GenAI in occupational pensions: the greatest potential lies hidden

When people talk about GenAI in the insurance industry, the focus is often on high-profile use cases: digital assistants in customer portals, automatically generated correspondence, and sales support. In occupational pension schemes (bAV), however, the greatest potential for improvement often lies where hardly anyone looks – in the depths of group contracts, in the daily volumes of documents, and in the many small verification steps that specialist departments still laboriously carry out manually today.

In the occupational pension collective business in particular, several challenges converge: complex contracts, numerous stakeholders (employers , intermediaries , consortia, employees) and a high level of manual administrative effort. At the same time, employers and their employees are increasingly demanding that occupational pension schemes are not only technically sound, with new applications processed swiftly and contract amendments implemented immediately, but also provide a positive digital experience.

Rather than testing “a bit of AI here and there”, it is worth taking a targeted look at three levers that can deliver rapid, tangible results: intelligent input management as a foundation, GenAI-supported capture of collective agreements, and automated deadline and data checks across the existing portfolio.

AI in input management: from inbox to process control

Anyone looking at day-to-day occupational pension management in an insurance company will quickly come across the inbox: documents arrive in a wide variety of formats and qualities. Collective agreements, new registrations, change notifications for employees, enquiries from employers and intermediaries, claims with extensive documentation – all of this must first be understood, correctly categorised and then forwarded to the appropriate department or a specialist working group.

Gone are the days when every single letter was opened, read, marked and routed to the relevant department by the insurer’s mailroom. Yet, at , this supposedly digital process still involves media breaks, manual intermediate steps and sources of error.

With AI-supported input management, it is precisely these steps that are consistently automated: the solution ‘reads’ incoming documents and recognises what they concern: new policies, changes to existing policies, premium adjustments, claims, etc. The document is then automatically assigned to the correct process.

The data relevant to the respective process is then extracted from the document: employer name, contract number, personal details, type of change requested, etc. This also works with documents that are not perfectly structured, such as scans or spreadsheets. The information obtained is processed and formatted in a way that specialist systems can process immediately. A ‘document’ is thus transformed into a data record – an important basis for dark processing. The structured data is fed directly into portfolio, quotation or workflow systems. Where it makes business sense, an intermediate step for business verification is retained.

The difference from traditional mail handling is clear: instead of managing documents and emails, controllable, structured information is available from the outset, which can be processed automatically. This is the crucial first step towards further value creation in occupational pension schemes – particularly in relation to group or consortium contracts.

Capturing collective and consortium agreements with GenAI: from PDF to a robust data model

Collective agreements are far more than just a ‘document in an attachment’. They are the binding set of rules governing cooperation with the collective or consortium agreement partner.

In practice, contract documents are available as PDFs, Word files or scans. Staff have to leaf through collective or consortium agreements, including addenda, page by page. The relevant content of the collective or consortium agreement is often entered manually into the partner, portfolio or management system. Where anything is unclear, staff have to make enquiries or discuss the matter internally.

This process is not only slow but also risky. Individual interpretations or transcription errors can have negative consequences, even if it is ‘only’ a matter of additional error correction.

GenAI can achieve two things simultaneously here:

Automatically understand contracts

Instead of manually entering collective and consortium agreements page by page, GenAI extracts the relevant content directly from physical or digital documents:

• Contracting parties
• Start date of the collective or consortium agreement
• Syndicate shares
• Insured groups and areas of coverage
• Premium and benefit logic
• Agreements regarding medical examination
• Rights and obligations
• Special provisions

The information is stored in a structured manner within the system and is immediately available to the relevant departments. Manual ‘typing out’ and interpretation of contract content by administrative staff is largely eliminated, which makes data entry significantly more efficient, saves time and frees up capacity for complex cases. At the same time, automated preparatory work reduces data entry and verification times, thereby shortening processing times in portfolio management. This also creates a consistent, analysable database that enables better control of portfolios.

Stabilising bulk processing

In group business in particular, heterogeneous input files (Excel, CSV, PDF) are the natural enemy of high dark processing rates.

GenAI can normalise these inputs: semantically recognise fields, assign them correctly, flag inconsistencies and automatically generate error logs with specific correction suggestions.

This makes ‘bulk processing’ more robust, reduces rework and makes group business scalable.

Automatically monitor deadlines and data in group contracts

Once a collective or consortium agreement has been drawn up, the actual long-term relationship begins. Spanning years – often decades.

In many collective and consortium contracts, terms and conditions are directly linked to volumes: the employer only receives particularly attractive terms if certain thresholds are reached within a defined period – for example, a minimum number of insured persons or a specific number of contracts taken out. If this threshold is met, special terms apply; if it is not met, less favourable conditions apply, such as higher premiums or the loss of discounts in employees’ individual policies.

Today, this is often managed very manually. Someone in the specialist department makes a note of the agreement, enters the deadline as a reminder in a calendar or system, and produces a collective evaluation at the end of the period. Based on this analysis, a manual check is carried out to determine whether the requirements for the special terms have actually been met. If this is not the case, the measures stipulated in the contract must be implemented – such as switching the affected individual policies to the standard, less favourable terms. In practice, however, it is just as common for such deadlines not to be recorded at all, or only incompletely, in the quiet hope that the agreed thresholds will be met. If they are then actually missed, special terms may continue to apply incorrectly for years. The financial consequences can be significant and, in extreme cases, harm the insured community because premiums have been calculated too low.

This is precisely where GenAI comes in as an automated “deadline monitor” and analytical tool. Based on the structured contract data, the solution identifies – as soon as the group contract is created – which volume-dependent special terms have been agreed and which deadlines are associated with them. These conditions are no longer maintained in individual lists or calendars, but are systematically stored and monitored within the administration system. On the relevant cut-off date, the solution automatically accesses the current portfolio and transaction data, independently checks whether the agreed number of insured persons or contracts has been reached, and generates a group evaluation.

The result is a clear, technically comprehensible audit report with an unambiguous conclusion: conditions met or not met – including specific recommendations for action for the relevant department, such as ‘continue special terms’ or ‘switch individual contracts to standard terms’. The responsibility for the decision remains with the insurer, but GenAI provides the complete, structured data basis and ensures that no volume-dependent special terms and conditions continue “by mistake”. This reduces financial risks, protects the insured community and, at the same time, makes the review process significantly more efficient and transparent.

What adesso achieves in occupational pensions with GenAI

adesso has been supporting insurers for many years in the digitalisation of occupational pension schemes – from analysing existing group processes and introducing AI-supported input management to automated contract and deadline checks. In projects with major life insurers, we have integrated GenAI solutions into the existing system landscape in such a way that unstructured documents are transformed into reliable data models and volume-based special terms are systematically monitored. On this basis, we work with our clients to develop bespoke business and IT architectures, implement solutions effectively and support organisational change – ensuring that GenAI in occupational pension schemes is not just a pilot project, but a genuine driver of efficiency and control.

Conclusion

GenAI is not a ‘nice-to-have’ in occupational pensions, but a genuine competitive factor. Where files, Excel spreadsheets and manual collective contract analyses still set the pace today – in document intake, when recording complex collective agreements and with volume-dependent special terms – GenAI can bring about tangible changes to processes: incoming data becomes actionable information, contracts become clearly structured control metrics and critical deadlines become automatically monitored guardrails.

The result: faster decisions, less risk, greater transparency – and an occupational pension scheme that, for employers and intermediaries, looks less like ‘administration’ and more like a modern, reliable partnership. Insurers who leverage these opportunities now will secure a clear advantage: professionally, economically and in their market presence.