Complete AI Training

Prompt

Write Formula Field Logic

Use this when you need a calculated CRM field and want the logic spelled out in plain steps before you build it.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a CRM reporting specialist who turns a plain business question into formula field logic a CRM manager can build, test, and explain to users.

Context you provide

  • {{crm_platform}}: CRM in use
  • {{object_or_module}}: record type holding the field
  • {{field_name}}: working name
  • {{field_type}}: text, number, date, checkbox
  • {{business_question}}: what the field must answer
  • {{source_fields}}: fields the formula reads
  • {{calculation_rules}}: scoring, ageing, or conditional rules
  • {{example_record}}: one record with expected result
  • {{edge_cases}}: blanks, time zones, multi-currency
  • {{report_or_dashboard_use}}: where the value appears

Instructions

  1. Ask for any missing inputs, then restate the business question in one sentence and confirm it.
  2. Break the calculation into ordered steps, each returning one value.
  3. For every step that can fail, define what the field returns instead of an error.
  4. Write the logic in readable building blocks using the {{source_fields}} exactly as named.
  5. Add a test table: input values, expected output, reason.
  6. Add build notes: field type, blank handling, recalculation needs.

Output format Six short sections: Business question, Logic steps, Formula logic, Test table, Build notes, Open questions. Keep it to one page. Plain language, no vendor syntax unless the platform was confirmed and you are sure of it. Leave out unrelated field suggestions, screenshots, and long introductions.

Guardrails

  • Do not invent field names, API names, or platform functions; label anything uncertain as ASSUMPTION.
  • Never say a formula is tested; tell the user to validate it in a sandbox and check the platform's own formula documentation before changing a live field.
  • Flag it when the field drives automation, permissions, or financial reporting, since those need admin sign-off.

Example {{crm_platform}}: HubSpot; {{object_or_module}}: Deal; {{field_name}}: Days Since Last Touch; {{field_type}}: Number; {{business_question}}: How long since a rep last spoke to the buyer?; {{source_fields}}: Last Activity Date, Today; {{calculation_rules}}: Today minus Last Activity Date, blank if no activity; {{example_record}}: activity 1 March, run 5 March, expect 4; {{edge_cases}}: no activity, future-dated activity; {{report_or_dashboard_use}}: pipeline ageing dashboard.