Urząd Komisji Nadzoru Finansowego to expand AI platform with security integration

Urząd Komisji Nadzoru Finansowego to expand AI platform with security integration

Urząd Komisji Nadzoru Finansowego is procuring hardware, software and training to expand its AI platform with 36-month support and Polish-language models.


More on Spotlight   Back to News & Insights

Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.

Poland’s financial supervisor has launched a comprehensive plan to grow its in‑house artificial intelligence capability. The programme will add equipment, software and support to an existing platform, integrate new security modules with the authority’s systems, and train staff. It will also produce two concrete Polish‑language models: speech‑to‑text with summarisation, and a RAG‑based chatbot. Read the notice: Expansion of AI Platform.

What the authority is buying

The contract covers end‑to‑end delivery and implementation. The selected supplier will prepare the technical design, deliver and install equipment, provide licences, expand the platform in the authority’s environment, and support testing, training and documentation.

Key elements include:

  • Development of a Technical Project to guide delivery and integration.
  • Supply and on‑site installation of equipment.
  • Licences or subscriptions for perpetual software licences with manufacturer technical support at a 5/8 standard for 36 months from delivery.
  • Licences or subscriptions for container orchestration on all cluster servers listed in the agreement’s annex, with manufacturer support for 36 months.
  • Implementation of the platform expansion and integration of its security modules with the authority’s security systems.
  • Comprehensive documentation: as‑built, administrative, user documentation, and a post‑expansion technical card.
  • Preparation of test scenarios, assistance during internal and acceptance testing, and test reporting.
  • Support during security testing and remediation of any vulnerabilities identified.
  • Training for employees under the agreement’s training provisions, plus 20 annual e‑learning subscriptions from the platform producer (administration, usage, and model lifecycle) in Polish or English.
  • Up to 480 man‑hours of consultations on administration and use.
  • Access, throughout the 36‑month support period, to a catalogue of ready‑made components for building AI solutions as set out in the agreement.
  • Workshops on building and deploying models on the expanded platform, delivering two models: Polish speech‑to‑text with summarisation, and a Polish chatbot using Retrieval Augmented Generation.

Some technical specifics, such as the exact cluster server inventory and the component catalogue requirements, sit in the referenced annexes and are not disclosed in the public summary.

Security, testing and operability

The notice places clear weight on secure implementation and operational readiness. The contractor must integrate the expanded platform’s security modules with the authority’s existing security systems. It must also assist with security tests and fix vulnerabilities found during those tests.

There is a structured testing regime: internal tests and acceptance tests must be designed and supported, with scenarios prepared and outcomes documented. This is in addition to the security testing requirements, which aim to validate both resilience and compliance before the expanded platform is handed over to daily operations.

On the operations side, the contract embeds a long tail of support and enablement. Manufacturer support at a 5/8 standard and the 36‑month term apply to core software and to the container orchestration layer. The package also includes consultations, formal training and e‑learning subscriptions, and access to a catalogue of components to help teams assemble solutions quickly and consistently.

Polish‑language use cases at the core

The workshops will deliver two concrete model implementations:

  • A speech‑to‑text model for Polish that generates a transcription from a recording and then produces a summary.
  • A chatbot in Polish for a selected topic area, using Retrieval Augmented Generation to ground responses in curated sources.

These are practical, language‑specific capabilities that point to near‑term applications in document handling, case triage and information retrieval. The inclusion of both model building and deployment during the workshops suggests the authority wants working exemplars embedded into its environment, not just proofs of concept.

Container orchestration and platform foundations

The requirement for container orchestration licences across all cluster servers indicates a commitment to running AI workloads in a managed, scalable environment. While the notice does not name the stack, the combination of orchestration, security integration and documentation signals a move toward standardised, maintainable operations for model training and inference.

Access to a catalogue of ready‑made components over three years should shorten delivery times for new use cases. Combined with training and consultations, the authority appears to be building internal capability to design, deploy and run AI services on a steady footing.

How this fits a wider public‑sector push

The move sits alongside a growing number of AI and infrastructure procurements across Poland’s public sector.

Also in October 2025, Narodowy Bank Polski launched a procurement to acquire an AI environment covering servers, a management platform, software, technical support and implementation services.

In June 2025, the Ministerstwo Rodziny, Pracy i Polityki Społecznej sought delivery, installation, configuration and commissioning of a complete AI platform, including a server and software for operating AI models.

In April 2024, Centralny Ośrodek Informatyki tendered infrastructure and software for production environments, including expansion of OpenShift platforms, alongside Oracle and IBM software with support.

In June 2025, Polska Agencja Żeglugi Powietrznej went to market for the delivery and implementation of an AI computing cluster with technical support and training.

In August 2025, Naukowa i Akademicka Sieć Komputerowa sought delivery, installation, configuration and maintenance of a research platform for AI, covering server and network hardware, warranty and support.

On the application side, in July 2025 the City of Poznań planned a SaaS rollout that included voicebot and chatbot tools, AI and CRM components, and training and support for customer service (Miasto Poznań - Urząd Miasta Poznania). And in February 2025, Śląski Uniwersytet Medyczny w Katowicach sought AI chatbot management software with analysis, design, training and support.

Together, these projects show a pattern: public bodies are pairing compute and orchestration foundations with targeted AI applications, training and long‑term support.

What to watch

The authority’s notice sets clear parameters for what must be delivered and how it will be tested. Details on the cluster footprint and the component catalogue sit in annexes, so the exact scale and available building blocks will emerge later.

Points to monitor include:

  • How the container orchestration licences are deployed across the cluster estate described in the agreement.
  • The integration depth achieved between the platform’s security modules and existing systems.
  • The outcomes of acceptance and security testing, which will shape go‑live.
  • How the training, e‑learning and consultations translate into sustained in‑house capability.
  • The performance and adoption of the two Polish‑language models delivered through the workshops.

With 36 months of support built in, this expansion is set up as a multi‑year operational effort rather than a narrow pilot. The next milestones will be the approved technical design, the installation of equipment and software, and the first results from testing and training.

Urząd Komisji Nadzoru Finansowego to expand AI platform with security integration

Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.