A multi-step agent chain where every image has to land on the right product.
A chain of agent steps that took product images from background removal to a staging store to production, built around one requirement: every asset reaches the right product.
What needed
solving.
Product images had to pass through several steps before they were live: background removal, upload to a staging store, assignment to a product for review, then sync to production. In a chain like that, the risk is less that a step fails and more that an asset ends up on the wrong product.
What I decided,
and why.
- 01
Chain the steps with agents: AI background removal, staging upload, assignment to a product, production sync.
- 02
Put a review step before production: assets are assigned to products for review before anything syncs.
- 03
Use the same approach for mass content migrations between Shopify stores.
What shipped,
and what isn’t proven.
Every asset reached the right product, and the same approach carried mass content migrations between Shopify stores. Not proven: repeatability. So far I have used it one-off, not as a standing process, and nothing about it was measured.