Orbiqent is a governance layer for AI systems that make investment decisions. It records what the system was looking at and why, at the moment it decided. It is not reconstructed afterwards. And it holds authority over money outside the model, so that what a system can do and what it is permitted to do move separately.
Autonomy levels
Levels are defined in advance, along with the conditions that unlock each one. Movement between them is not a setting. The naming here describes the model; the platform's own levels are defined against its specific trading context.
When the system proposes a name, it persists what surfaced it: the screen, the named factor behind it, and the conditions evaluated at that moment. The record is written as the decision happens, so it is contemporaneous rather than assembled later. A reviewer can ask what the lens was: value, discounted asset value, relative strength. The answer is a named factor rather than a score.
The system operates at a defined level of independence. The model defines progression around measured performance against conditions fixed in advance, and a human review recorded alongside the evidence. The evaluation runs today; the enforcement path from evaluation through to approval is still being built. A model can become dramatically more capable without gaining any additional authority over money. Authority is held outside the model.
The console is written for a reviewer with no knowledge of the internals. That includes an explicit account of what the platform does not yet measure and why, because a governance layer that reports only its strengths is not a governance layer.
Contact
Orbiqent is in development and in active conversation with securities-registered firms about what a governance record needs to contain to be useful. If you supervise an AI system in an investment process, or expect to, the conversation is useful in both directions.
Built in Prince Edward Island, Canada. In development, with discovery conversations underway with securities-registered firms.