Turn a plain-English workflow into an automation recipe
A scenario prompt that converts a described VIP-customer workflow into an implementation-ready trigger/condition/action recipe, resolving the ambiguities the plain description leaves open.
Works with Claude / GPT960 uses★ 4.5
The prompt
workflow-to-automation-recipe
A store owner describes their process like this: "When someone spends over $150 in
one order, I want to tag them as a VIP in our email tool, send them a thank-you email
with a discount code for next time, and post a note in our team Slack so support
knows to prioritize them if they reach out."
AVAILABLE TOOLS: [WooCommerce (trigger), Uncanny Automator, Klaviyo (email/tagging),
Slack]
Turn this into an implementation-ready automation recipe, thinking through what the
plain description leaves ambiguous:
1. Trigger: the exact WooCommerce event to fire on. "Spends over $150 in one order"
needs to be order total after discounts, not before, and you should specify which
order status should fire it (pending vs. processing vs. completed), since firing
on "pending" risks tagging abandoned or failed orders.
2. Conditions: state the exact conditional logic, including the edge case of a
partial refund that drops the order below $150 after the automation already
fired. Should the VIP tag be reversed? Decide and state your reasoning.
3. Actions in order, each mapped to a specific tool: Klaviyo tag, then Klaviyo email
send (sequence matters if the email template pulls the tag), then a Slack message
with the customer's name and order total.
4. Failure handling: if the Klaviyo tagging step fails, should the email and Slack
steps still fire, or should the whole recipe halt? Recommend one and say why.
5. One edge case specific to this exact workflow that the plain description doesn't
address, such as a customer who crosses $150 through two separate same-day
orders.
Replace the workflow description and available tools with your own.
Most automation-tool UIs force you to think in trigger/action blocks from the start,
which loses the actual business logic buried in edge cases like the refund scenario
above. Writing the workflow in plain English first and having this surface the gaps is
faster than discovering them after the recipe misfires on a real customer. See
Uncanny Automator for the most common no-code recipe
builder this maps directly onto.
Pro tip: always ask for the failure-handling step explicitly. It’s the part every
manually-built automation recipe forgets, and the one that causes silent data loss
later.