What is generative AI for procurement, and how do you actually learn it?

Generative AI tools — ChatGPT, Claude, Copilot, Gemini — produce new text and analysis from a prompt. In procurement, that means drafting, summarizing, and analyzing RFPs, contracts, and supplier communications on demand, not just retrieving stored data.

What does generative AI actually do in procurement?

Unlike traditional procurement software, which retrieves stored data or runs predefined reports, generative AI creates something new each time you use it. Give it a lengthy supplier agreement and ask it to list every indemnification obligation — it reads the document and writes a summary that didn't exist before. Give it competing quotes in different formats and ask it to flag pricing discrepancies — it builds a comparison you never templated.

The practical applications span drafting RFP evaluations, summarizing contract terms for internal stakeholders, normalizing supplier data across inconsistent formats, and composing first-draft correspondence. For a fuller catalog, see AI procurement use cases.

None of this replaces a buyer's judgment. The output is a first draft, not a decision. But it eliminates a large share of the manual reading, reformatting, and re-typing that fills a procurement day.

How is generative AI different from the procurement tools you already have?

Your ERP runs workflows. Your sourcing platform manages events. Your CLM stores contracts and tracks approvals. Generative AI does something none of those systems do: it reads unstructured text — a supplier's email, a long MSA, a price-increase letter — and produces a written response, a summary, or a structured comparison on request.

It does not replace those systems. It handles the reading and writing work that happens between them — the part that, until now, has been done by hand.

What should generative AI training for procurement teams include?

Several universities and course platforms already offer programs specifically titled "generative AI for procurement." Most follow a similar structure: a theory module explaining how the underlying models work, followed by a demonstration of one vendor's product.

The gap in most of those programs is practice on real procurement documents. A team that watches a demo of contract summarization learns that the capability exists. A team that briefs an AI tool on a supplier agreement — even a synthetic one — and then verifies the output against the source learns where the tool is useful and where it breaks.

Effective training should cover how to brief the tool with a structured prompt rather than a one-off question, how to attach context without exposing company data, and how to judge what comes back. It should also work across platforms, because most teams have access to more than one and the method needs to transfer.

What does LucyWorks offer?

LucyWorks teaches generative AI for procurement in two formats.

The team workshop is a 2-hour live session (virtual or on-site) for up to ≤15 participants from a procurement or supply-chain team. Each person builds a working AI agent tied to one of their own recurring tasks and leaves the session with it running — not with a certificate. The method works on Copilot, ChatGPT, or Claude, because vendor academies can only teach their own tool; this trains the method. €2,500 pilot price for the first three teams, in exchange for honest feedback and a reference conversation. Guarantee: if the session does not give your team at least one workflow they will actually run next week, you do not pay.

The 20 Lessons is a self-serve course at €199, one-time. Twenty short lessons, each running one real procurement task across ChatGPT, Claude, Gemini, and Copilot, with real screenshots and synthetic case files. Immediate access after purchase.

See pricing for the full breakdown.

Built by a practitioner running AI adoption inside a Fortune 500 procurement organization.