Skip to content
Trust layer evolution

How trust evolves

Two event types. One trust signal.

Trust is not static. It shifts in response to definition changes (severity events) and agent decisions (inference / decision events). Trigger both kinds below and watch the same consumption confidence signal absorb each one.

Sample metricMonthly recurring revenue
CC92%
GradeA
Last eventBaseline, no events yet
Event type 1

Definition changes

When a metric definition changes, the severity classifier counts blast radius across static signals (lineage, freshness, ownership, discoverability) and responds proportionally.

What changed?

Select one or more definition changes. Blast radius and consumption confidence response update below.

Blast radius
No change selected
Select one or more changes to see the blast radius.
Lineage

Upstream sources and transformations

Unaffected
Freshness

Recency of the underlying data

Unaffected
Ownership

Accountable human or team

Unaffected
Discoverability

Name, description, tags, docs

Unaffected
Projected CC response
Now92%
Δ-
After92%

Select one or more changes to preview the impact

Event type 2

Agent decisions

Every agent action produces two correctness events, inference (the read on the world) and decision (the action that followed). The quadrant separates them so robustness becomes measurable and decision logic becomes falsifiable.

Scenario

Agent retrieved Monthly Recurring Revenue at $4.2M with trust score 87. Question: should it auto-act or escalate?

Inference

Was the read on the world correct?

Decision

Was the action that followed correct?

Projected CC response
Now92%
Δ-
After92%

Select an inference and a decision to preview the impact

Unified signal

One signal, two streams

Both event types feed the same consumption confidence signal. The architecture has one trust state that absorbs both schema types.

CC over time
Event timeline

No events yet. Apply a definition change or log an agent decision above.

The unified abstraction

Severity events and decision events have different schemas. The architecture treats them as different event types. They share one trust signal.
That is the unified abstraction.

Part of: Stage 04 · Catching Drift

Back to the map

A definition changes on a Tuesday and every downstream number quietly means something else. The layer classifies the change by blast radius and responds in proportion, instead of sounding one binary alarm, and watches the consumer, because a VP who hesitates is a faster signal than any dashboard.

Read the essay