Eighty per cent of German companies use or are testing artificial intelligence (AI) tools in their occupational pension schemes, according to a survey by WTW Germany.

Its Artificial Intelligence in Occupational Pension Scheme Administration 2025 report found that this includes digital assistants, chatbots and knowledge management tools.

The study, undertaken this Autumn, included responses from 34 managers across 24 companies and in 11 different industries. This represents more than 2.2 million employees and pensioners.

WTW Germany senior director outsourcing, Dr Franziska Kühnemund, said companies are at a “turning point” as AI is “no longer a topic for the future” and is more widely used than a year ago.

"At the same time, however, it is also becoming apparent that transparency, traceability and regulatory requirements define the limits of what is feasible, especially in more complex processes,” he added.

Around a quarter of companies use chatbots to answer simple HR or occupational pension scheme queries. Just under a fifth are also driving forward AI projects in knowledge management to facilitate access to knowledge and make expertise available in the long term.

However, more complex applications, such as those used in administration, data verification, or reporting, are used by only a small minority.

WTW senior director outsourcing, Dr Claudio Thum, noted that handling more advanced cases requires “interdisciplinary teams that combine occupational pension expertise with technological competence”.

“Many companies reach their limits here,” he stated.

Around half of companies are also developing plans to broaden their use of AI-enabled services within occupational pension schemes, with digital assistants for handling enquiries emerging as a key priority.

More than 60 per cent of firms looking to expand their AI capabilities are considering such tools, often enhanced with advisory features. Interest in applying AI to knowledge management is similarly strong, with around 45 per cent of companies exploring this area.

In addition, many organisations are taking initial steps towards automating processes or using AI for data-driven decision support, typically beginning with small pilot projects.

The majority of companies continue to see AI as offering clear opportunities. Around two-thirds expect efficiency gains, while half of those surveyed see potential to enhance both service and process quality.

A similar proportion believes that self-service options for applicants and pensioners could be significantly improved through AI-led solutions.

Expectations also extend to easing workforce pressures: just under half of respondents anticipate a noticeable reduction in routine tasks for skilled workers.

"AI can significantly improve the quality of occupational pension services, for example, through faster response times and more consistent information," emphasises Kühnemund, adding: "However, in order to fully exploit the potential, data quality and governance must be systematically developed further."

Despite strong interest and clear potential, companies also face a number of barriers. Data protection remains the dominant concern for a clear majority, followed by worries about output quality and the limited explainability of AI expenditure.

Around four in 10 companies also point to regulatory uncertainty as a reason for caution. Economic challenges such as tight IT budgets and a lack of interdisciplinary resources further complicate implementation efforts.

In the short term, most companies expect costs to rise as they establish the necessary infrastructure, expertise and processes. Over the longer term, however, respondents anticipate that AI will begin to reduce operating costs, even if opinions differ on whether these efficiencies will materially lower the overall expense of providing occupational pension services.

Looking ahead, companies expect the digital transformation of occupational pensions to progress gradually over the coming years. AI is expected to speed up processes, increase quality and improve service, but few foresee disruptive change in the near term.

---