Clear operational instructions
How can you turn scattered instructions into a short, understandable and verifiable procedure with the help of AI?
1250 € per session, excl. VAT
In short
This half-day session of 3 hours explores how generative AI can help document a procedure for receiving supplies. Working from fictional notes, participants define the scope of the activity, organise the steps, write a first version and then check that it is understood through a peer simulation.
Why this training?
An overly vague procedure forces a new starter to interpret the instructions: what should be checked, in what order, and who should be notified if there is a discrepancy? Describing observable actions reduces ambiguity and makes sticking points visible. AI can suggest a structure and rephrase notes, but it can also add responsibilities or rules that do not exist. The training therefore distinguishes writing support from subject-matter validation, which remains essential before use. Running case study: The fictional organisation Talmérie Fournitures welcomes a new employee responsible for receiving deliveries. Instructions are scattered across notes and messages. The group must explain how to compare a delivery with the order, record a discrepancy and pass the information to the procurement contact, without inventing missing rules. Transfer to the workplace: After the session, choose a routine, low-risk activity, reuse the template and have the procedure reviewed by someone who performs that activity. Submit the rules and responsibilities to the subject-matter lead before any distribution; keep unresolved questions separate from the applicable instructions.
Objectives of this training
- Define the start, end and intended audience of a procedure.
- Extract the necessary actions, responsible people and information from the notes.
- Formulate an AI prompt specifying the sources and boundaries.
- Write numbered steps using action verbs.
- Identify ambiguities through a simulation and correct the document.
Who is it for and what are the prerequisites?
Roles and tasks concerned
- Procurement assistants and administrative staff required to document routine instructions.
- Operations or logistics staff wishing to hand over a simple activity.
Prior skills
- No prior AI knowledge required: the first interactions are guided.
- Ability to enter and edit text on a computer.
What you will work on
Define the scope of the receiving process to be described
Objective: Choose a sufficiently narrow scope for a short procedure. Difference between a procedure and a general description; trigger, end point, reader and available sources.
Discover assisted writing
Objective: Obtain an outline without delegating business rules to AI. How a generative assistant works, limitations of its responses and how to formulate a prompt: context, sources, format, prohibition on inventing information.
Write actionable steps
Objective: Turn the outline into precise instructions. Action verbs, chronological order, responsible people, required documents and wording of the conditions for raising an alert.
Simulate, check and correct
Objective: Assess understanding of the document without verbal explanation. Step-by-step reading, identifying omissions, traceability of corrections and validation before distribution.
What you take away
- Scope sheet for receiving supplies.
- Writing instructions for the AI, explicitly prohibiting invented information.
- A corrected short procedure, marked as a draft awaiting validation.
- Completed review checklist and list of subject-matter questions.
Method and assessment
- Short teaching inputs followed by immediate practice using the running case study.
- Individual work followed by work in pairs, with participants reviewing each other’s outputs.
- Guided group discussion: what works, what needs correcting and what remains a human decision.
- Preparation: Arrange access to a generative AI tool authorised by the organisation; no specific commercial tool is required. Prepare a fictional or anonymised example of instructions, without personal data or confidential information; a training pack is provided.
Observable criteria: The trigger, intended audience and expected outcome are explicitly stated. ; Each step specifies an observable action and the person responsible for it. ; Checks and the handling of a discrepancy follow only the information provided. ; Missing information is flagged, without arbitrarily adding a rule. ; The simulation records the steps understood without help and the sticking points; each sticking point leads to a correction or a question..
Frequently asked questions
Do you need to have used AI before?
No. The first prompt is developed collectively from the case study notes.
Can you use an internal procedure?
Only in anonymised form and in accordance with the organisation’s rules; the fictional case study remains available.
Can the procedure produced be applied immediately?
It is an operational draft that must be validated by the subject-matter lead before distribution.
What should you do if AI fills in a missing rule?
Remove the addition, flag the missing information and ask the relevant person responsible for confirmation.
Last updated: