Reliable functional documentation

How can you turn product information into a usable user guide without letting AI invent behaviour?

1250 € per session, excl. VAT

In short

This half-day session of 3 hours offers a guided introduction to generative AI for writing functional documentation. Using fictitious specifications, participants structure a user guide, describe user journeys and verify every statement. They leave with a reviewed document, a reusable prompt and a checklist.

Why this training?

An incomplete user guide leaves users guessing which permissions they need, which steps to follow or what the consequences of an action will be. Distinguishing the standard journey, restrictions and failure scenarios reduces these ambiguities from the drafting stage. AI can help organise notes and rephrase instructions, but it can also add a button or rule that does not exist. The work therefore focuses as much on the traceability of statements as on the readability of the text. Running case study: The fictitious organisation Brumelis Archives is adding a temporary folder-sharing feature to its portal. Only the owner can create a link, which is valid for seven days and can be revoked. The recipient can view but not edit. Participants write the user guide using the mock-ups and rules provided, with some information deliberately left out. Application at work: For a simple feature in their organisation, participants will be able to reuse the template, gather the approved sources and have the user guide reviewed by a product contact before publication. The checklist will allow them to compare the draft and the corrected version.

Objectives of this training

  • Identify the reader, their needs and the limits of the feature.
  • Classify information into confirmed facts, assumptions and open questions.
  • Write an AI prompt based on a defined set of sources.
  • Describe a standard user journey and two specific situations.
  • Check the accuracy and readability of a user guide using a checklist.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Product and support staff who need to explain a feature to users.
  • Developers and testers wishing to contribute to an accessible functional user guide.

Prior skills

  • No prior AI knowledge required; the first prompts are created step by step.
  • Ability to describe the steps of an action in an application, without programming skills.

What you will work on

  1. Define what the user guide needs to explain

    Objective: Separate available facts from missing information. Target reader, purpose of use, reference sources; introduction to plausible but unverified AI responses.

  2. Build an outline and guide the assistant

    Objective: Obtain a structure suited to temporary sharing. Sections: purpose, access requirements, steps, outcome, limitations; explicit instruction not to fill in missing information.

  3. Write user journeys without extrapolating

    Objective: Produce verifiable and understandable instructions. Numbered steps, interface terminology, expected outcome; coverage of expiry and revocation without inventing system messages.

  4. Review from a user’s perspective

    Objective: Measure the draft’s shortcomings and correct them. Checks for accuracy, completeness and readability; distinction between reviewing documentation and validating actual behaviour.

What you take away

  • Reader scoping sheet and sources table.
  • User guide template and annotated writing prompt.
  • Corrected temporary-sharing user guide, with unresolved questions.
  • Checklist completed before and after correction.

Method and assessment

  • Short teaching inputs followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to a generative AI assistant approved by the organisation and a text editor. Read the fictitious case file provided; any additional example must be anonymised and approved before use.

Observable criteria: Each rule described refers to a supplied source; unknowns are flagged. ; The user journey specifies the required permissions, actions and expected outcome. ; Expired-link and revocation scenarios are distinguished without inventing behaviour. ; Labels match the mock-ups and instructions use action verbs. ; The number of criteria met is recorded before and after correction, with justification..

Frequently asked questions

Do you need to know how to code?

No. The exercise describes usage and visible rules, not the technical implementation.

Can you use internal specifications?

Only if their use is authorised and their content anonymised; the fictitious case file is sufficient.

Does AI validate the feature?

No. It helps with drafting; sources and human review remain decisive.

Can the user guide be published straight away?

It is an exercise deliverable. Adapting it for real use requires confirmation of the unknowns and product team validation.

Last updated: