Awareness Avenue, n8n Pipeline

Reads the collection row from Airtable, then runs four Perplexity deep research passes in parallel: competitor landscape, material science, customer sentiment, and Google Shopping keywords.

The four results merge, then Gemini writes two documents from them: a brand identity doc and a set of GMC title directives. Both are created in Google Drive, written as Google Docs, and their IDs are written back to the Airtable row.

29 nodes. The output is a reusable brief, not a one-off answer.

Picks up where the first left off. Reads the ready collections from Airtable and loops one product at a time.

Images come from either source: if the row points at Canva, it triggers a Canva export, polls the export status and waits for the assets; if it points at a Drive folder, it lists the images there instead.

It then pulls the brand doc and the GMC directives back out of Google Docs, consolidates everything, and hands it to Gemini for the product copy.

Before anything is created, the workflow builds a title query, checks the existing Shopify PDP titles, and resolves a unique one.

This is the node people skip. Two products with the same title is a support problem later and a Merchant Center problem immediately.

Builds the Shopify payload, creates the product, waits for it to settle, then assigns the image to every variant rather than only the first.

It then ensures the variant set is complete, determines the Google product category and sets it on the product. The Airtable row is updated with the result, so a rerun knows what is already done.

38 nodes, including its own retry and wait handling, because the Canva export and the Shopify create are both slower than the rest of the flow.