Skip to main content
The example to try first. It needs two lookups joined — what people are waiting for, and what is actually in stock right now — which is exactly the thing a dashboard cannot do for you. Saved prompt: find_best_opportunities

​
The prompt

You
Screenshot to add — images/c1a-prompt.png: the prompt typed into the chat, with the agent beginning to work on it.

​
What the AI does

  1. Calls get_analytics with the most_wanted report for the variants with the most demand.
  2. Takes the top variants and calls shopify_product_lookup with inventory included, reading live stock per location straight from Shopify.
  3. Compares the two:
    • Long queue + stock available somewhere → revenue sitting still. Notify these now.
    • Long queue + no stock anywhere → your restock list, ranked by demand.
Screenshot to add — images/c1b-tool-calls.png: the two tool calls in sequence — get_analytics for most_wanted, then shopify_product_lookup with live inventory.
Screenshot to add — images/c1c-result.png: the finished answer: a ranked list of variants to notify now versus restock.

​
Why it beats the report page

The waiting-list report knows demand but not current stock. The Shopify inventory screen knows stock but not who’s waiting. Joining them by hand means two tabs and a spreadsheet; here it’s one question, and the ranking comes with it.

​
Good follow-ups

Stay in the same chat — the agent already has the context:
→ get_entity_details returns that variant’s sign-ups (up to 100).
→ send_test_notification, if your token allows change tools. It sends a real message and uses your allowance, and it only goes to your store’s own addresses. The AI previews it first and sends only after you confirm.
→ export_analytics_csv returns a download link, valid for 24 hours. The most-wanted list itself can’t be exported through the connector — download it from the app’s Analytics page.