Every workflow running here was configured in
Zen Studio, and started as a conversation with
Zara. This is where it actually lives.
Every deployed workflow, at a glance
Filter by specialty. Check status, version, and when it last changed, without opening a single run.
v162a workflow updated in June, still running the case it was built for
100%online and ready, shown per workflow, not assumed
Liveevery version timestamped: when it changed, when it was added
Every case, tracked while it's actually happening
Click into a workflow and the individual pipeline runs underneath it are right there, named, timed, and stepped through in real time, not reconstructed after the fact.
Three named cases can be running at once, each with its own progress bar, its own duration, its own step count. One file doesn't wait behind another. They move through the same governed cycle in parallel, and both are visible the entire time.
Nobody can show the check stops being the objection the moment you can click a node and read exactly what it found.
Every decision, explained in language a person can actually read
Click any node in a completed run. The rationale is a sentence, not a log a specialist has to translate for the compliance team.
Plain languageA rationale a reviewer can actually read
Not a confidence score standing alone. A sentence explaining what was checked, what tool ran, and what it led to.
Underneath itThe numbers behind the sentence
Risk level, risk score, the exact node that produced the figure, and the inputs it used to get there, all on the same screen.
Most audit trails get reconstructed after someone asks for them, stitched together from logs nobody designed to be read by a person. Here, the record exists because it was built the moment each step ran, not because someone went looking for it three months later when a regulator finally called.
The control plane: always on, owned by nothing
The parts of the platform that stay separate from any single workflow, because governance a workflow could switch off isn't governance.
WatchingSupervisor agent
Monitors every running node for stalls, drift, and schema violations, and flags in real time.
ControlledKill switch
Controlled termination of a running workflow, so a runaway run costs seconds, not an afternoon.
AttributableOpenTelemetry tracing
Every step traceable to the node, the model, and the agent that produced it.
One numberRisk Control Index
Covers guardrail coverage, escalation discipline, and evidence completeness in a single score.
Silent Agents: work that starts before anyone has to ask
Proactive triggers that sit above your systems of record. Event or threshold based, and governed by the same SPAR cycle as human-initiated work.
Webhook / EventEvent triggers
Starts the moment a core banking, claims, or CRM event fires. Zero polling lag.
Cron / ScheduleSchedule triggers
Runs recurring sweeps: daily sanction list diffs, weekly portfolio stress tests, monthly compliance reports.
ThresholdDrift triggers
Activates when an operational metric breaches tolerance (e.g. claim volume spike, fraud confidence dip).
Common questions about Zen Pilot
Can regulators access Zen Pilot audit trails directly?
Yes. Read-only compliance auditor roles can export self-contained defence files containing model versions, prompt snapshots, and cryptographic proof of human sign-off.
What telemetry standards does Pilot support?
Full OpenTelemetry (OTel) compliance. Spans and metrics can be piped directly into Datadog, Dynatrace, Prometheus, or Grafana.
How fast does the kill switch take effect?
Instantly. The runtime halts downstream agent execution within 50ms, revokes active tool tokens, and commits the state up to the halt point.
