The gap between "ChatGPT doesn't work for me" and "ChatGPT saves me two hours a week" is usually one thing: specificity. Generic input produces generic output. If you walk up to a consultant and say "help me with procurement," you'll get a blank stare. Give them context, a role, and a specific ask — and they deliver something useful. Prompts work the same way.
Every prompt below follows the same logic: assign a role, give context, specify the output. Copy them, adapt the bracketed parts to your situation, and use them. The explanations are there to help you understand when each one is worth reaching for.
How to use ChatGPT in procurement: where it actually helps
Most searches for "ChatGPT for procurement" come from people who tried it once, got a bland answer, and are wondering whether it is worth the effort. It is, but only for the right jobs. ChatGPT is a drafting and thinking partner. It is not a database, a price index, or a decision-maker. Use it where a smart colleague with no access to your systems would help, and nowhere else.
- Drafting: RFQ emails, supplier follow-ups, escalation letters, award and rejection notices. First drafts in seconds, your edits on top (prompts 2 and 6).
- Structuring decisions: supplier evaluation scorecards, negotiation positions with fallbacks, consolidation options (prompts 1, 4 and 10).
- Reading: turning a dense contract clause or policy paragraph into plain language so you know what to ask Legal (prompts 3 and 7).
- Preparation: questions for a supplier review, a market briefing before a sourcing event, a cost breakdown to challenge a price increase (prompts 5, 8 and 9).
Where it does not belong: anything that depends on live or verified facts. ChatGPT does not know your supplier's current price, their financial health, or whether a regulation changed last month, and it will answer confidently anyway. Market data, supplier due diligence and final contract wording need a source you can check.
The ten prompts
Supplier evaluation scorecard
When to use: you've received a proposal and need a structured assessment before the decision goes up. Give the AI a clear set of criteria and the proposal text.
You are a procurement analyst. Evaluate this supplier based on the following criteria: [list your criteria, e.g. price competitiveness, delivery reliability, payment terms flexibility, technical capability, references]. Here is the supplier's proposal: [paste anonymized proposal text]. Rate each criterion 1–5 and explain your reasoning in 2–3 sentences per criterion. End with an overall recommendation.What you get back: a structured scorecard with ratings and written rationale. Useful for internal alignment before a supplier review meeting, or as a first pass when evaluating several proposals simultaneously.
RFQ email draft
When to use: you need to send an RFQ and want a professional structure without spending 30 minutes formatting it from scratch.
Draft an RFQ email for [product or service description]. We are a mid-size manufacturer operating in EMEA. We need [quantity] by [delivery date]. The email should include sections for: technical requirements, delivery terms, payment conditions, and validity period. Tone should be professional but direct. Maximum 400 words.What you get back: a ready-to-edit RFQ email. You'll still need to add your specific technical specs and adjust the tone, but the structure is done and you won't start from a blank page.
Contract clause plain-language summary
When to use: legal sent you a 40-page contract and you need to understand what a specific section actually means before the review meeting. Works well on standard boilerplate sections.
Summarize the following contract section in plain language. Flag any clauses that could create financial exposure, limit our ability to terminate, or restrict price renegotiation. Explain each flagged clause in one sentence. Here is the contract section: [paste section text]What you get back: a plain-language summary and a short list of flags. Not a substitute for legal review on anything significant, but useful for pre-reading before you go into a meeting.
Spend category consolidation analysis
When to use: you're preparing a category review and want to identify consolidation opportunities from spend data. Use anonymized or aggregate numbers.
I have spend data for the category [category name, e.g. MRO, packaging, logistics]. Here are the top 10 suppliers and their annual spend in EUR: [paste table or list]. Identify consolidation opportunities based on this distribution. Suggest 3 strategic questions I should ask in the next supplier review. Note any spend concentration risks.What you get back: a quick read on spend spread and concentration, plus questions you can use directly in a supplier performance review. Most useful when you're new to a category or haven't done a proper spend analysis in a while.
Supplier performance review meeting prep
When to use: you have a QBR or performance review tomorrow and want to go in with specific questions rather than generic ones. This prompt keeps the conversation productive without making it adversarial.
I have a supplier performance review tomorrow with [supplier type, e.g. our main packaging supplier]. Their key metrics this quarter: [paste or describe, e.g. on-time delivery 78% vs 95% target, 3 quality incidents, invoice accuracy 91%]. Write 5 specific questions that address the underperformance areas directly without being confrontational. Each question should invite the supplier to explain root cause and propose a recovery plan.What you get back: five questions ready to use. You'll probably reword a few of them, but the structure is there. Saves you the prep time and keeps the meeting focused on action, not complaints.
Supplier email response draft
When to use: you've received a difficult supplier email — a price increase notification, a capacity constraint warning, a dispute over payment terms — and you need a response that's firm without burning the relationship.
I received this email from a supplier: [paste the email text]. Draft a professional response that [describe your goal, e.g. pushes back on the proposed 12% price increase, requests a detailed cost breakdown, and keeps the door open for a call this week]. Tone: firm but collaborative. Maximum 200 words.What you get back: a draft that handles the emotional framing so you don't have to. Still review it carefully — AI can occasionally soften something you actually need to keep sharp.
Policy interpretation check
When to use: you need to check whether a planned action is covered by your procurement policy, and the policy document is long enough that finding the relevant section takes time. Use internal documentation only in approved tools.
I need to check our procurement policy on [topic, e.g. single-source approvals above €50,000]. Here is the relevant policy text: [paste section]. What does it say about [specific situation, e.g. emergency procurement when a sole supplier is the only viable option]? List any approval requirements or exceptions explicitly mentioned.What you get back: a direct answer about what the policy actually says, with the relevant parts highlighted. Faster than reading the whole document, and useful when you need to brief a manager quickly on what approvals are required.
Cost breakdown negotiation analysis
When to use: a supplier has shared a detailed cost breakdown and you want to identify which line items are most open to challenge before you walk into the negotiation room.
A supplier sent this cost breakdown for [product/service]: [paste breakdown]. Identify which cost elements are most open to negotiation and explain why for each. Note any elements where the margin appears higher than typical for this type of supply. Suggest 2–3 specific challenges I can raise in the negotiation.What you get back: a structured view of where the money is going and which elements have room. The AI won't know your specific market benchmarks, so use its output as a starting point, not a final answer.
Supply market research summary
When to use: you're starting a sourcing project and need a quick orientation on the market before you design your RFQ. This is one of the few prompts where you don't need to paste in your own data — ChatGPT's training data is actually useful here.
Summarize the key supply market trends for [category, e.g. industrial packaging, MRO consumables, freight forwarding] in Europe in 2025–2026 that a procurement manager should know before a sourcing project. Focus on: supply risk factors, pricing pressure drivers, main supplier consolidation trends, and any regulatory changes affecting the category. Keep it to one paragraph per topic.What you get back: a useful market briefing that gives you enough context to ask better questions in supplier conversations. Verify anything time-sensitive with actual market sources — AI training data has a cutoff date.
Negotiation position framework
When to use: you have a significant negotiation coming up and want to map your positions before you sit down. Particularly useful when you're negotiating something outside your usual category.
I am about to negotiate a [type of contract, e.g. 2-year maintenance services agreement] with [type of supplier, e.g. a mid-size regional provider who is currently our sole source]. My goals: [list your goals, e.g. price freeze for year 1, 3% cap in year 2, improved SLA response times, quarterly performance reviews]. Their likely position: [describe what you expect, e.g. they want a 5% increase annually and longer payment terms]. Give me 3 opening positions and 3 fallback positions with a brief rationale for each.What you get back: a structured set of opening and fallback positions that you can walk into the room with. Most experienced negotiators will adjust these — that's expected. But having the framework written down before you go in is better than improvising.
ChatGPT, Copilot or a custom GPT?
The prompts work in any of them. The difference is data, not intelligence. Consumer ChatGPT is fine for anonymised drafting. ChatGPT Team or Enterprise plans do not use your business data for training by default, which makes them usable for real supplier data if your IT has approved them. Microsoft Copilot runs inside your Microsoft 365 tenant and can read your mail, files and meetings, which is why it is usually the easiest option to get approved in a manufacturing company. If you want the Copilot route, start with five Copilot workflows for procurement. If you want your team writing better prompts in any tool, read prompt engineering for procurement.
A note on how to use these
None of these prompts are magic. They produce a first draft, a structure, or a set of questions — not a finished deliverable. What they eliminate is the blank page problem. You get something to react to, edit, and improve instead of starting from nothing. For most procurement tasks, that's where the time goes.
The other thing worth saying: the output quality improves significantly when you iterate. If the first response misses what you needed, tell the AI what was wrong and ask it to try again with the correction. One round of feedback typically gets you somewhere useful.
Prompting is a skill, not a feature. The people getting the most out of AI in procurement are the ones who learned to give it context, not the ones with the best tool subscription.