AI Ads

How do you swap a product in an existing ad without re-shooting it?

Last updated August 1, 2026

Upload the existing ad to an AI agent as a reference video: it detects scene breaks, product changes, and lighting shifts, splits the ad into segments, and you drop in reference images of the new product per segment. Pick and lock the best regenerated version of each segment, then stitch. Documented cost: ~$30 and 30 minutes per swap once the workflow is established.

Start by uploading your winning ad as a reference video and letting the AI analyze it before generating anything. invideo is an agentic video creation tool with the current video and image models built in, and the invideo agent runs this workflow end to end. In one documented run, the invideo agent watched the uploaded ad, figured out that the model wore one outfit in the first half and a second outfit in the second half, and split the video into two clips on its own — zero manual editing steps to segment the ad.

Step 1 — Load brand context once. Feed the invideo agent your visual guidelines and product catalog before the swap so every regenerated segment holds your brand's look without re-prompting.

Step 2 — Upload the ad as a reference video. The invideo agent detects every scene break, outfit or product change, and lighting shift, and splits the ad into segments automatically.

Step 3 — Drop in the new product. Upload reference images of the replacement product for each segment — the documented workflow used 4 angles per new outfit (front, side, back, and a fabric close-up). For packaged products, include a shot of the product held in a hand so the model reads true scale.

Step 4 — Generate, pick, lock. For each segment, the invideo agent generates options of the new product in the same scene, preserving the original camera work, pacing, and framing. Pick the best version, lock it, and repeat for the next segment. Generating multiple options and locking the best is the selection method the documented run used — the output was judged visually comparable to the original shot ad in side-by-side review.

Step 5 — Stitch inside invideo. Ask the invideo agent to stitch the locked clips, then trim timing and audio in Slate, invideo's built-in timeline, so the whole swap happens in one platform.

Step 6 — Repeat across the catalog. The workflow persists: upload new products and the invideo agent reruns the same sequence. The first swap took about 2 hours in the documented production; every subsequent swap took about 30 minutes at 115 credits ($30), supporting up to 12 product-swap ads in an 8-hour day — and 8–10 variations of one winning ad per day when combined with market localizations (which run ~5x the credits per ad). One winning ad can be replicated across an entire product catalog without shooting again.

Beyond the swap itself: the traditional route — masking the product frame by frame and compositing a replacement in an editor — still works for a single shot but doesn't scale across segments or catalogs. And when you ship the swapped version to Meta, upload it as a new ad rather than editing the live one, so you don't disturb the original's optimization.

Watch some of these to see what works for you:

See the invideo agent run a full product-swap across an existing winning ad
Watch the invideo agent analyze a reference ad and regenerate it shot by shot

getting the first swap to take place took me about two hours. But once I had the first product swap in place and the workflow in place, every new product swap took me only about 30 minutes.

— invideo's creative team

Share

More on AI Ads