AI LEAD EVIDENCE REVIEW — Lonta AI — https://lonta.io/blog/ai-search-lead-attribution/
Original worksheet CC BY 4.0. Keep actual customer records private.

Review period and cutoff/time zone:
Outcome definition and currency:
Technical source rules/version:
Person responsible for capture and job matching:

For each inquiry:
Inquiry ID and timestamp:
Job ID, or UNMATCHED (keep outside job totals):
Captured technical marker and original capture system/time:
Customer's own source explanation (optional):
Known staff/test/bot flag and reason, if known:
Other source markers (including paid contact evidence):
Job outcome as of cutoff: paid / pending / lost / unresolved
Observed amount collected; UNKNOWN if not yet known:
Match evidence and unresolved conflicts:

Reconcile before reporting:
Raw contacts:
Unique jobs:
Known excluded jobs:
Held/conflicting jobs:
Unmatched inquiries (separate):
Eligible jobs:
T = eligible jobs with recognized technical markers:
R = eligible jobs with affirmative customer-reported AI assistance:
O = eligible jobs appearing in both lists:
U = T + R - O (unique jobs in either list):
Paid jobs within U and observed collected amount:
Pending/unresolved outcomes within U:
Missing capture and known limitations:

Report source rules, cutoff, matching gaps and holds beside totals.
Use evidence-associated job counts; this worksheet does not estimate jobs caused by AI.
Do not add crawler requests, AI impressions, visits and jobs into one total.
