Intelligence
Intelligence
Intelligence is what the platform knows about the data estate it manages: data quality, the sources and how data flows, capital-governance (CapEx) data, the business rules, and the playbooks that act on them. Most of it is a Flutter port of the partner product Atlas; Estate is ported from FACE. Every figure comes from TIDE’s own services: a figure TIDE cannot compute is drawn as absent, with the reason, never as zero.
The Intelligence rows, in rail order
| Row | ?tab= | What it answers |
|---|---|---|
| Intelligence overview | dqOverview | How good is the estate’s data, through each reader’s lens, and over time? |
| CapEx | capexMonitor | What do the CapEx feeds say, where does a finding come from, and what fixes are proposed? |
| Business rules | rules | Which rules are scored, who owns them, does code match policy, and what is in the knowledge corpus? |
| Playbooks | orchestrator | Which chains dispatch TIDE’s agents, and what did they do? |
If Intelligence is not set up yet, the overview’s tab strip names the first missing step and links to where it is done. The first step is a connection, made under Integrations.
The Data audit, data lineage, and deploy lineage.
Every gRPC-web RPC with its policy class, plus the HTTP routes.
The error messages, what causes them, and the fix.
Three rules every page follows
- Absent is a state, not a zero. A figure that cannot be computed is a
Kpiwith no value and a non-emptyabsent_reason. Each response also carries aProvenance:LIVE,DEFAULT,EMPTY(nothing has been uploaded) orUNAVAILABLE(the store could not be read). - Reads never dial. A page load, a timer or a read RPC never contacts a tenant source. Only an admin’s confirmed Estate action (or a CapEx pull) dials.
- Writes are attributable and audited. Every non-read attempt is recorded in the console’s audit chain, including refusals. With no sign-in configured, every write is refused.