GLOBALTECH
Understand through an example

Hector I.A.: an n8n demonstration of Qualiopi preparation

A simulated document review with Hector I.A.: requirements, evidence, gaps, corrective plan and human approval to prepare the Qualiopi quality process.

Simulated demonstration · fictional data · no connected services. This page processes no real records, messages, purchases or accounts.

Observe the workflow and decisions

Start the demonstration, review the fictional matrix, then click “Simulate my approval”. The actual corrections still need to be completed and checked.

Founder’s avatar

Stéphane Judicaël introduces Hector I.A.

AI portrait of Stéphane Judicaël, the founder’s avatar

Video presentation coming soon.
The written presentation is available below.

Read Stéphane Judicaël’s presentation

Hello, I am the avatar of Stéphane Judicaël, founder of GlobalTech. I am introducing Hector I.A., our virtual agent for compliance support and Qualiopi preparation.

His role is to bring together applicable requirements, our practices and available evidence. He helps identify inconsistencies, prepare corrections and make a document review easier to verify.

This animation is a mock-up: it uses a fictional record and performs no real operations. The illustrated connections remain to be configured.

Each block represents a step. n8n organises the flow of information between QualiProcess, Avatar Hub, Hector’s analysis and my approval.

It starts with a request: prepare a review of documents for a training session. QualiProcess provides references to the programme, agreement and assessment index.

Before any consultation, the workflow checks access rights. Hector must work only within the authorised organisation and session.

Avatar Hub then provides his profile: mission, tone and limits. His approach should be precise, explanatory and rigorous, without unnecessary jargon.

Hector checks the requirements configured for the review and the accessible documents. He compares each expectation with the evidence found and records the versions examined.

In our example, he identifies different dates in the programme and agreement. He also cannot find assessment evidence in the documents consulted. This calls for verification: a document not found does not mean that it does not exist.

These observations feed a matrix: the requirement reviewed, available evidence, observed gap and proposed action. Hector also prepares a corrective plan: check the right dates, align the relevant documents and search for the assessment evidence. Marie or Alice can be asked to complete the administrative record.

The workflow then stops for my approval. I can examine the sources and proposals. The “Simulate my approval” button represents this human decision in the demonstration.

After my simulated approval, the plan is recorded in QualiProcess with document references and the approval record. A review report gathers observations, versions and decisions. The actual corrections still need to be completed and checked.

If interpretation is uncertain, a branch provides for a decision by me. It is not executed in this example.

Hector does not issue certification or guarantee compliance. He never fabricates evidence, signatures or attendance.

The intended benefit is more structured preparation: finding supporting documents, understanding points to check and tracking corrections. Hector prepares the review; I retain the decision.

Prepare a document review without fabricating evidence

For a training provider, preparing a review means comparing the requirements examined, described practices and available documents. Hector represents assistance with that preparation: identifying references, recording versions and outlining points to check. His avatar is used on QualiProcess IA; the n8n workflow shown here remains a simulation of connections to configure.

A fictional record with two concrete observations

Programme v2 and agreement v1 show different dates. Assessment evidence is not found in the consulted index. The matrix links each observation to a source and proposes an action: verify dates, align the relevant documents and search for the missing evidence. A document not found does not prove it does not exist; the scope and available records must be checked before drawing conclusions.

Human approval before recording the plan

Stéphane examines references and proposals before approving the plan. The demonstration then simulates recording the plan and preparing a review report. It does not carry out corrections, issue certification or guarantee Qualiopi compliance. Marie or Alice can be asked for further documents; the optional decision branch is not executed in this example.

How can this process fit your business?

GlobalTech can help define a use case: the agent mission, authorised resources, steps suitable for automation, human decisions and traceability. Real integration requires application access, validated rules and testing within a controlled scope. This page explains the process; the animation is not an importable n8n workflow.

AI Agent and Automation ↗Discuss my AI project