AI automation
ArtLifting
Unblocking a catalog migration with 94% automatic image matching
Automated reconciliation of approximately 20,000 artwork images unblocked the catalog migration onto ArtLifting's new ERP, turning a weeks-long manual audit into days.
Service: Prototype to Product
~20K
Archival images reconciled
94%
Matched automatically
Weeks → days
Audit time
The problem
ArtLifting was launching a new ERP and had to reconcile approximately 20,000 archival artwork images from ten years of history against its database. Images lived across multiple systems with inconsistent naming. A manual audit would have taken weeks of staff time, and reconciliation was blocking the catalog migration.
Match what is clear
Automatic matching
- Exact, pattern, and fuzzy rules
- Matches above a confidence threshold
Review the exceptions
Human review
- Lower-confidence cases
- Rules evolve as edge cases surface
New edge cases informed matching rules.
What I did
I built a matching system that combined exact, pattern, and fuzzy matching to handle inconsistent names. Matches above a confidence threshold were applied automatically; cases below that threshold went to human review.
The rules engine refreshed on a short cycle so matching rules could evolve as edge cases surfaced.
The outcome
The reconciliation completed with 94% of the full set of approximately 20,000 images matched automatically. The audit took days instead of weeks of manual work.
That unblocked the catalog migration onto ArtLifting's new ERP.
“20K+ images, 94% matched automatically — Don saved us weeks of work.”