Create a copy-paste AI prompt library for a [PROCUREMENT TEAM] working in [INDUSTRY]. The library should cover their recurring work without treating AI output as a sourcing decision, legal advice, or authority to negotiate or approve spend. Team responsibilities and current process: [TEAM CONTEXT] Categories and typical purchases: [CATEGORIES] Policies, approval limits, and required controls: [POLICIES] Available data and document types: [AVAILABLE INPUTS] Produce exactly 10 prompts, each 120–180 words, for these tasks: intake clarification, requirements definition, supplier longlist research plan, RFP question set, bid-comparison matrix, total-cost-of-ownership analysis, negotiation preparation, contract red-flag checklist, supplier performance review, and stakeholder update. Each prompt must be written for a procurement professional to paste into ChatGPT or Claude. Include only the square-bracket placeholders needed for that task, a precise requested output, and instructions to separate facts, calculations, assumptions, and recommendations. For sourcing and supplier comparison, require objective criteria, weighted scoring only when weights are supplied, and a record of missing evidence. For contracts, require the model to identify review issues for legal or procurement review, not interpret law conclusively. For cost analysis, require currency, period, volume assumptions, one-time costs, recurring costs, transition costs, and sensitivity items. After the 10 prompts, add a one-page “Safe Use Rules” checklist covering confidential data, human approvals, supplier conflicts, audit trail, and verification of external claims. Before answering, check that no prompt asks AI to invent supplier data, pricing, contract terms, or policy requirements. Flag tasks that should go to legal, finance, security, or an authorized approver. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PROCUREMENT TEAM] | Enter the team’s size, roles, regions, and level of procurement maturity. | Five-person indirect procurement team serving the United States and Canada |
| [INDUSTRY] | Enter the industry and any regulated or operational context that matters. | Business software company handling customer account data |
| [TEAM CONTEXT] | Describe the intake, sourcing, contracting, and supplier-management process used today. | Business owners submit requests in Jira; procurement runs sourcing above $25,000; legal reviews nonstandard paper; finance approves budget before purchase orders. |
| [CATEGORIES] | List common spend categories and examples of typical purchases. | SaaS, recruiting agencies, marketing services, laptops, facilities maintenance, and professional services |
| [POLICIES] | Paste relevant approval limits, competition rules, security reviews, and recordkeeping requirements. | Two competitive bids for purchases above $25,000 unless a sole-source exception is approved; security review for vendors handling customer data; retain bid records for three years. |
| [AVAILABLE INPUTS] | List the data and documents the team can safely provide to an AI tool. | Approved intake forms, redacted proposals, supplier scorecards, spend exports, contract templates, and internal policy excerpts |
How to use
- Replace policy placeholders with your current rules before sharing prompts with the team.
- Test the bid-comparison and total-cost prompts against a completed purchase to find missing columns or assumptions.
- Keep supplier names, pricing, and customer data out of public AI tools unless your approved tool and data policy permit them.
- Follow up with: “Adapt prompts 4–6 for facilities services purchases with multi-site labor and insurance requirements.”
Variations
RFP builder
Use when an approved requirement needs a structured request for proposal.
Create an RFP outline for [PURCHASE] using [REQUIREMENTS], [EVALUATION CRITERIA], [TIMELINE], and [POLICY RULES]. Include supplier instructions, scope, mandatory requirements, response format, pricing schedule, implementation questions, evaluation process, and submission dates. Separate mandatory from desirable requirements. Do not set scoring weights unless supplied. Flag requirements that are ambiguous, favor one supplier without justification, or conflict with [POLICY RULES].
Bid comparison
Use when supplier proposals are ready for an evidence-based comparison.
Build a bid-comparison matrix for [PURCHASE] from [SUPPLIER RESPONSES], [CRITERIA], and [VOLUME ASSUMPTIONS]. Return a table with evidence, gaps, commercial terms, implementation dependencies, total-cost inputs, and risks for each supplier. Calculate only from supplied figures and show each formula. Do not infer compliance from silence. End with questions to resolve before recommendation and flag any criterion that requires legal, security, or finance review.
Negotiation brief
Use before a supplier negotiation meeting.
Prepare a negotiation brief for [SUPPLIER] based on [CURRENT OFFER], [MARKET EVIDENCE], [BUSINESS NEEDS], and [APPROVAL LIMITS]. Produce objectives, walk-away conditions, prioritized asks, tradeable items, questions, and a meeting agenda. Separate verified leverage from assumptions; do not claim competing offers or authority I do not have. Flag commercial changes that require legal, finance, security, or executive approval before agreement.
Tips
- Make an intake prompt demand a measurable business need, expected volume, budget owner, and decision date before it generates an RFP.
- Compare bids on normalized units and contract periods; headline annual prices often omit implementation, usage, renewal, or exit costs.
- Keep evaluation notes tied to published criteria so the decision record is defensible and fair to suppliers.
- Use AI to expose missing evidence, then obtain the evidence from the supplier or internal owner rather than filling the gap with a guess.
FAQ
Can AI choose a supplier for us?
It can organize evidence and identify gaps, but authorized people must make the recommendation and approval under your policies.
Can we paste supplier proposals into an AI tool?
Only if your data-handling rules and the approved tool permit it. Redact confidential pricing, personal data, and protected contract information when required.
Can AI review a contract?
It can create a practical issue checklist, but legal counsel should review legal meaning, risk allocation, and nonstandard terms.