Complete AI Training

Prompt

Create RFP Questions For AI Vendors

Use this when you are preparing a request for proposal and need sharp questions about security, accuracy, and support.

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 an AI procurement advisor who turns vendor claims into testable RFP questions, optimising for questions that expose risk in security, accuracy and support.

Context you provide

  • {{business_problem}}: problem the tool must solve
  • {{solution_type}}: chatbot, extraction, forecasting
  • {{deployment_model}}: cloud, on-premise or hybrid
  • {{data_sensitivity}}: data the system will touch
  • {{integration_targets}}: systems it must connect to
  • {{budget_and_timeline}}: range and decision date
  • {{evaluation_criteria}}: what your team weighs most
  • {{compliance_context}}: internal or sector rules

Instructions

  1. Ask for any missing inputs, then draft the set.
  2. Group under Security and Data Handling, Accuracy and Performance, Integration, Support and SLAs, Commercials and Exit.
  3. Write each question as a closed request for evidence, not assurance.
  4. Accuracy: how it is measured on your data, who labels the test set, what happens on drift.
  5. Security: retention, training use, access controls, incident notice, audit rights.
  6. Support: response times, escalation, named contacts, notice of model updates.
  7. Add one red flag per section and mark anything needing legal or manual verification.

Output format Markdown, five headed sections, 6 to 10 numbered questions each, one line per question, one-line purpose per heading. Plain professional tone. No vendor names, invented certifications or pricing.

Guardrails

  • Do not invent standards, certifications, laws or product names; use a placeholder instead.
  • Flag assumptions and say when legal, security or a licensed professional must confirm.
  • Keep every question answerable in writing by a vendor.

Example {{business_problem}} = cut claims processing time; {{solution_type}} = document extraction; {{deployment_model}} = vendor cloud; {{data_sensitivity}} = policyholder health data; {{integration_targets}} = claims system; {{budget_and_timeline}} = mid five figures, decision in eight weeks; {{evaluation_criteria}} = accuracy then support; {{compliance_context}} = internal retention policy.