EU AI Act

This page states how Workforce AI is classified under the EU Artificial Intelligence Act (Regulation (EU) 2024/1689), what the system is for, and what it does not do.

Last reviewed: 2 October 2026.

Provider

Workforce AI is the provider of this AI system. We are established in the European Union. For questions about the AI Act, or to report misuse, email privacy@workforceai.com.

What the system is for

Workforce AI is an AI-native workforce health administration platform for HR, occupational health and leadership teams. It combines workforce structure, psychosocial and psychometric measurement, risk stratification, EAP/service orchestration, organisational analytics, regulatory reporting and governance in one operating environment. Every administrative capability can be operated manually through the application or conversationally through our specialised AI agents.

Its purpose is organisation-level prevention and support for work-related health risks. The instrument in use today is the BAT-23, which measures occupational burnout. The system helps an employer run assessment cycles, read group results, and plan prevention and support services for groups of employees.

The platform does measure and stratify risk. Fixed scoring rules do that, not AI. Each employee sees their own result, the employer sees group results only, and no output is for a decision about an individual's employment. The sections below give the detail.

How a result is produced, and who sees it

Employees answer the BAT-23, a validated self-report questionnaire, voluntarily and with explicit consent. The Healo platform computes each score and risk tier with fixed BAT-23 scoring rules. These rules are not an AI system. Each employee sees their own result in the Healo app.

The employer sees group results only. A group smaller than the minimum group size is not shown. The minimum is 5, and a customer can set a higher value but not a lower one. When figures shown together could be combined to work out the result of specific people, those figures are hidden as well. The AI assistant applies the same rules to its answers.

What the AI does, and does not do

The AI assistant helps HR to:

set up and run assessment cycles;
explain group results and the cost of inaction for the organisation;
define service rules for cohorts.

It is instructed to ask HR before significant changes, such as launching or closing a cycle.

The AI does not:

compute, analyse or assign a burnout score or risk tier;
score, assess, rank, triage or monitor any person;
decide which named person receives which service;
take any decision about an individual employee.

People who use the AI assistant are told in the product that they are interacting with an AI system.

Uses the system is not for

Do not use the system, or any output of it, for:

performance evaluation;
assessment of conduct or behaviour;
task allocation;
promotion;
pay;
discipline or termination;
recruitment or selection of candidates.

No biometric input

The system takes no biometric input: no camera, microphone, voice, typing-cadence or wearable data. Its only inputs are a self-report questionnaire and typed text.

Our classification

We assess Workforce AI as not a high-risk AI system. It is outside Annex III point 4(b) of the AI Act, for three reasons:

the system evaluates health, not work performance or behaviour;
an individual's result never reaches the employer;
the AI agents work on group results and do not evaluate people.

This is our own assessment. We review it when the product changes, and we update this page.