AI in time tracking
Here’s what it’s all about:
- How AI Can Effectively Complement Time Tracking and Workforce Planning
- COMING SOON: 3 AI Features from ZMI
- How Permissions, Approvals, and Human Oversight Govern AI Use
- What Role Data Protection, the EU AI Act, and Employee Participation Play
- What Companies Should Consider When Implementing AI in Time Management
Artificial intelligence is increasingly finding its way into HR software and is opening up new possibilities for time tracking as well. The goal here is not to completely replace proven processes with AI. Of particular interest are features that support employees and HR managers with recurring tasks, make information more readily accessible, or enable existing data to be used for planning and analysis.
However, particularly in time management, the use of AI is closely linked to issues of data protection, access permissions, and employee participation in decision-making. Working hours, absences, and vacation balances are personal employee data. Companies should therefore not only consider what AI is technically capable of, but also what data it accesses, what actions it is permitted to perform, and how human oversight is ensured.
📣 With this in mind, ZMI will soon be releasing various AI features for time management. These include the ZMI Chatbot, the ZMI Assistant, and an AI-powered absence forecast. In this article, they serve as concrete examples of how AI can support various tasks in time tracking and workforce planning in the future. Learn more.
🤝 How can AI help with time tracking?
The three new features address different tasks. The chatbot provides direct access to information about the software. The assistant works with company data within existing permissions. The absence forecast combines actual and projected absences to assess staffing risk.
ZMI – Chatbot: Get Answers to Your Questions Right Away
The ZMI chatbot answers questions about how to use the software in natural language. Instead of searching for specific terms in manuals, users can, for example, ask how to correct a posting error or enter a holiday. The answer is provided as specific step-by-step instructions with menu paths.
The ZMI manuals and FAQ entries serve as the knowledge base. When this documentation is updated, the chatbot’s knowledge base is also updated with the new versions. This approach differs from a general AI chatbot: the focus is on specific product knowledge.
👉 For HR departments, this is particularly relevant when it comes to recurring operational questions. The AI does not make any personnel decisions or alter time-tracking data. It simply makes existing knowledge more readily accessible.
ZMI – Assistant: Retrieve Data and Perform Tasks
The ZMI Assistant takes it a step further. While the chatbot provides product information, the Assistant is designed as an AI agent with access to ZMI functions. This allows it to work with specific data within the scope of its authorized permissions.
Use cases include, for example, questions about remaining vacation days or absences within a department. CSV exports and searches for specific qualifications at a given location are also possible.
A key feature is the authorization model. The Assistant can only access information that the logged-in user is already authorized to view. Data access is also visible in the activity log.
An additional verification step is provided for write operations. For example, if the Assistant is supposed to enter a specific transaction for multiple employees, it first displays a preview of the changes. The data is not actually written until explicit confirmation is given.
👉 This ensures that the decision to approve remains with the responsible person. This separation between preparatory support and the actual modification of data is particularly relevant for AI functions with direct data access.
AI Absence Forecasting: Detect Staffing Shortage Risks Earlier
The third area of application concerns workforce planning. The AI absence forecast indicates, on a day-by-day basis, the risk of understaffing within an organizational unit. A traffic-light system indicates the risk: green for low, yellow for moderate, and red for high.
The assessment takes into account already approved vacation and sick days, as well as requested vacation days that are still pending. This data is supplemented by a forecast of unplanned absences. This forecast is based on aggregated and anonymized data. Individual profiles or evaluations of specific employees are not intended.
The traffic light icon appears in two places: immediately when submitting a vacation request in Employee Self Service (more on this) and in the monthly view of the shift and vacation planner. This allows managers to jointly review planned absences and the projected staffing risk.
🚨 It’s important to clarify: The forecast serves as a planning tool. It does not predict which specific person will be absent. The risk is assessed at the organizational unit level, not based on the individual behavior of specific employees.
⚠️ Why Data Privacy Is Especially Important in AI-Based Time Tracking
Time and attendance systems contain personal data closely related to the employment relationship. This includes working hours, absences, vacation balances, and, depending on the system, other personnel data. As soon as AI accesses such data, the purpose, legal basis, access rights, and technical processing must be reviewed in the same way as for other processing operations.
The GDPR does not contain a general special rule permitting or prohibiting the use of AI in time tracking. The specific nature of the processing is decisive. If processing—particularly when new technologies are used—is likely to result in a high risk to the rights and freedoms of natural persons, Article 35 of the GDPR requires a data protection impact assessment to be conducted in advance.
👉 The ZMI AI functions are guided by three key principles: Data processing takes place within the EU; the functions are designed to meet the requirements of the GDPR and the AI Act; and predictions of unplanned outages are based on aggregated and anonymized data.
📜 What role does the EU AI Act play?
The EU AI Regulation has generally been in effect since August 2, 2026. Some provisions have been in effect since 2025, and different transition periods apply to certain areas of regulation. Companies should therefore not only check whether a function uses AI, but also determine the specific purpose for which the AI system is used.
This distinction is particularly relevant for HR. Annex III of the AI Act covers, under certain conditions, AI systems in the field of employment and human resources management—for example, systems used for personnel selection, as well as certain systems that assign tasks based on individual behavior or personal characteristics, or that monitor and evaluate employee performance and behavior. Not every AI function within HR software is therefore automatically a high-risk AI system. The purpose and specific use are decisive.
👉 A chatbot designed to explain software functions is therefore classified differently under the regulations than AI that evaluates employees or influences personnel decisions. It is precisely this functional distinction that should be part of the implementation process.
🔍 AI in Time Tracking and the Works Council: What Should You Keep in Mind?
In addition to data protection and the AI Act, employee participation in the workplace plays an important role. Section 90(1)(3) of the Works Constitution Act (BetrVG) explicitly mentions the use of artificial intelligence in the planning of work procedures and workflows. The works council must be informed of this in a timely manner, along with the necessary documentation.
In addition, there may be co-determination rights under § 87 of the Works Constitution Act (BetrVG). § 87(1)(6) of the BetrVG is particularly relevant in the case of technical devices designed to monitor the behavior or performance of employees. Issues regarding working hours and vacation policies are also expressly regulated in Section 87(1). Which specific participation rights apply depends on the employee’s role and the company’s organizational structure.
Furthermore, the Works Constitution Act explicitly takes AI into account when consulting experts. If the works council must assess the introduction or use of AI, consulting an expert pursuant to Section 80(3) of the Works Constitution Act is considered necessary in this regard.
👉 For practical reasons, there are therefore strong arguments for involving employee representatives earlier than just immediately before the technical implementation. A clear description of data sources, access permissions, processing purposes, and control mechanisms facilitates a joint assessment.
✅ What Companies Can Do to Prepare Before Implementing AI in Time Tracking
Companies don’t have to wait until specific AI features are available. Data protection, governance, and permissions can be reviewed even before implementation using a checklist for HR and IT:
🔵 Define specific use cases and distinguish them from general AI requests.
🔵 Determine which personal data is actually required for each use case.
🔵 Review the legal basis and data protection information for the planned processing.
🔵 Determine whether a data protection impact assessment is required under Article 35 of the GDPR.
🔵 Involve the works council or other relevant employee representative bodies at an early stage.
🔵 Review roles and authorization concepts for self-service and HR functions.
🔵 Equip AI-generated text with approval and control mechanisms.
🔵 Create internal rules for the permissible use of AI or supplement existing AI guidelines.
🔵 Define responsibilities for errors, approvals, and the review of AI results.
🔵 Train employees to use AI functions appropriately for their specific areas of application.
The last point is also gaining regulatory significance. Article 4 of the AI Act requires providers and operators of AI systems to take measures to ensure that individuals involved in the operation and use of AI systems possess a sufficient level of AI competence.
👉 This provision has been in effect since February 2, 2025.
🤓 Conclusion: AI supports time tracking, but the responsibility remains with the company
AI can usefully complement time tracking in several ways: It can make product knowledge more accessible, support administrative tasks, and make existing data usable for workforce planning. It is crucial that permissions, data access, and approvals remain clearly regulated.
Especially in time management, the focus should therefore not be on how much can be automated using AI. It is more important to determine for which tasks AI offers concrete benefits and what data is actually required for those tasks. Data protection, the AI Act, and employee participation in decision-making must therefore be integrated into the implementation process from the very beginning.
👉 With ZMI – Chatbot, ZMI – Assistant, and the AI-powered absence forecast, ZMI addresses three different areas of application. These solutions demonstrate how access to knowledge, administrative support, and predictive planning can be integrated into existing time management processes.
💡 FAQ: AI in Time Tracking
Is AI-powered time tracking possible in a way that complies with the GDPR?
Yes, the GDPR does not generally prohibit the use of AI in time tracking. However, companies must assess, among other things, the legal basis, purpose, data minimization, access rights, and technical processing for each specific use case. If a high risk is anticipated, a data protection impact assessment may be required under Article 35 of the GDPR.
What is the difference between a chatbot and an AI assistant in HR software?
A chatbot can access a defined knowledge base and, for example, answer questions about how to use a software program. An AI assistant can also work with approved company data and functions. The ZMI – Assistant, for example, enables queries and exports, as well as write operations upon explicit approval.
Can an AI-powered absence prediction tool evaluate individual employees?
That depends on the specific technical implementation. ZMI’s AI absence forecast does not create individual profiles. The forecast of unplanned absences is based on aggregated and anonymized data and represents the staffing risk for an organizational unit.
Does the works council need to be involved when AI features are used in time tracking?
Depending on the application, different participation rights may apply. Section 90(1)(3) of the Works Constitution Act (BetrVG) explicitly mentions AI in the context of planning work procedures and workflows. In the case of technical devices for monitoring behavior or performance, § 87(1)(6) of the BetrVG may also be relevant. The specific scope of employee participation must therefore be assessed on a case-by-case basis.
Jonathan Martin is a managing partner at ZMI GmbH. For many years, he has been working in the fields of time tracking, HR software, access control, and the digitization of HR processes. In his articles, he provides practical insights into current developments, legal requirements, and digital solutions for businesses.



