Documented purchasing process

How can you turn fragmented purchasing practices into a clear, verifiable process that teams can use?

1250 € per session, excl. VAT

In short

In a half-day session of 3 hours participants use generative AI to structure a purchasing process based on notes and business rules. They define the steps, assign responsibilities, document exceptions and test the consistency of the resulting documentation before human approval.

Why this training?

A purchasing process described solely as a task list often leaves decisions, handovers and expected evidence out of view. These omissions make it difficult to handle requests consistently, particularly when an approver is absent or a delivery is incomplete. AI can help organise information and identify contradictions, but it can also invent a missing rule. The training teaches participants to separate documented facts, assumptions and decisions requiring resolution, without confusing a description of current practice with desired improvements. Running case study: The fictitious organisation « Tavrion Pièces » receives purchase requests by email. Some are incomplete, approvals remain implicit and partial receipts are poorly tracked. The exercise pack contains fictitious interviews, authorisation rules and three contrasting requests, from the initial statement of need through to receipt. Transfer to the workplace: Reuse the template within a limited real-world scope using anonymised information, have the process reviewed by a buyer and a requester, then test one pathway before circulation. Record discrepancies, corrections and the business process owner's decision in the version history.

Objectives of this training

  • Define the trigger, expected outcome and boundaries of the process.
  • Extract the necessary activities, decisions and information from the sources.
  • Map the steps and assign responsibility for them.
  • Write a procedure incorporating checks, evidence and exceptions.
  • Assess the consistency of the documentation across three documented pathways.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Buyers and supply coordinators responsible for documenting their practices.
  • Operations managers and quality leads involved in handovers.

Prior skills

  • Have prior experience of writing and refining simple prompts in a generative AI assistant.
  • Know the usual stages of a purchase request and the people involved.

What you will work on

  1. Define the scope and assess the sources

    Objective: Set the boundaries of the process without inventing rules. Trigger, inputs, outputs, stakeholders; distinction between fact, assumption and proposal.

  2. Build the steps and responsibilities

    Objective: Make sequences and handovers explicit. Activities, decisions, feedback loops; person responsible for execution, approver and recipient of the information.

  3. Write a usable procedure

    Objective: Turn the process map into operational instructions. Structure of a step sheet: action, input, output, check, evidence; exception management and version control.

  4. Test and measure quality

    Objective: Identify ambiguities before business approval. Pathway simulation, coverage checks, role consistency and traceability of corrections.

What you take away

  • Scope sheet and register of points requiring resolution.
  • Text-based map of the steps and responsibilities table.
  • Version-controlled purchasing procedure with exceptions and expected evidence.
  • Completed checklist and corrections log.

Method and assessment

  • Short instructional sessions followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group review: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to an AI assistant authorised by the organisation; only fictitious or anonymised cases will be used. Identify a short process to document, without bringing contracts, supplier contact details or confidential data.

Observable criteria: The trigger and output clearly distinguish what belongs within the process. ; Each activity has an identified owner and each decision has explicit outcomes. ; Each rule is linked to a source in the pack or flagged as an assumption to be validated. ; The three pathways are tested; obstacles and corrections are recorded. ; The checklist measures the proportion of steps completed and lists the remaining inconsistencies..

Frequently asked questions

Do you need to know a modelling notation?

No. A text-based representation with steps, decisions and arrows is sufficient; no specialist notation is required.

Can AI define our authorisation rules?

No. It structures the rules provided; any missing threshold or approval authority still requires a decision.

Do we document current practice or a target process?

Current practice first. Proposed improvements are kept separate so that they are not presented as established practices.

Can the documentation be circulated immediately?

It must first be approved by the business process owner, particularly with regard to responsibilities, exceptions and checks.

Last updated: