Fautons
9 min readAI trainingProcurement

AI for procurement: what your team should learn first

AI for procurement: what your team should learn first

What AI for procurement means, in plain terms

AI for procurement means using general AI assistants and the AI built into sourcing software to do the reading and writing that fills a buyer's week. That covers drafting tender documents, researching suppliers, reviewing contracts and summarising spend. You probably don't need a new platform to start, because most organisations already pay for ChatGPT, Copilot or Claude, and many procurement suites now ship with an assistant of their own.

Search for the topic and you mostly find software vendors explaining what their platform automates. That is useful if you are choosing a platform. It says very little about the people who have to use it, which is where most teams get stuck. In practice, four jobs account for most of the early value:

  • Supplier research: a first pass on a market, a shortlist of possible suppliers and the questions to ask them, which a buyer then checks against real sources.
  • Tender and RFx drafting: a first draft of a specification, an evaluation question or a supplier questionnaire, written from your own templates.
  • Contract review: comparing a supplier's paper against your standard terms and listing where it differs, for a person to judge.
  • Spend summaries: turning a large export from the finance system into a short narrative of where the money goes and what has changed.

None of these hand the decision to the tool. They take the reading and the first draft off a person, so that person has more time for the judgement calls and the conversations with suppliers.

Why most procurement teams aren't using it yet

The hold-up is usually skills. Deloitte surveyed more than 100 chief procurement officers across Europe, North America and Asia-Pacific in early 2024. 92% were planning and assessing generative AI, but only 37% were piloting or deploying it in procurement. Respondents ranked the limited availability of generative AI talent and skills as a significant barrier, and nearly 71% described their knowledge of the technology as limited to moderate.

The vendors say much the same thing in their own way. Sievo's guide to AI in procurement says that high-performing procurement organisations "adopt AI earlier and more consistently", but it never explains how a team gets there. Precoro's guide comes closer: buying AI tools without the skills to run them, it says, "can be like buying a high-end sports car but never learning how to drive."

So if your organisation already has an AI assistant and your buyers aren't using it, the next move is usually a few hours of hands-on practice on real work, with clear rules about data, before anyone buys another licence. We wrote a wider guide on how to use AI at work without creating a problem that applies to any team.

What to teach a procurement team first

Start with tasks the team does every week, where a mistake is cheap to catch. Four skills cover most of the ground, and each one can be practised on the team's own documents in an hour or so.

Drafting a supplier questionnaire. Give the assistant your template, the category and what matters most on this contract, and ask for a first draft of the questions. Then ask it what a weak supplier might hide in their answers and which questions would expose that. The buyer edits, cuts and owns the final version. The skill being taught is writing a clear brief, which is most of what a good prompt is.

Checking a contract against your standard terms. Paste in your standard clauses and the supplier's version, and ask for a table of every difference, with the clause reference and what the change means for you. People need to learn two habits here: check every line of that table against the documents themselves, and never treat it as legal advice. The output is a reading aid for the person who signs off, and it doesn't replace them.

Turning a spend export into a one-page summary. Most AI assistants can now read a spreadsheet. Ask for the top suppliers by spend, the categories that grew fastest and anything that looks like duplicate or off-contract buying, and ask it to show how it worked each figure out. Then check two or three numbers by hand. This is the skill where overtrust does the most damage, so the checking deserves as much practice time as the asking.

Knowing what never goes into a public AI tool. This one is a rule rather than a technique, and it belongs in the training rather than in a policy nobody reads. It's covered in more detail in the next section.

If you are not sure which of these to start with, pick the one the team complains about most. Our guide on which workflow to automate first has a simple way to score the options.

What never goes into a public AI tool

Procurement handles more commercially sensitive material than most teams, so the data rules need to be taught early and practised, not just published. Agree these before the first session, using your organisation's own policy and approved tools:

  • Suppliers' bids and pricing during a live tender. If a bid leaks, the competition is compromised, and a public chatbot account is not the place for it.
  • Personal data, such as named supplier contacts, staff details or anything covered by UK GDPR.
  • Confidential contract terms and negotiation positions, unless you are using an enterprise account your organisation has approved for that kind of material.
  • Anything under an NDA, or anything a supplier has marked commercially confidential.

Where the real work is confidential, run the exercises on sanitised copies: a past tender with names and figures removed works well. People learn the same skills without any risk to a live procurement.

A one-day workshop plan for a procurement team

Here is how we would shape a single hands-on day for a buying team of 5 to 15 people, which is the group size where everyone can build something and still get individual help. It runs on the team's own material, ideally a recent tender or contract with the sensitive parts removed.

  • Morning, first hour: the rules. What data goes where, which tools are approved, and which outputs always need a human check before they leave the team.
  • Morning, second block: supplier questionnaires and specifications. Everyone drafts one from their own template, then the group compares results and tightens the prompts that worked.
  • Late morning: contract comparison. Each person runs a supplier's terms against the standard clauses and then checks the output line by line against the documents.
  • Afternoon, first block: spend. The team works on a real export, asks for a summary, and checks the figures by hand until they trust the method rather than the tool.
  • Afternoon, last block: each person picks one recurring task from their own week, builds a reusable prompt or workflow for it, and writes down how they will check its output.

By the end of the day everyone has something they will use the next morning, and the team shares one way of working with AI instead of each person improvising. That is the format of our AI team workshop. For the wider question of how to choose a format, see AI workshops: which format suits your team.

If you buy for the public sector

Public buyers work under tighter rules on fairness and transparency, so AI use needs more care on both sides of a tender. The Cabinet Office's procurement policy note PPN 017, Improving Transparency of AI use in Procurement, applies to central government departments, their executive agencies and non-departmental public bodies for procurements starting on or after 24 February 2025. It says plainly that suppliers' use of AI is not prohibited during the commercial process, and it compares the risk to a supplier using a bid writer. It also gives optional questions that buyers can add to a tender to understand how suppliers use AI.

For a public sector buying team, that means the training should cover two things: using AI safely on your own work, and reading bids that may have been drafted with it. Our page on AI training for public sector teams covers the format, and our explainer on how the UK government buys training covers the purchasing routes.

Course, platform training or a team workshop?

There are three common ways to build these skills, and they suit different needs. CIPS, the professional body for procurement, runs its own AI courses for procurement professionals. Your software vendor will usually train you on its platform. A live team workshop puts the whole team through the same practice on its own work.

OptionBest forWhat it won't do
A professional course, such as those CIPS runsAn individual who wants structured learning and a recognised route alongside their careerChange how the rest of the team works, or use your own documents
Your software vendor's platform trainingLearning the features of a tool you have already boughtCover the general assistants your team also uses, or the judgement about when not to trust the output
A live, hands-on team workshopGetting a whole team using AI on its own tenders, contracts and spend in a short timeGive anyone a credential

Many teams end up using more than one. A sensible order is the team workshop first, so everyone shares the basics and the data rules, then courses for the people who want to go further.

Where Fautons fits

We run live, hands-on AI training for teams across the UK, live online anywhere or in person around Birmingham and the West Midlands. A procurement session is built on your own templates, contracts and spend data, sanitised where it needs to be, and covers the tools you already have, whether that is ChatGPT, Copilot or Claude. For larger organisations rolling AI out across several functions, our corporate AI training programmes cover the wider plan.

If you want to talk through what a day would look like for your buying team, book a scoping call. We will tell you plainly whether a workshop is the right step or whether a course would serve you better.

Frequently asked questions

What is AI for procurement?

AI for procurement means using AI assistants such as ChatGPT, Copilot and Claude, and the AI built into sourcing software, to help with procurement work. The common uses are researching suppliers, drafting tender documents and questionnaires, comparing contracts against standard terms and summarising spend. A person still makes every decision. The AI does the first draft and the heavy reading.

Will AI replace procurement jobs?

It is changing the work more than removing the role. The tasks AI handles well are the first drafts and the reading, while negotiation, supplier relationships, risk judgement and sign-off stay with people. Buyers who learn to use AI on the routine parts of the job tend to get more time for the parts that need judgement.

Is it safe to put supplier contracts into ChatGPT?

Not into a personal or free account. Contract terms, bids and pricing are commercially sensitive, and personal data about supplier contacts is covered by UK GDPR. Use only the AI tools your organisation has approved for confidential material, follow your own AI policy, and practise on sanitised copies when you are learning.

How long does it take to train a procurement team on AI?

A single hands-on day is enough to get a team using AI safely on its main recurring tasks, if the day is built on the team's own documents. Building habits takes longer. Most teams need a few weeks of applying it to their real work afterwards, ideally with a check-in, before it becomes routine.

Do we need to buy AI procurement software first?

Usually not. Most organisations already have a general AI assistant that can draft, compare documents and read spreadsheets. Training the team to use that well is a cheaper first step, and it also tells you which tasks a specialist platform would need to handle if you decide to buy one later.

Sources

More from our Blog

November 18, 20269 min read

AI for small business: how to train a team of 10 to 20

Most guides to AI for small business are lists of tools. The tools are the cheap part. This is how a business of 10 to 20 people gets everyone using AI well on real work, in about a month.

AI trainingSmall business
Read the article
November 16, 20269 min read

AI in learning and development: the two jobs L&D now has

Most guides to AI in L&D are about building courses faster. That is half of it. The other half is that L&D is now the team expected to teach everyone else to use AI well.

AI trainingL&D
Read the article
October 18, 20268 min read

How to roll out AI across your team: an adoption plan that sticks

Most AI rollouts stall because they're run like a software purchase when they're really a change programme. Here's a five-step plan for People and L&D leaders: start with a reason, pilot on real work, set guardrails, train people properly and measure weekly usage.

AI adoptionAI training
Read the article