AI use-case directory · Procurement

Spend classification: real AI examples and a starter prompt

AI sorts messy purchase lines and free-text descriptions into a standard spend taxonomy so you can see what you actually buy, from whom, and where to consolidate.

Best for: Procurement leaders, Buyers

Real-world examples

Fortune Global 500 manufacturer (published Oraczen case study)

Classification accuracy reached 95% (misclassification fell from 20% to under 5%), manual effort dropped from 12,000 to 1,800 hours a year, and $30M in annual savings opportunities were identified.

Source: Oraczen case study

Tools

Free or low-cost options first, then enterprise platforms.

General-purpose AI assistant (ChatGPT, Claude, Gemini, Copilot; free tiers)
Excel or Google Sheets + AI formulas
Sievo
Coupa
SAP Ariba

Starter prompt

Replace the [BRACKETS]. Remove names and confidential numbers first.

Here are [N] anonymized purchase lines (description, supplier type, amount). Assign each to one category from this taxonomy: [TAXONOMY]. Return a table with line, category, confidence (high/medium/low), and a one-line reason. Flag anything you can't classify.

Related Brief coverage

#AI adoption

Physical Intelligence opens its robot models to outside developers (Issue #2)
Practitioner tip: "boring AI" is the procurement win (Issue #1)

More procurement use cases

Vendor names are examples, not endorsements or paid placements. Links marked "affiliate link" may earn us a commission at no cost to you; that never affects which tools we list. Outcomes are as reported by the linked source. Anonymize data and use your company's approved AI tools.