FURN | AI SEARCH OBSERVATION RECORD
Guide: https://furncmo.com/blog/furniture-ai-search-optimization
Updated: October 21, 2026

1. SCOPE AND ACCESS
Owner / date / product family / target public pages:
Intended search access policy:
Separate training permission decision:
Provider documentation URL and check date:
Deployed robots / page directives / hosting response:
Verified crawler evidence or unavailable logs:
Approved changes / owner / follow-up:

2. PRODUCT READINESS
SKU and variant / source record:
Dimensions and units / materials / care:
Price / stock / delivery / returns / verification date:
Visible facts and structured data agree? Evidence:
Internal links / image fidelity / unresolved questions:

3. REPEAT FOR EVERY PROMPT
Exact prompt / buyer question / expected supporting page:
Service / model or mode / search enabled / unknown settings:
Date and time / location / session context:
Full response record / source links / destination checks:
Brand mentioned? Linked retailer citation? Record separately:
Fact checked / approved source / correct, incorrect or unknown:
Test unavailable? Reason (do not record as zero citations):
Observed error / proposed correction / owner:
Repeat conditions / next review date:

4. SEPARATE OUTCOME RECORD
Prompt audit scope and counts (not market share):
Referral visits / period / source visibility / attribution limits:
Qualified outcome definition / verified completions:
Promotions, inventory, other campaigns and measurement changes:
Costs and comparison design before any causal or ROI claim:
No citation guarantees. A crawler request is not a shopper.
Do not include private customer information in test prompts.