Artificial intelligence is gradually gaining ground in human resources departments across Germany. New Bitkom figures show that companies are increasingly using AI for employee training, employment references, onboarding and internal HR services. At the same time, concerns over sensitive personal data and uncertainty surrounding the EU AI Act continue to slow adoption.

Artificial intelligence has already transformed numerous corporate functions, yet its uptake in human resources remains comparatively cautious. New figures from German digital association Bitkom show, however, that adoption is beginning to accelerate: 14 percent of German companies already use AI to draft employment references, while more than half are open to doing so in the future.

The findings are based on a representative survey of 602 German companies with at least three employees, conducted by Bitkom Research between calendar weeks 18 and 25 of 2026.

The results point to a broader development. AI is increasingly moving beyond experimental applications and into everyday HR processes – from personalised employee training and onboarding to internal service requests and even the assessment of workloads and employee performance.

At the same time, the growing role of automated systems in personnel management raises sensitive questions surrounding data protection, transparency, governance and regulatory compliance.

HR Remains a Sensitive Area for AI Adoption

Compared with other corporate functions, HR poses particular challenges for the deployment of artificial intelligence.

Personnel departments handle large volumes of sensitive information, ranging from employment records and salary information to performance assessments and personal data. As a result, companies face a significantly higher threshold when introducing AI systems into HR processes.

“Especially in human resources, the barrier to using AI is higher than in many other areas of a company,” says Dr Bernhard Rohleder, CEO of Bitkom.

According to Rohleder, many businesses remain cautious partly because they fear breaching data protection requirements when processing sensitive employee data.

Another source of uncertainty is the practical implementation of the EU AI Act. Certain AI applications used in employment and workforce management fall into categories subject to particularly stringent regulatory requirements. For companies considering automated decision support in recruitment, performance assessment or other personnel processes, compliance therefore becomes a central design requirement rather than an afterthought.

Personalised Training Leads AI Adoption in HR

The most widespread HR application identified in the Bitkom survey is individual employee training.

Currently, 16 percent of German companies use AI for personalised professional development, up from 12 percent in 2024. A further 58 percent say they could imagine using AI for this purpose in the future.

The appeal is straightforward. AI systems can potentially tailor learning materials, recommendations and training programmes to individual employees, making corporate learning more adaptive and accessible.

The figures indicate significant potential for further expansion: almost three quarters of companies either already use AI for individual training or are open to doing so.

For HR departments, this could shift training away from standardised programmes towards more personalised learning environments – provided that the quality of recommendations, employee data protection and human oversight remain properly managed.

One in Seven Companies Uses AI to Draft Employment References

AI-generated or AI-assisted employment references have also become an established use case.

According to Bitkom, 14 percent of companies already use artificial intelligence when drafting employment references. Another 52 percent can imagine doing so.

Interest has increased markedly compared with 2024, when 45 percent of companies said they could envisage using AI for this task.

Employment references are particularly well suited to language-based AI assistance because substantial parts of the process involve structuring information and formulating standardised professional language.

Generative AI can help HR professionals prepare drafts, ensure consistent wording or accelerate repetitive writing tasks. Human review nevertheless remains crucial, particularly because employment references can have significant consequences for an employee’s future career.

AI may therefore reduce administrative effort, but responsibility for the accuracy and appropriateness of the final document remains a human task.

AI Is Becoming an Onboarding Assistant

Another rapidly developing field is employee onboarding.

Currently, 14 percent of companies use AI to support the induction of new employees, while another 36 percent could imagine introducing such systems.

That represents clear growth compared with 2024. At that time, 11 percent already used AI for onboarding and 28 percent were open to doing so.

AI-based onboarding assistants can provide newcomers with information about processes, organisational structures or frequently asked questions. Chatbot-style systems can make this information available at any time rather than requiring HR staff or colleagues to repeatedly answer routine queries.

“Anyone who is new to a job has lots of questions. A well-trained AI can provide answers, help with orientation and is always available,” Rohleder explains.

The potential benefit is not necessarily to replace personal onboarding. Instead, routine information can be handled digitally, leaving more time for human interaction in areas where individual support is genuinely required.

This distinction is likely to become important as organisations decide which HR interactions should be automated and which should remain explicitly personal.

Internal HR Services Are Moving Towards Automation

Companies are also beginning to automate internal personnel enquiries.

Thirteen percent of surveyed businesses already use AI to answer or process internal HR requests, while a further 53 percent can imagine doing so in the future.

This represents one of the areas with the highest potential acceptance rates in the survey.

Typical applications could involve recurring administrative queries concerning internal processes, payroll-related information or standard HR procedures. By automating routine questions, HR teams could reduce administrative workloads and concentrate on more complex cases.

However, the usefulness of such systems depends heavily on the quality of the underlying information. Incorrect or outdated answers regarding employment conditions, payroll or company policies can quickly create operational and legal problems.

Reliable data sources, clear escalation mechanisms and human oversight therefore remain essential.

AI Is Increasingly Being Used to Assess Workloads

A more sensitive trend emerges in the use of AI to evaluate employee workloads.

Here, adoption has more than doubled within two years.

In 2026, 13 percent of companies use AI in connection with assessing employee workloads, compared with just 6 percent in 2024. Another 28 percent are open to doing so, up from 19 percent two years earlier.

Such systems could potentially identify workload imbalances, patterns of overcapacity or indications that teams are under excessive pressure.

Yet workload analysis also illustrates why AI in HR requires particularly careful governance. Depending on the data analysed, systems could potentially draw on information about working patterns, productivity or employee behaviour.

The transition from supportive analytics to employee monitoring can therefore become difficult to define.

Businesses introducing such technologies must consequently consider not only technical performance but also proportionality, transparency and the legitimate expectations of employees.

Performance Assessment Raises Even Greater Questions

The use of artificial intelligence in employee performance assessment moves automation into an even more consequential area.

According to the survey, 12 percent of German companies already use AI for evaluating work performance, with another 27 percent considering future use.

AI can potentially process large quantities of performance-related data and identify patterns that might otherwise remain hidden. However, the more directly an algorithm contributes to assessments affecting careers, remuneration or employment opportunities, the more important issues such as bias, explainability and human oversight become.

A system may generate a score or recommendation, but that does not automatically make the result objective.

The quality of AI-based evaluation depends on the data used, the criteria selected and the organisational assumptions embedded in the system.

For security, compliance and HR leaders alike, this means that deploying AI for workforce decisions requires considerably stronger governance than using a language model simply to draft internal text.

The EU AI Act Changes the Risk Equation

The figures illustrate why regulation is becoming increasingly relevant to HR technology.

AI-based personnel applications range from relatively low-impact assistance, such as drafting text or answering routine questions, to systems that can influence hiring, evaluation or working conditions.

These applications cannot all be treated in the same way.

The closer AI moves towards decisions that affect an individual’s employment prospects or working life, the greater the need for structured risk management, documentation, transparency and human control.

For businesses, the challenge is therefore no longer simply whether AI can improve an HR process.

The more important questions are which data the system processes, how decisions are generated, who remains accountable and whether the application complies with applicable regulatory and data-protection requirements.

From HR Efficiency to AI Governance

Bitkom’s figures demonstrate that AI is steadily becoming part of German personnel management.

Personalised training currently leads adoption at 16 percent, while employment references and onboarding each stand at 14 percent. Internal HR requests and workload assessment follow at 13 percent, with AI-supported performance evaluation already used by 12 percent of companies.

Perhaps more significant than current adoption rates is the level of interest among businesses that have not yet implemented these technologies. Across several applications, large shares of respondents say they can imagine using AI in the future.

This suggests that the next phase of AI adoption in HR may develop rapidly.

For companies, however, the decisive factor will not simply be how many processes can be automated.

In HR, artificial intelligence operates particularly close to people, their careers and their personal information. Successful implementation will therefore depend on combining efficiency with data protection, regulatory compliance, transparency and meaningful human oversight.

The technology may increasingly draft documents, answer routine questions and analyse workforce data. Responsibility for the consequences, however, will remain with the organisation deploying it.

Survey methodology

The findings are based on a representative survey conducted by Bitkom Research on behalf of German digital association Bitkom. A total of 602 companies in Germany with at least three employees were surveyed by telephone between calendar weeks 18 and 25 of 2026.

Companies were asked whether they currently use artificial intelligence for a range of HR tasks or could imagine doing so in the future.