Trace models how tasks flow between departments and then routes each one to whichever should handle it, an agent or a person. It connects to Slack, Notion, Jira, Google Drive and HubSpot, monitors workflow health, raises agent alerts and keeps a full trace of what happened, with approval gates in front of the steps that need them.
Treating people and agents as peers in the same routing layer is the honest design. Most orchestration tools assume the agent does the work and a human handles exceptions, which inverts within weeks of real use; a system where either can be the right destination for a task survives contact with how the work actually divides.
It is on a waitlist with no pricing published, which is the main thing to know: this is not generally available. Full traceability is a claim to test rather than accept, orchestration only helps where processes are defined well enough to model, and routing between agents and humans needs someone owning the rules.
