⏱️ The Essentials in 3 Minutes

• The knowledge base is managed exclusively by admins and forms the foundation of Didask AI.
• The AI draws on the base's content to answer questions: if information is not explicitly present, it cannot respond accurately.
• Didask training courses are automatically integrated. You can enrich the base with your business documents.
• Base quality depends on explicit titles, absence of redundancies, and removal of outdated content.

🧠 Understand the Pedagogical Value of the Knowledge Base

The knowledge base is the heart of Didask AI: it contains the expertise specific to your organization. Without it, the AI can neither answer your collaborators' questions nor support their skill development.

The challenge is not simply to store documents, but to make your company's knowledge accessible at the right moment, to the right person, in the right context. As an admin, you are the guardian of this quality: well-organized content allows the AI to guide your learners with precision, while disorganized content generates vague or incorrect responses.

🛡️ Understand Who Can Access the Knowledge Base

Management, organization, and improvement of the knowledge base is reserved for platform administrators.

As an admin, your role is to organize this base so that it contains quality information for your collaborators.

🔎 Understand How Didask AI Uses Your Knowledge Base

When a learner asks a question, the AI searches your base to identify the sources closest to their need. It selects the most relevant ones, filters out those that seem distant, and formulates a response. It provides:

A reliability level indicating whether the sources found allowed it to respond
The sources used, so the learner can verify the context of the response

Important: if information is not explicitly present in your base, the AI will not be able to respond accurately. It does not make inferences beyond what your base contains.

📌 Example: if the question is "where is our data hosted?" and your base only contains the name of the hosting provider, the AI will not deduce the server location. The information must be explicitly stated: "our data servers are in Ireland, and our application servers are in France."

🎓 Benefit from a Pre-Populated Base with Your Training Courses

If you use the Didask authoring interface, your training content is automatically integrated into the knowledge base in real time. This integration requires no action on your part and includes all the text content from your projects.

Your learners can immediately ask questions about their training courses to:

Clarify a concept covered in the course
Get additional examples adapted to their context
Review specific points before an evaluation
Deepen a topic they are interested in

You can enrich this base with your specific business content: internal procedures, operational guides, technical documentation, resources from your collaborative tools.

🚀 Improve Your Knowledge Base Performance

A high-performing base is one with little "competition" between knowledge items. When a learner asks a question about a product feature, several documents may compete: the resource center article, the technical meeting notes, the newsletter mentioning that feature's release. The AI has filtering mechanisms, but you can further improve its performance with three simple actions.

Give each document an explicit title:

Good: "Data servers: location and infrastructure", "Cybersecurity fundamentals workshop", "Design review meeting notes - R&D project HERA"
To avoid: "Tech doc", "Workshop 14_Dupont", "R&D and design sync"

Remove redundant or contradictory information:

If two documents written by different teams cover the same topic with contradictory versions, the AI cannot identify the source of truth. Delete the outdated version.

Avoid time-limited sources: