AI Agent Delivers 2× Faster Shelf Audits with Video Recognition

Effie’s merchandising AI agent now runs on live video, not photos.

The rep scans the shelf. The agent watches it, flags the fix, and checks the work.

Until now, capturing a shelf meant stopping to frame and shoot photo after photo. We changed that. With video recognition, the merchandiser moves the phone along the shelf in a single pass.

When something is off, a gap, a misplaced SKU, or a facing that doesn’t match the plan, the app highlights it on screen right where the rep is looking.

The part that matters most to us is what happens after the fix. The agent validates the completed task and writes it back as structured execution data you can rely on, closing the gap between telling a store what to do and knowing it actually happened.

What this gives field teams:

• One continuous scan instead of a stack of photos per section
• On-screen guidance during the visit, not a report afterward
• Shorter visits with fewer manual steps
• Every completed task checked by the agent rather than self-reported

Video Recognition cuts shelf-audit time by more than half compared with traditional photo capture. A shelf audit requiring six photos takes approximately 120 seconds with Image Recognition, versus 45–60 seconds with Video Recognition, while maintaining 95–99% product-recognition accuracy.

For the business, this means more consistent Perfect Store execution and near-real-time visibility into what is actually on the shelf. For Shelf Execution, it introduces a new way for AI Agents and merchandisers to work seamlessly together.

We’re not interested in giving reps one more AI tool to run. The point is an agent that works beside them: it reads the shelf, points to the work, and verifies it as it happens. That’s what collaboration in the field should look like.

See it in action: