Documentary research with AI

How can you use AI to explore a documentary research question without confusing a plausible answer with verified information?

1250 € per session, excl. VAT

In short

In a half-day session of 3 hours discover how AI can help refine a question, suggest avenues for documentary research and prepare a synthesis. Working from a fictional case, you will produce a referenced briefing note, verify the selected claims and assess the quality of the result using a simple assessment grid.

Why this training?

A fluent answer can conceal a non-existent reference, outdated information or an overly broad interpretation. Assisted research therefore requires a distinction between what AI suggests and what the documents actually support. The educational value lies in practising a short workflow: question, research avenues, consultation, verification, synthesis. Participants discover where AI helps with exploration and rephrasing, while retaining responsibility for selecting sources and drawing conclusions. Running case study: The fictional association Avelune Ressources is preparing a workshop on repairing small objects. Its coordinator wants to understand the practical requirements for organising it before bringing volunteers together. Participants explore this question using an anonymised learning pack, then write a briefing note distinguishing established information from points requiring further investigation. Application at work: After the session, revisit a non-sensitive documentary research question from your work. Reuse the scoping approach and evidence table, then ask a colleague to review the note and find the passages supporting three claims.

Objectives of this training

  • Formulate a documentary research question with an explicit scope.
  • Write an AI prompt specifying the context, expectations and limits.
  • Distinguish between a research suggestion, a consulted source and a verified claim.
  • Check the existence, date and relevance of a reference.
  • Produce a short briefing note with accessible supporting evidence.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Staff who need to prepare initial documentary research for their team.
  • Assistants, coordinators and project officers starting to use AI assistants.

Prior skills

  • No prior AI knowledge required.
  • Ability to consult a digital document and carry out a simple search in a browser.

What you will work on

  1. Understand what the assistant produces

    Objective: Identify the limitations of a generated answer to a documentary research question. Difference between text generation and consulting sources; fabricated references, currency and confidentiality.

  2. Define the scope of the research

    Objective: Turn a vague need into a workable question. Context, audience, scope, keywords and expected format; an explicit request to flag uncertainties.

  3. Return to the documents

    Objective: Verify information before reusing it. Author, date, primary source, independence of sources and correspondence between claim and extract; handling contradictions.

  4. Synthesise and check

    Objective: Write a note that faithfully reflects the selected sources. Synthesis limited to the documents provided, attribution of information and identification of gaps; claim-by-claim checking.

What you take away

  • Documentary research scoping sheet and initial AI prompt.
  • Table of sources, supporting extracts and identified limitations.
  • One-page synthesis note with source references.
  • Completed checklist with priority corrections.

Method and assessment

  • Short instructional inputs followed by immediate practice using 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; check connection requirements before the session. Provide an anonymised learning pack with dated documents, accessible references and sample answers to examine.

Observable criteria: The question specifies the need, scope and exclusions. ; Each selected reference has been opened and its origin identified. ; Three central claims are linked to a specific passage in a consulted source. ; The note separates documented facts, interpretations and missing information. ; The grid counts supported claims and flags those to remove or rephrase..

Frequently asked questions

Does the assistant need access to the web?

No. The exercises can use the pack provided; available capabilities are checked before any online research.

Can internal documents be used?

The exercises use only fictional or anonymised cases, within tools authorised by the organisation.

Is a reference cited by AI sufficient?

No. You must locate the document and check that its content actually supports the claim.

How can you assess the final quality?

By checking the traceability of claims, the relevance of sources and the visibility of uncertainties, rather than just the fluency of the text.

Last updated: