Pilot evidence

Measure the phone work, not the pitch.

The pilot has not started. This evidence board reads only durable live CALL-E records, excludes every fixture, and pre-registers the measures before consenting calls begin.

Awaiting consenting pilot calls

No real-world performance claim is published yet.

Loading verified records

Median sourcing time

Request creation to terminal CALL-E result

Loading verified records

Completed-result rate

Completed supplier records ÷ attempted suppliers

Loading verified records

Quote completeness

Required structured fields returned

Loading verified records

Compatible options

Fitment-confirmed offers across live requests

Loading verified records

Average price spread

Verified high-to-low spread where two quotes exist

Loading verified records

Human intervention

Requests with failure, missing quote, or fitment uncertainty

Evidence denominator

Every attempt stays visible.

Counts come from durable live CALL-E records. Fixtures, failures, and incomplete outcomes are never silently converted into successes.

Open machine-readable evidence snapshot
Live requests
0
Completed requests
0
Supplier attempts
0
Completed supplier results
0
Fixtures excluded
0
Snapshot generated
Preparing snapshot

Minimum credible study

Five requests, consenting businesses, one fixed rubric.

Each request should use a real vehicle and part, contact two or more businesses that agreed to receive an AI-assisted sourcing call, and preserve the resulting structured record.

Fixture runs, unanswered calls, failures, and incomplete quotes stay in the denominator. No synthetic outcome will be presented as pilot evidence.