Spot catalog gaps from Instagram, Pinterest and competitors with Shopify and GPT-4o

Quick overview

This workflow runs every Monday to collect fashion signals from Instagram and Pinterest via Apify plus competitor product feeds, compares them to your Shopify catalog and historical trends in Google Sheets, and uses OpenAI to cluster trends and generate gap-focused launch briefs that it posts to Slack.

How it works

1. Runs every Monday at 7 AM IST and loads configuration values such as brand focus, batch sizes, and source actors.
2. Reads trend history, the design backlog, your active Shopify product catalog, and the active trend sources list from Google Sheets.
3. For each active source, fetches recent Instagram hashtag posts and Pinterest search pins via Apify and pulls competitor product lists from the configured URLs.
4. Combines all collected items, filters out weak or stale entries, de-duplicates repeated items, and posts a Slack alert if too few usable items are found.
5. Sends the remaining items to OpenAI (gpt-4o-mini) to extract structured fashion tags and a canonical style key, then scores and ranks the strongest style keys across all sources.
6. Uses OpenAI (gpt-4o-mini) to cluster the top style keys into named trends, reusing prior cluster names where possible, and calculates coverage against your Shopify catalog plus momentum versus previous weeks.
7. Appends this week’s trend metrics to the Trend History sheet, generates OpenAI-written product launch briefs for the highest-priority uncovered trends, posts a block-formatted report to Slack, and appends new ideas to the Design Backlog sheet while avoiding recently-added repeats.

Setup

1. Add credentials for OpenAI, Slack, and Google Sheets, and set the target Slack channel ID used by the report and low-volume alert.
2. Create a Shopify custom app with product read access, store its Admin API token as an HTTP Header Auth credential, and replace the REPLACE_STORE.myshopify.com value in the configuration.
3. Add an Apify API token as an HTTP Header Auth credential (Authorization: Bearer <token>) and set the Pinterest actor ID in the configuration if you don’t use the default.
4. Create a Google Sheets file with Trend Sources, Trend History, and Design Backlog tabs (with columns matching the workflow’s appended fields) and replace each REPLACE_WITH_GOOGLE_SHEET_ID placeholder.
5. Populate the Trend Sources sheet with source_type (instagram/pinterest/competitor), query, optional label and weight, and an active flag, then run the workflow once to validate the Slack output and sheet writes.