INTERVAL EVIDENCE / LOCAL ONLY

Build Compression Ledger

Separate a shorter calendar from less human work. Every result traces back to occupied time.

Human Agent Fallback

Import CSV

BUILD COMPRESSION LEDGER

Awaiting a ledger

No evidence computed

Calendar compression -- Baseline -> comparison

Human effort reduction -- Merged human intervals

Unattended agent work -- Agent union minus overlap

Outage fallback work -- Explicit fallback rows only

Load the synthetic comparison or import your own two-project ledger.

The audit will expose calendar, actor, overlap, and fallback intervals here.

Results describe supplied evidence; they do not verify the source anecdote or prove causation.

WHY THE NUMBERS DIFFER

The source motivates the question. The ledger supplies the evidence.

Overlapping rows for one actor occupy the same hour. The engine merges them before totaling active effort.

actor effort = hours(union(actor intervals))

Agent time is not automatically unattended. Human-agent intersection is removed from the agent union.

unattended = agent union - intersection

A shorter calendar can also reflect different scope, quality, staffing, or missing rows. The export preserves that limitation.

comparison is evidence, not attribution

Assumptions

UTC intervals, user-supplied actor labels, merged same-actor overlaps, and explicit fallback labels.

Limitations

No live telemetry, agent execution, historical verification, scope normalization, quality adjustment, or causal inference.

Transfer

Use two comparable project exports and explain whether the observed change is calendar compression, human-effort shift, or both.