The situation report drafted on the canvas, as an operational picture and in the standard message format, with the decisions still waiting for the coordinator.
- The reported positions plotted on the map with their times and sources, and a plain note on the events that had no position to plot.
- A dashboard across every open incident on the watch, from a distress case at sea to a radar intrusion in a coastal exclusion zone.
The agent never declare a phase, task a unit or send a message on their own, and where the data does not support a position or a number, they say so instead of inventing one. The same agent harness runs for a border sector, a port, a coastal watch or any control room working live data against its own procedures.
Availability
OpCopilot 1.5 is available now as OpCopilot Edge and OpCopilot Server, deployed as sovereign AI on your own infrastructure.
We deliver it through forward-deployed engineering. Our engineers embed with your operations team from day one, on your infrastructure, and configure OpCopilot to your doctrine, your systems and your sensors alongside the people who will use it. Engagements are staged, from proof of concept to proof of value and then to full deployment, and each stage stands on its own.