The campaign process

From frozen build to flight evidence.

Every FlightLab campaign begins with a specific decision, a controlled software build and an agreed evidence plan.

Campaign principles

The smallest safe campaign capable of changing the decision.

Every claim connects to a measurable test. Every result connects to an approved build. Every anomaly preserves its context. Every limitation is stated clearly.

Decision first

Establish what the customer or decision authority needs to learn.

Measurable claims

Connect every permitted claim to an acceptance criterion and test.

Configuration control

Keep each result traceable to the build, hardware and operating context.

Explicit limits

Record what the campaign did not test and where conclusions stop.

Seven decision gates

A controlled path from readiness to evidence delivery.

01

Flight Readiness Sprint

Evaluate workload, trigger, dependencies, data path, compute needs and mission constraints. Exit: fly, defer or remain ground-only.

02

Evidence-plan definition

Agree criteria, methods, outputs, operating limits and permitted claims. Exit: approve the campaign baseline.

03

Ground-twin acceptance

Execute the signed build on matched hardware with representative vectors and injected faults. Exit: admit, remediate or reject.

04

Mission acceptance

Complete operator, safety, cybersecurity, export and configuration review. Exit: accept for orbital execution.

05

In-orbit execution

Run scheduled windows under independent supervision and resource quotas. Exit: complete, extend or terminate safely.

06

Evidence correlation

Compare flight and ground behaviour across timing, resources, anomalies and recovery. Exit: approve findings.

07

Evidence delivery

Deliver the signed dossier, executive proof pack and recommendations. Exit: authority evaluates readiness implications.

Controlled boundary

Campaign timing depends on platform availability, integration complexity, safety review and launch schedule. The Sprint establishes customer-specific gates and assumptions.

The next controlled step

Do not build an entire space mission to answer one software question.

Start with a bounded workload, a defined evidence requirement and a clear decision trigger.