There are roughly 200 AI prompts we use with procurement clients on any given week. We've published all of them, twice, once for teams running on Claude, once for teams running on OpenAI Codex/ChatGPT, and grouped them by role so each person on a procurement function can find the prompts that apply to their actual work in under 30 seconds.

This piece is the map. It explains why we organized our prompt library by role instead of by workflow, what's in each role guide, and how to run any of these prompts well enough that the output is something you can ship rather than something you have to rewrite from scratch.

If you want to skip the map and go straight to the downloads, scroll to the bottom. Both libraries are free.

Why generic AI prompts fail in procurement

Most prompt engineering content teaches you the same three things: specify the format, give examples, break complex tasks into steps. All true. All useless if you're a procurement professional drafting an RFP or scoring suppliers, because none of them address the actual reason your AI output keeps coming back as glossy generic text that wouldn't survive a stakeholder review.

The reason is context, and procurement is the most context-heavy function in most organizations. A category manager's "draft an RFP" is not the same as a marketing lead's "draft an RFP." Your category, your supplier base, your contract template, your evaluation rubric, your code of conduct, your internal stakeholders, your regulatory environment, all of that is context the model needs before it can produce something usable. Generic prompt training teaches you syntax. It doesn't teach you how to package procurement context into a prompt so the output sounds like it came from someone who actually works on your team.

Three structural problems make procurement different:

The work is sequential and collaborative. A sourcing lead's RFP draft becomes a category manager's evaluation rubric becomes a contract manager's redline. If each person uses AI in a different way, the output quality degrades down the chain. The fix is a shared prompt library the whole team uses, not 30 individual notebooks of clever prompts that no one else can find.

The work is governance-sensitive. Pasting a supplier name into the wrong AI tool can violate an MNDA. Pasting contract terms into a consumer-tier AI can void enterprise data handling protections. Prompts have to be written assuming the right data classification and the right workspace tooling. Generic prompt examples ignore this entirely.

The work has a high cost of bad output. A marketing email AI got slightly wrong is annoying. A supplier scorecard that hallucinates a financial metric is a real problem. Prompts have to be written to make hallucination harder, to demand sourcing, and to flag uncertainty rather than mask it.

Once you've internalized those three things, "good prompts for procurement" becomes a recognizable shape, not a single magic template, but a pattern. The prompt libraries we've published are 200+ instances of that pattern across the workflows procurement teams actually run.

Why we organize by role, not by workflow

There are two ways to organize a procurement prompt library. By workflow (RFP, contract review, spend analysis, supplier scoring, etc.) or by role (sourcing lead, category manager, contract manager, etc.). Both have published examples. We picked role-based for both our playbooks, and the reasoning is worth a paragraph because it changes how you'd use the library day to day.

Procurement work is role-anchored more than workflow-anchored. A category manager doesn't sit down on Monday morning and ask, "what workflow am I doing today?" They sit down and look at the categories they own. The work flows out of the role. A workflow-cut library makes a category manager hunt through 7 different sections to find the 15 prompts that apply to them. A role-cut library hands them one PDF that covers their whole job.

There's a secondary benefit. A role-cut library makes it easier to share. The contract manager on your team gets the Contract Manager guide. The sourcing lead gets the Sourcing Lead guide. Nobody has to read prompts that don't apply to them. Adoption goes up because the friction comes down.

(The workflow cut still has its place, it's the right organization for a single individual learning a new skill area. If you're a tactical buyer learning category strategy, a workflow library is easier to read end-to-end. We have an unpublished workflow-cut draft we use internally for that case. But the role cut is the right primary library for a team.)

The 7 procurement role guides, for both Claude and ChatGPT/Codex

What follows is the role-by-role map. For each role: a 1-paragraph framing of what the role does and where AI helps most, a list of the prompts we've packaged for that role, and a featured prompt published in full so the page has real prompts a reader can use right now without filling out a form.

For each role we link to two free downloads: the Claude version (from the Claude Cowork Playbook for Procurement Teams) and the OpenAI version (from The Codex Playbook for Procurement Teams). Same workflows, prompts tuned for each tool's strengths.

Sourcing Leads, the RFP, evaluation, and negotiation cycle

If your job is running sourcing events, you spend most of your week on a small number of high-leverage workflows: drafting RFPs and RFIs, scoring supplier responses, building negotiation strategies, debriefing calls while the details are still fresh, summarizing shortlists for steering groups, and drafting the award memo at the end. Every one of those is a high-context, high-stakes deliverable that benefits from AI as a structured first draft.

The 15 prompts in the Sourcing Lead guide: Supplier Comparison Matrix, RFP First Draft, RFP Evaluation Scorecard, Negotiation Strategy Builder, Negotiation Debrief, Should-Cost Model, RFI for Market Sounding, Supplier Due Diligence Report, Bid Clarification Questions, Contract Redline Summary, Supplier Shortlist Recommendation, Sourcing Pipeline Status Report, Award Justification Memo, Post-Award Transition Plan, Supplier Communication Templates.

Featured prompt, RFP First Draft:

You are a senior procurement sourcing specialist drafting an RFP for a competitive tender. We are a [industry] company running a formal sourcing process for [category/service]. The scope is: [2-3 sentence description]. Estimated annual contract value is $[amount]. Key requirements include: [list 3-5 must-haves]. Draft a complete RFP document with these sections in sequence: 1. Introduction and company overview 2. Scope of work and detailed requirements (translate the must-haves into specific, measurable requirements a supplier can price against) 3. Technical approach (what we want suppliers to describe about their methodology, tools, and team) 4. Pricing structure (provide a pricing table format that forces like-for- like comparison) 5. Implementation timeline (with key milestones) 6. Evaluation criteria and weightings 7. References (request at least two of comparable scope) 8. Terms and conditions (standard commercial terms) 9. Submission instructions (format, deadline, contact for questions) Use a formal tone suitable for [type of supplier]. Ensure every section gives suppliers enough detail to respond specifically. Flag any areas where I need to provide additional internal detail before issuing.

Downloads: Sourcing Lead, Claude version · Sourcing Lead, Codex version

Category Managers, strategy, market intel, supplier landscape

Category managers run the most research-heavy procurement role and benefit the most from AI as a research compressor. The work is structured around recurring cycles: refresh the category strategy, update the supplier scorecards, prep the next category council deck, monitor the supplier landscape, build the savings business case. Each of those is a multi-hour-to-multi-day exercise that AI can compress to under an hour when the prompts are right.

The 15 prompts in the Category Manager guide: Supplier Research Brief, Category Strategy Draft, Supply Market Intelligence, Kraljic Portfolio Mapping, Supplier Scorecard, Quarterly Business Review Prep, Stakeholder Requirements Brief, Supplier Relationship Health Check, Category Risk Assessment, Savings Business Case, Supplier Innovation Review, Market Price Index Tracker, Stakeholder Update Email, Buy Analysis, Supplier Consolidation Plan.

Featured prompt, Category Strategy Draft:

You are a senior category manager building a refreshed category strategy for [Category] for [Year]. Annual spend in this category is approximately $[amount] across [N] suppliers. Our top 3 suppliers are [list]. Our strategic priorities for this category are: [list 2-3]. Draft a 3-page category strategy document with these sections: 1. Category at a glance (spend, suppliers, business unit footprint) 2. Market context (5 bullets on the state of the supply market for this category) 3. Strategic positioning (where the category sits on a Kraljic matrix and why) 4. Strategic objectives (3 to 5 measurable objectives for the next 12 months) 5. Initiatives (the 3 to 5 initiatives that ladder up to the objectives, with owner, timing, and resourcing) 6. Risks (top 3 risks and our mitigation approach) 7. Decisions needed from category council Tone: senior procurement to executive audience. Lead with conclusions, support with data, recommend action. Maximum 3 pages.

Downloads: Category Manager, Claude version · Category Manager, Codex version

Contract Managers, clauses, risk, renewals, redlines

Contract work is the most rules-heavy procurement role and the one where AI's reliability matters most. The work is data-dense: read a long document, extract specific fields, compare against a template, flag what's missing or unusual. AI compresses this from hours to minutes when the prompts ask for specific extraction with quoted sources, not narrative summaries.

The 15 prompts in the Contract Manager guide: Contract Clause Extraction, Contract Risk Flag Review, Renewal Pipeline Summary, Redline Summary, Contract Amendment Draft, Termination Notice Draft, Obligation Extraction, NDA Review, Supplier Compliance Certificate Request, Contract Performance Summary, Contract Counterparty Research, Liquidated Damages Exposure Model, Post-Award Contract Briefing, SLA Breach Letter, Contract Handover Pack.

Featured prompt, Contract Risk Flag Review:

You are a commercial contracts analyst conducting a risk review. Read the attached contract and produce a structured risk register covering: 1. Termination provisions (notice required, conditions, financial consequences) 2. Liability and indemnification (cap amount, exclusions, who indemnifies whom for what) 3. Auto-renewal and notice periods (any auto-renewal clauses, exact notice windows, where they live in the document) 4. Change-of-control consent rights 5. Intellectual property ownership and license grants 6. Data protection and confidentiality obligations 7. Audit rights 8. Force majeure 9. Insurance requirements 10. Subcontracting restrictions For each, quote the exact contract language with the section/page number, assess the risk level (low/medium/high), and recommend a counter-position if the risk is medium or high. Do not invent terms. If a provision is missing entirely, flag it as "NOT PRESENT, recommend adding standard language."

Downloads: Contract Manager, Claude version · Contract Manager, Codex version

Procurement Analysts, spend, data, dashboards, audit

Analysts spend most of their week with data: refreshing the spend cube, cleaning supplier master data, building reports for category managers, reconciling invoices, preparing for audits. AI compresses the data wrangling so analysts spend more time on the questions and less on the prep. The role has the highest leverage from prompts that handle structured data well, which is where ChatGPT/Codex's Python and data analysis tooling tends to win.

The 15 prompts in the Procurement Analyst guide: Spend Category Summary, Tail Spend Identifier, Price Variance Analysis, Contract Clause Extraction, Savings Tracker Update, Supplier Master Data Cleanup, PO Compliance Check, Invoice Reconciliation Report, Data Quality Audit, Weekly Procurement Dashboard Summary, Actuals Variance Report, Supplier Onboarding Checklist Generator, Ad-hoc Spend Query Response, Contract Register Maintenance, Audit Preparation Pack.

Featured prompt, Tail Spend Identifier:

You are a procurement analyst identifying tail spend opportunities. Attached: our spend extract for the last 12 months (vendor name, category, amount, business unit). Tail spend definition for our function: any supplier with annual spend under $[threshold]. Identify: 1. All suppliers below the threshold (the long tail) 2. Total tail spend amount and percentage of total 3. Tail suppliers grouped by category, which categories have the most tail concentration? 4. Top 10 tail suppliers by spend (the ones closest to consolidation candidates) 5. Specific consolidation hypotheses: which tail suppliers could plausibly move under existing strategic supplier contracts? For each consolidation hypothesis, name the strategic supplier they could move to, estimate the addressable spend, and flag any obvious risk (geographic coverage, specialization that would be lost, etc.). Output as a memo with an executive summary at the top and supporting tables. Suitable for sharing with the category lead for that area.

Downloads: Procurement Analyst, Claude version · Procurement Analyst, Codex version

Supplier Relationship Managers, QBRs, scorecards, escalations

SRMs sit at the intersection of procurement and the supplier itself. The work is relationship-heavy and context-rich: prepping for QBRs, scoring supplier performance, drafting corrective action plans when something goes wrong, escalating issues to the right level. AI compresses the prep work so SRMs can spend more time in the actual relationship and less time assembling the materials.

The 15 prompts in the SRM guide: QBR Preparation Pack, Supplier Performance Scorecard, Corrective Action Plan Request, Relationship Health Assessment, Executive Sponsor Briefing Note, Issue Escalation Brief, Annual Supplier Review, Joint Business Plan Outline, Supplier Innovation Request, Supplier Development Action Plan, At-Risk Supplier Intervention Plan, Supplier Segmentation Review, SRM Meeting Minutes, Supplier Satisfaction Survey, Preferred Supplier Justification.

Featured prompt, QBR Preparation Pack:

You are an SRM preparing for an upcoming quarterly business review with [Supplier Name]. Annual contract value is $[amount]. Our last QBR was [date] and the previous quarter's performance against our SLAs was: [paste SLA performance summary]. Build a QBR preparation pack with: 1. Performance review (their delivery against each SLA, what changed quarter-over-quarter, anything trending the wrong way) 2. Outstanding items from last QBR (status of every commitment from both sides) 3. New items to raise this quarter (top 3 to 5, prioritized) 4. Asks of them (what we need from this supplier in the next 90 days) 5. Anticipated asks from them (what they're likely to push for and our response) 6. Agenda for the QBR (with time allocations, owners, expected outcomes) 7. Stakeholders to brief beforehand internally (and what to brief them on) Tone: factual and direct. SRM-to-supplier-account-team. No softening language on performance gaps.

Downloads: SRM, Claude version · SRM, Codex version

Tactical Buyers, POs, expediting, three-quote, daily ops

Tactical buyers handle the highest-volume, lowest-individual-value workflows in procurement: requisitions, purchase orders, three-quote comparisons, delivery follow-ups, invoice queries, end-of-month open PO reviews. The gains from AI here are in the small repeated tasks that consume the day, not in any single high-leverage exercise.

The 15 prompts in the Tactical Buyer guide: Purchase Order Confirmation Email, Three-Quote Comparison, Preferred Supplier Check, Price Benchmark Check, Delivery Delay Escalation Email, Invoice Query to Supplier, Emergency Purchase Justification, Approval Pack for Manager, Goods Receipt Discrepancy Note, End-of-Month Open PO Review, Supplier Acknowledgement Follow-up, Purchase Request Triage Note, Catalogue Item Description, Monthly Spend Summary for Manager, Spend Consolidation Opportunity Brief.

Featured prompt, Three-Quote Comparison:

You are a tactical buyer reviewing three quotes for [item or service] with budget approximately $[amount]. Required by [date]. Attached: three supplier quotes. Build a comparison covering: 1. Headline total for each quote (apples-to-apples; flag if a supplier excluded something the others included) 2. Per-unit pricing comparison 3. Delivery terms and dates (and which quote is fastest, which is slowest) 4. Payment terms (and which is most favorable to us) 5. Warranty and support (and the differences) 6. Any contractual flags or unusual terms Recommend a winner and explain why in 2 to 3 sentences. If the lowest price is not the recommended winner, explain the total cost reasoning. Output as a 1-page memo suitable for manager approval.

Downloads: Tactical Buyer, Claude version · Tactical Buyer, Codex version

Procurement Executives, board updates, exec briefings, strategy

The executive role doesn't benefit from prompts in the same way the operational roles do, there's no recurring deliverable that AI replaces. But the work that wraps the strategic decision-making, drafting the board update, summarizing the supplier risk landscape, prepping the CEO on a key supplier situation, reviewing the prior quarter's category outcomes, all benefits from AI as a structured first draft generator. The Executive Briefing in our Claude Cowork playbook covers this; the Codex equivalent does as well.

Downloads: Procurement Executive Briefing, Claude version · Codex Playbook, for procurement leaders

How to run these prompts effectively

Three principles run through every prompt in both libraries. If you internalize these, your output quality will improve more than any specific prompt below.

Context beats cleverness. Every prompt in the libraries assumes you attach real files, your category taxonomy, your supplier list, your contract template, your previous RFP. The prompt is the steering wheel. The attached context is the engine. Skimp on context and even the best prompt produces generic output.

Specify the output shape. Every prompt tells the model what the output should look like, a table with specific columns, a memo with specific sections, a 5-bullet summary. Models default to wandering narrative output when you don't constrain them. Constraint produces quality.

Iterate, don't accept. The first output of any prompt is a draft. Procurement work is too important to ship a first draft. Use the model's response to identify what's missing, then come back with: "That's good. Now redo it with X tightened and Y added." Two iterations usually beats one super-prompt.

Get the full library, for both ecosystems

Both playbooks are free. Same role coverage. Same workflows. Prompts tuned for each tool's strengths.

Download the Claude Cowork Playbook for Procurement Teams →, 7 role-specific PDFs, 105 prompts, 25K words, downloaded by 400+ procurement professionals.

Download The Codex Playbook for Procurement Teams →, 17,000-word playbook + free starter pack with 8 production-tested Codex skills, AGENTS.md template, governance checklist, and folder structure for ChatGPT/Codex users.

If your team is on Claude, start with the Cowork Playbook. If your team is on ChatGPT, start with the Codex Playbook. If you're a mixed-tool function, common in larger organizations, install both and let each role use the prompts written for the tool they actually use.

Frequently asked questions

Do these prompts work in both Claude and ChatGPT?

The patterns translate, but each playbook is tuned for one ecosystem. The Cowork Playbook is written for Claude's strengths in long-document reasoning and Claude Skills as the workflow surface. The Codex Playbook is written for ChatGPT/Codex's strengths in parallel cloud sandboxes, Goal mode for long-horizon runs, and the AGENTS.md custom-instructions model. You can use either prompt in either tool with light adaptation, but the native version will perform better than the cross-applied version.

Can I share these prompts with my team?

Yes. Both libraries are free to share inside your organization. We ask for attribution if you republish externally, but inside your team, paste them into Slack, drop them into your internal wiki, fork them into your own variants, do whatever helps your team adopt AI faster.

Where do I add my company's context?

Every prompt has bracketed placeholders ([Category], [Supplier Name], [amount], etc.). Fill those with your actual values. Then attach the files referenced, your supplier list, your category taxonomy, your contract template, your evaluation rubric. Context attachment is what separates the 30-second generic answer from the 5-minute usable draft.

Why organized by role and not by workflow?

Procurement work is role-anchored. Each person on the team thinks in terms of "the work I own," not "the workflow categories." A role-cut library means each person grabs their guide and starts using it. A workflow-cut library makes them hunt through 7 sections to find the prompts that apply to them. The role cut wins on adoption.

How do I get more advanced prompts?

If you've worked through the role guide and want to go deeper, turning these prompts into skills, building shared infrastructure across your team, customizing them for your specific categories and supplier base, that's what our 1-week team training engagement is built for. We also run a Procurement AI Academy for individual professionals, modules per Procurement OS skill, monthly newsletter, monthly practitioner roundtable. $49 to try a single module.

What if my team is on Microsoft Copilot, not Claude or ChatGPT?

Copilot is improving fast and is the right answer for organizations already standardized on Microsoft 365. The prompt patterns in both playbooks translate to Copilot with adaptation, though Copilot's enterprise-rollout maturity for procurement workflows is closer to where ChatGPT was 9 months ago. For most M365 organizations, the right move today is to use Copilot for organizational fluency and supplement with Claude or ChatGPT for the procurement-specific workflows where the gap is widest. (Our Claude vs ChatGPT for Procurement comparison has the decision framework.)

Start with one prompt this week

The fastest way to test whether AI is worth investing real time in is to run one prompt against real work. Pick the role guide closest to your job from either playbook. Pick the workflow you spent the most time on last week. Run the matching prompt with your actual context attached. Compare the output to what you would have produced from scratch.

If the output is 80% there in 5 minutes when it would have taken you 60, install the rest of the playbook. If it's not, the prompt is missing context. Add it and iterate. Three iterations gets most procurement professionals to "this is genuinely faster."

Want help rolling these prompts out to your team with shared infrastructure and a 30/60/90 adoption plan?

Book a 20-min scoping call