Usable and prioritised backlog

How can you turn scattered requests into clear, ordered and verifiable backlog items with the help of AI?

1250 € per session, excl. VAT

In short

In half a day of 3 hours you turn a collection of requests into a usable product backlog. Generative AI helps group, rephrase and question the items; you decide on their scope, order and acceptance criteria. You leave with a documented backlog and a quality checklist.

Why this training?

A backlog full of ambiguous requests perpetuates duplicates and makes trade-offs difficult. Distinguishing between a user problem, a feature and a technical task allows you to discuss value before detailing a solution. AI can suggest groupings and reveal imprecise wording, but it can also invent business rules. Every proposal must remain traceable to a source request, and every assumption must be flagged before validation. Running case study: The fictitious organisation « Brévance Location » is preparing an equipment booking portal. Fifteen requests mix availability, extensions, invoicing and incidents. The team must clarify an initial scope centred on booking, without confusing commercial requests, user problems and solutions already suggested. Application at work: At the next refinement session, take five anonymised requests from your team, apply the checklist and compare their completeness before and after rewriting. Have the acceptance criteria and dependencies validated by the relevant product and technical staff.

Objectives of this training

  • Classify requests by user problem and identify duplicates
  • Break down a feature into sufficiently bounded items that deliver value
  • Write user stories with testable acceptance criteria
  • Order items according to explicit criteria and their dependencies
  • Measure backlog quality using a checklist

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Product owners and product managers responsible for backlog refinement
  • Business analysts and project managers involved in clarifying needs

Prior skills

  • You have already used a generative AI assistant to rephrase or summarise a text.
  • You are familiar with the concepts of backlog, user need and product iteration.

What you will work on

  1. Sort requests and frame the value

    Objective: Separate needs, suggested solutions and out-of-scope requests. Product objective, users concerned, source traceability; AI instructions limited to the supplied case materials.

  2. Break down and clarify items

    Objective: Obtain understandable and testable stories. Splitting by user value; distinction between story and task; acceptance criteria, edge cases and unknown rules.

  3. Order without concealing uncertainties

    Objective: Justify a coherent order of work. User value, urgency, risk and dependencies; limitations of scores; effort estimates to be confirmed with the team.

  4. Check and prepare for refinement

    Objective: Measure quality and prepare the remaining validations. Checklist covering completeness, ambiguities, testability and traceability; distinction between documentation quality and business validation.

What you take away

  • Request map with sources, groupings and exclusions
  • Ordered backlog with five detailed user stories
  • Log of assumptions, dependencies and trade-offs
  • Completed quality checklist and reusable AI instructions

Method and assessment

  • Short explanations 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 an AI assistant authorised by the organisation and an editable table format. Prepare a few anonymised requests if desired; the supplied fictitious case materials allow all exercises to be completed without real data.

Observable criteria: Each retained item refers to a source request and an explicit user need. ; The five detailed stories include verifiable criteria, including a relevant error scenario. ; The proposed order is based on documented criteria and takes account of the identified dependencies. ; Missing information is marked as assumptions or questions, without becoming business rules. ; The percentage of items meeting the checklist requirements is calculated before and after correction..

Frequently asked questions

Does AI decide the priorities?

No. It proposes comparisons; the product lead makes the trade-offs according to the documented objectives and constraints.

Is specific backlog management software required?

No. A table is sufficient; importing into the team's software remains out of scope.

Can internal tickets be used?

Only when anonymised and in authorised tools, without technical secrets or personal data.

Will the backlog be ready for development?

It will be ready for refinement. The business rules, estimates and dependencies still need to be validated by the team.

Last updated: