Meta's new shopping agent is recommending products from whatever it can scrape off your site, so the optimiser's job is to stop it guessing and hand it the facts directly.
Yesterday Marpipe.com announced a partnership with Meta on a new initiative called Open Catalogs. It's the tool I use for image optimisation in Meta feeds, and one I recommend to all brands with 50+ SKUs investing in Meta carousel ads.
The context. Meta launched Muse, its personal AI agent, in the US in September. According to Marpipe, shopping has become its main use case: millions of shoppers skipping Google and going straight to Muse to find products, with an average order value around $200. Their figures, not mine, but the direction of travel is clear.
The problem. Muse currently learns about your products by crawling the web. It reads what it can find and fills the gaps itself. Your stock levels, live offers and the reasons someone should buy are all second-hand.
The fix. Open Catalogs lets you pipe real-time product data straight to Muse: availability, use cases, active discounts. In Marpipe.com that's one toggle, the Meta Agentic Catalog. Meta says brands that switch it on will see better visibility in Muse.
The optimisation. This is feed optimisation with a new reader. An agent can only recommend what it understands. So when you flip the switch:
Match your feed to your product pages. Price, stock, variants. An agent flags a mismatch, it doesn't forgive it.
Write use cases into your product data. Who it's for and what problem it solves is exactly what the agent is matching against.
Keep discounts in the feed, not just on site banners.
Toggle it on. Then treat the catalog as a channel you optimise every week, not a file you upload once.
Ian