Skip to content
KiweelyKiweelyKiweely

Prompt Patterns That Stop Claude Code Inventing Workday AMD Attributes

Good Extend prompts require the agent to read existing components or Developer Agent tools before writing AMD/PMD JSON—and to validate before deploy.

  • By Kiweely
  • Updated
  • 1 min read
On this page

Generic models invent plausible AMD keys. Prompts have to make invention more expensive than lookup.

Patterns that work better

  1. “Open the existing business object component and list current fields before proposing changes.”
  2. “Use Developer Agent / WDCLI tools; do not freehand schema.” (when entitled)
  3. “Quote the Workday component reference you relied on; if none, stop.”
  4. “Run Validate; paste errors; do not deploy on red.”
  5. “After deploy, read back via API/UI as persona X.”

Patterns that fail

  • “Generate a complete peer recognition app” with empty context.
  • “Fix it” after a validation wall of text with no component attached.
  • Mixing two clients’ field names in one conversation.

Workday’s own Developer Agent positioning exists because general tools lack Workday domain precision until connected—build that fact into the prompt, do not fight it.

Your Workday work, done by AI.

Kiweely Vine is the AI agent that builds, deploys and tests Extend apps as real people, makes changes in any Workday task screen, answers from your real tenant and builds reports.

Get the next one in your inbox.

One email a month for Workday teams: what’s changing in Workday and what to retest, one tip tried in a real tenant, and what’s new at Kiweely.

We’ll email you once to confirm. Unsubscribe in one click from any issue.