Collect: title, specification, price, URL
Normalize: brand, unit, pack
Candidate: code, attributes, similarity
Review: exact, comparable, not comparable
Monitor: version, evidence, change
Prove comparability before comparing price
One name may hide volume, flavor, pack count, gift, membership, or delivery differences. Comparing the lowest visible price creates false gaps.
Map internal SKUs to exact, unit-adjusted, substitute, or incomparable competitor products and retain evidence.
Normalize decisive attributes
Extract brand, model, volume, unit, pack, color, audience, and barcode. Convert 500ml×2 and 1L to unit price without ignoring channel and pack differences.
Store list and paid price, promotion conditions, shipping, tax, membership, and collection time.
Combine rules and review
Barcode and model are strong signals; name and attributes retrieve candidates. Have operators review high-value or uncertain items.
Version match owner, confidence, and validity. Recheck after URL or specification changes.
Separate collection errors from market change
Layout, stock, region, and login can change collection. Verify screenshot, URL, and job status before acting on an anomaly.
Deduplicate repeated events and disclose freshness and retry behavior.
Accept with difficult products
Test bundles, gifts, member prices, import tax, sizes, similar models, and relisting.
BI0 collection and price-analysis directions can be evaluated; site compliance, login, anti-bot, matching, and latency require explicit confirmation.
Public references
Take the next step with BI0.AI
Talk through a real business scenario and see how governed AI BI can fit your team.
Explore BI0.AI