Recommendations / Employer, management and HR

Recommendations · Employer, management or HR

Deploying AI in a team: five steps for the employer

An AI project passes through five gates before it is deployed, then it is monitored over time. A closed gate calls for the actions shown before moving on to the next one.

By Dr Charles Broutin, occupational physician · Updated on

Read the pathway ↓Open the flowchart →

The essentials

The principle, and what holds in every case.

The steps that depend on no answer.

The principle

“Bringing AI into a team is an organisational decision, not just a choice of tool.”

To do whatever the situation:

  • For all uses. Map the data: purpose, legal basis, flows, access, processors, retention period. Check whether a data protection impact assessment (DPIA) is required: AI does not automatically call for one, the risks of the processing decide.
  • In all cases. Involve the employees concerned in choosing and configuring the tool, and call on the occupational physician or the multidisciplinary team of the occupational health service while the project can still change.
  • Follow-up schedule. Record the effects at T1 (4 to 8 weeks), T2 (3 months), T3 (6 months), then every 6 to 12 months, and after any change of model, data, objective or population, or after an incident. Indicators remain collective: never used to assess or rank people.

The pathway

Question by question, the recommendation that follows.

The same content as the flowchart, as continuous reading. Each recommendation points to the page of the site it is drawn from.

AI project in a team

Gate 1Proven need

  1. 1.1

    Does the project start from a specific work problem, with alternatives examined?

    If no

    Describe the task, the people concerned, the expected result and the points where an error would have consequences. “Saving time” is not a task. Name who will be able to launch, change or stop the pilot.

    Sources: Law and governance · Understanding AI · five questions

Gate 2Clearly defined use

  1. 2.1

    What will the system do?

    Obligations depend on the function the system performs, not on the software brand.

    Drafting, summarising, searching

    Usage note, data entry rules, human verification protocol and review time built into the workload.

    Recruiting, assessing, managing mobility

    Sensitive use: AI Act classification (some employment uses are high-risk, with rules applying from 2 December 2027), non-discrimination, informing the works council and candidates in advance, a way to challenge decisions.

    Labour Code L. 2312-38 · L. 1132-1

    Planning or allocating tasks

    Mapping of work before and after, rules for contesting, attention to autonomy, fairness and work intensification.

    Monitoring activity or performance

    Purpose, necessity and proportionality test, informing employees in advance, consulting the works council before implementation. No fully algorithmic individual assessment.

    Labour Code L. 1222-4 · L. 2312-38 · GDPR art. 22

    Touching on health or fitness for work

    Legal, clinical, technical and organisational analysis before any pilot: health data, professional secrecy, separation of roles.

    Source: Law and governance · use cases

  2. For all uses

    Map the data: purpose, legal basis, flows, access, processors, retention period. Check whether a data protection impact assessment (DPIA) is required: AI does not automatically call for one, the risks of the processing decide.

    GDPR art. 35

    Source: Law and governance · GDPR

  3. 2.2

    Does the system act without prior human approval?

    • A0 Information
    • A1 Recommendation
    • A2 Supervised action
    • A3 Autonomous action
    • A4 Delegation

    “Yes” means levels A3 and A4 on the autonomy scale in the works council guide.

    If yes

    Define in advance what it can do on its own, require explicit approval for actions with serious consequences, keep a log of its actions and keep a real means of stopping it. The higher the level, the better equipped human oversight must be.

    Sources: Works council and AI · autonomy · Reading · Human in the loop

Gate 3Organised dialogue

  1. 3.1

    Does the project change jobs, the organisation, working conditions or the monitoring of activity?

    If yes

    In companies with at least 50 employees, inform and consult the works council (CSE) early enough for its opinion to still be able to change the project, including for a tool presented as optional (2026 court decisions). Provide a description of the need, the groups of employees concerned, the data and the supplier’s documentation.

    Labour Code L. 2312-8 · L. 2312-14

    Sources: Works council and AI · Law and governance · social dialogue

  2. In all cases

    Involve the employees concerned in choosing and configuring the tool, and call on the occupational physician or the multidisciplinary team of the occupational health service while the project can still change.

    Sources: Psychosocial risks · autonomy · Works council and AI · occupational health

  3. 3.2

    Will roles, occupations or jobs change?

    If yes

    Inform early and precisely: distinguish what has been decided, what remains uncertain and when the next information will come. Include AI in strategic workforce planning (GEPP), open training to everyone and preserve the positions where people learn the job.

    Labour Code L. 2242-20

    Sources: Psychosocial risks · insecurity · Economic and social risks

Gate 4Reversible pilot

  1. 4.1

    Have you assessed the effects on real work with the people concerned?

    If no

    Before the pilot (T0), fill in the nine-axis grid: intensity, emotions, autonomy, relationships, values, insecurity, contestability, supervision load, skills. Record the risks identified in the DUERP (the company’s occupational risk assessment document) where the framework requires it.

    Labour Code L. 4121-3 · R. 4121-1

    Sources: Assess an AI project · Law and governance · DUERP

  2. 4.2

    Do you plan to raise targets thanks to the time saved?

    If yes

    Not before measuring the complete task, checking and corrections included, task by task and across different groups of staff: a positive average can hide negative effects. Decide with the team what share of the gains goes to production, quality, learning and recovery.

    Sources: Reading · AI and real work · Psychosocial risks · intensity

  3. 4.3

    Can employees set aside an AI output without penalty, and obtain a human review?

    If no

    State in writing that professional judgement takes precedence over the AI output, guarantee the right to correct, refuse or report without sanction, and designate who can review a decision, and within what time.

    Source: Assess an AI project · autonomy, contestability

  4. 4.4

    Do those who approve have the time, the skills and the authority to refuse?

    If no

    Match the volume of outputs to the real capacity for checking, count supervision in workload and staffing, and train during working hours. Plan periods when no response is expected, even if an agent is blocked.

    AI Act art. 4 · art. 26

    Sources: Reading · Human in the loop · Reading · AI agents and disconnecting

  5. 4.5

    Does the pilot have a duration, a scope, stop thresholds and a tested fallback?

    If no

    Prepare the eight documents of the pilot file: usage note, current and planned work, AI Act classification, data mapping, impact assessments, information and consultation file for the works council, reversible pilot protocol, monitoring and incident plan.

    Source: Law and governance · pilot file

Gate 5Documented follow-up

  1. Follow-up schedule

    Record the effects at T1 (4 to 8 weeks), T2 (3 months), T3 (6 months), then every 6 to 12 months, and after any change of model, data, objective or population, or after an incident. Indicators remain collective: never used to assess or rank people.

    Source: Assess an AI project · schedule and safeguards

  2. 5.1

    Do the observed effects remain compatible with health and the quality of work?

    Workload, autonomy, skills, incidents, contested decisions.

    If yes

    Continue the use, subject to conditions, with the next measurement already scheduled.

    If no

    Correct the organisation, resources or rules; reduce the scope or suspend the use.

    Source: Assess an AI project

A reasoned decision, recorded and shared with the works council

Taking action

The site’s tools for this audience.

Assess an AI project18 questions on nine axes, before the pilot and then at follow-up. Nothing is sent. Pilot fileThe eight documents that make a pilot open to discussion, traceable and reversible. Works council checklistThe ten questions the works council will ask before giving its opinion.

Key answers

Three frequent questions.

Short answers, drawn from the pathway above.

What are the obligations before deploying AI in a team?

They depend on the function the system performs, not on the software brand. For all uses, data must be mapped and the need for an impact assessment checked. In France, from 50 employees, the works council (CSE) is informed and consulted when the project changes employment, organisation, working conditions or the monitoring of activity. The risks identified are entered in the risk assessment document (DUERP) where the framework requires it.

Can targets be raised thanks to the time saved with AI?

Not before measuring the whole task, checking and corrections included, task by task and person by person: a positive average can hide negative effects. The share of gains allocated to output, quality, learning and recovery is decided with the team.

Can an employee set aside an AI recommendation?

They must be able to do so without penalty. The flowchart recommends writing down that professional judgement prevails over the AI output, guaranteeing the right to correct, refuse or report without sanction, and naming who can review a decision, within what time.

Scope of these pointers

These pointers restate the flowcharts published on this site, which are written for the French legal framework. They are prevention pointers to adapt to each work situation: they replace neither legal advice, nor the opinion of the occupational physician, nor dialogue with employee representatives.

Update log

  1. First continuous-reading version, drawn from the flowchart updated on 13 November 2026.