Poland’s Social Insurance Institution (ZUS) and artificial intelligence: how AI is changing the social insurance system

Poland’s Social Insurance Institution (ZUS) and artificial intelligence is no longer just a future-oriented topic in public administration. In Poland, it is becoming a real part of how social insurance matters are handled in practice. The Social Insurance Institution (ZUS) is no longer communicating only general digitalisation plans. It is now pointing to specific implementations, pilot projects and rollout schedules for solutions based on AI. At the same time, ZUS clearly emphasises that artificial intelligence is meant to support employees, not to make decisions independently in customers’ cases.

From the perspective of entrepreneurs, this means one thing above all: contact with ZUS will become increasingly digital, automated and procedurally faster. However, this does not change the basic obligations of contribution payers or the rules of responsibility for the accuracy of documents and settlements. AI may improve service, data organisation, communication and document analysis, but it does not replace the obligation to register employees correctly, submit data on time and settle social security contributions properly.

---

Where ZUS stands today

ZUS confirms that some solutions using AI elements are already in operation, while others remain at the pilot stage, are being prepared for implementation or are still under conceptual development. In official statements, the institution explains that when building such tools it takes into account data security, privacy protection, model transparency, resilience to misuse, legal compliance and human oversight. ZUS also points out that each concept is assessed in terms of compliance with the General Data Protection Regulation (GDPR), the AI Act, IT architecture and the risk of incorrect answers.

This is an important signal for businesses, because it shows that public administration in Poland is not treating AI solely as a tool for automating contact with clients. In practice, the change is broader. It concerns the wider reorganisation of how the authority works: from sorting applications and correspondence, through document analysis, to the development of electronic channels and a more transaction-based service model within eZUS.

---

What is already operating at Poland’s Social Insurance Institution (ZUS)

The most concrete officially confirmed example is a system supporting the handling of reports submitted by employees registered in the KSI User Portal. ZUS indicates that trained algorithms recognise and categorise the most common types of reports, which means that around 25% of them are automatically directed to the appropriate support groups without going through the first support line. In addition, algorithms are already being used to recommend solutions for the most common problems.

For entrepreneurs, this is not yet a function directly visible to the contribution payer or insured person. However, it still has practical importance. It shows that ZUS is introducing AI first where it can accelerate internal processes and reduce repetitive work, without moving into the area of automatically deciding customers’ rights and obligations.

---

What the Social Insurance Institution in Poland plans to implement next

The most important project that may have a direct impact on the efficiency of proceedings concerns the initial analysis of documentation for medical examiners. According to official information, the system is intended to analyse collected documents, organise them chronologically, group them, identify relevant medical facts and make it easier to find information in a virtual file. ZUS has stated that the system has completed the pilot stage, and full production implementation is planned for Q2 2026.

A second solution is a translator for foreign-language documents in a text-to-text model. It is intended to support employees in translating documents submitted by customers, including in benefit-related cases. ZUS states that this system is currently at the pilot stage.

From a business perspective, both projects may result in:

• faster analysis of documents in benefit-related cases,
• a lower burden on employees performing technical tasks,
• more efficient handling of foreign-language documentation,
• better access to information during proceedings.

At the same time, it should be stressed that these are still solutions being implemented or finalised, not yet a fully established operational standard across the entire system. For that reason, businesses should not automatically assume that every matter handled by ZUS will already be processed faster solely because AI is being used.

---

Other AI solutions currently being tested by ZUS

Apart from projects with a defined timetable, ZUS is also carrying out pilot work and conceptual development on several tools that may be particularly relevant from the perspective of contribution payers and individuals dealing with the authority.

Voicebot for booking appointments

ZUS has tested voicebot functions intended to allow appointments to be booked automatically by phone, 24 hours a day, 7 days a week, with integration into the appointment calendar in eZUS and SMS confirmation. However, conceptual work on introducing this solution is still ongoing.

System supporting email handling

ZUS handles more than 1.3 million email enquiries per year, and that volume is one of the main arguments for implementing a system that analyses the content of messages and proposes a response template to an employee. The employee would still approve or modify the final response. This solution has completed the pilot stage, but according to the institution’s communication it has not yet been fully implemented.

Assistant for reliefs and remissions

The planned system is intended to support ZUS employees in the initial analysis of documents and in formulating repayment proposals tailored to the customer’s situation. From the perspective of entrepreneurs, this is one of the more interesting projects because it concerns debt-related matters, including contribution arrears, payment reliefs and remissions. For now, however, it is still at the stage of pilot preparation.

Intelligent Knowledge Base

This is an internal solution for ZUS employees, equipped with an intelligent search engine and an AI assistant designed to summarise, compare and condense documents. In practice, it may improve the consistency of answers and reduce the time needed to search for information in the institution’s dispersed resources. Documentation of requirements is currently being prepared.

---

Will AI at Poland’s Social Insurance Institution (ZUS) make decisions instead of humans?

The official answer is no. ZUS explicitly states that its general rule is to use AI to support human decisions, not to make automated decisions independently. This is highly important from both an organisational and legal perspective.

It also matters in the context of personal data protection. The General Data Protection Regulation (GDPR) provides protection against decisions based solely on automated processing where such decisions produce legal effects for an individual or similarly significantly affect them. For that reason, a model in which AI prepares an analysis, suggests a response or organises documents, while the final decision remains with a human, is a safer regulatory approach for a public social insurance system.

---

The AI Act and ZUS: what it means for businesses in Poland

The issue of AI at ZUS should also be assessed through the lens of EU law. The AI Act is being applied in stages. Some provisions started to apply from 2 February 2025, while the general application date for many key obligations is 2 August 2026. Some provisions for high-risk AI systems will apply even later.

For entrepreneurs, this means that while ZUS is testing and implementing AI tools, it must also build them in a way that complies with requirements relating to security, accountability, transparency and human oversight. As a result, the pace of implementation in areas involving customer data will naturally be more cautious than in internal administrative processes.

---

What this means for entrepreneurs and contribution payers

The main conclusion is practical: AI at ZUS does not currently change the rules of settlements, but it may change the way businesses interact with the authority and the speed at which some matters are handled. From a company’s perspective, it is worth preparing for several consequences.

1. Data and document quality will matter even more

The more digital and automation-supported processes become, the more important the quality of documents, completeness of data and consistency of information submitted to ZUS will be. This applies both to benefit-related documentation and to correspondence, applications and debt-related matters. Poorly described, inconsistent or incomplete data will still create a risk of delays or the need for additional clarification.

2. Faster processing does not remove responsibility

The fact that ZUS is developing AI tools does not mean responsibility for the correctness of settlements is transferred to the authority. Entrepreneurs still remain responsible for proper registrations, contribution settlements, deadlines and the accuracy of the data entered into the system.

That is why, in companies that want to reduce the risk of mistakes in this area, well-organised payroll in Poland and HR administration are especially important.

3. Contact with ZUS will become more digital

The development of eZUS, changes to the portal interface, the expansion of electronic functions and the planned introduction of bots all indicate that communication with the authority will increasingly take place through remote channels. For companies, this is a strong reason to organise account permissions, document workflows and internal responsibility for contact with ZUS.

4. The biggest changes will appear first in service processes, not in substantive law

At present, AI at ZUS mainly affects work organisation, access to information, document analysis and customer service. It does not yet change the legal model for determining entitlement to benefits or contribution obligations under Polish law. For entrepreneurs, this is an important distinction, because it separates service technology from the substance of legal obligations.

---

Should businesses be concerned about AI at the Social Insurance Institution in Poland?

At the current stage, a pragmatic approach is more appropriate than an alarmist one. ZUS is developing AI primarily to relieve employees, improve the flow of information and shorten the time needed for repetitive tasks. The institution itself points out that one of the reasons for accelerating this work is the need to support areas under staffing pressure, including medical certification. At the same time, it emphasises the need for human oversight, legal compliance and caution when dealing with customer data.

From a market perspective, this suggests a gradual shift in the operating standard of public administration in Poland rather than a sudden revolution. Simpler and more repetitive matters are likely to speed up first, such as booking appointments, handling some correspondence or retrieving information. More legally sensitive and operationally complex areas will be introduced more slowly.

---

Key conclusions for businesses

Three key conclusions can be drawn.

First, ZUS is genuinely implementing solutions with AI elements rather than merely discussing them.

Second, the dominant use cases today are those that support employees and internal processes, not systems that automatically resolve customer cases.

Third, for entrepreneurs, the key issue is not so much interest in the technology itself as organisational readiness: orderly documentation, accurate data, efficient information flow and properly managed payroll Poland and HR administration.

---

In practice, companies that already maintain high-quality HR and payroll processes, timely settlements and consistent communication with public authorities will be best placed to benefit from the further digitalisation of ZUS. In that context, support with payroll settlements and contact with the Social Insurance Institution (ZUS) can help reduce operational risk and better prepare the company for further administrative changes in Poland. Contact getsix®.

---

If you have any further questions or require additional information, please contact your business relationship person or use the enquiry form on the HLB Poland website.

Contact us »

***