Training
SKILLS DEVELOPMENT PROGRAMME. AI training for public administration. Practical. Safe. Human-centred
The programme combines everyday AI practice, security and law, change implementation and collaboration between institutions. Each course can be ordered separately or combined into a coherent development path
Programme facts
- 8 courses
- ISO 42001 — full compliance
- 3 × 80 min — convenient blocks
- Mode — Online
- Group — 8–18 people
- PLN 0 — up to 5 people
Skills that translate into action
Almost half of public-administration staff already use generative AI at work, 87% want training and 70% cite missing internal procedures as an implementation barrier (NASK, 2026). The Ministry of Digital Affairs guide says institutions should train staff in how AI works, its limitations and safe use
What makes the courses different
For public administration
Exercises reflect administrative work and the four NASK concepts: citizen chat, content creation, RAG search and AI agents
A practical outcome
Every course produces a document the institution can use: a checklist, oversight map, tool register, pilot charter or action plan
Manageable workload
Up to four hours of learning in three 80-minute blocks with breaks
Tool-independent
We teach capabilities and judgement, not one product; exercises can use existing tools and Polish PLLuM and Bielik models
The voice of officials
6,093 officials told us what they need. The programme responds point by point
Based on the NASK report “AI in public e-administration—the perspective of officials and institutions” (March 2026; CAWI N=6,093 and 24 focus groups, commissioned by the Ministry of Digital Affairs) and the Ministry’s public-administration AI guide. Every figure below maps to a course
87% want training in AI tools
The most frequently reported need. The Ministry guide adds that institutions must provide training on AI operation, limitations and safe use
Related courses: 01 FOUNDATIONS. + the complete path
46% already use generative AI at work—77% encounter it daily
Mostly for information search and writing, often through public tools and private accounts. Practice has moved ahead of policy
Related courses: 01 FOUNDATIONS, 03 SECURITY
70% cite missing internal rules as a barrier
An institution cannot resolve unclear legislation, but it can remove its own procedural gap in a four-hour workshop
Related courses: 03 SECURITY. → draft policy for internal adoption
72% see the greatest potential in document search and analysis
Natural-language search is the highest-rated concept when answers include a source and date and respect permissions
Related courses: 02 KNOWLEDGE. → metadata-based source catalogue
81% want routine work automated so they can focus on expert tasks
Yet agents raise questions about existing systems, liability and whether oversight costs more than it saves
Related courses: 05 AUTOMATION, 04 PILOT
91% oppose algorithms replacing people in citizen decisions
This is a boundary condition, not resistance. The programme is built around human oversight (AI Act, Article 14): information provider, not decision-maker
Related courses: 06 CITIZEN SERVICE, 03 SECURITY
Sources
- NASK, “AI in public e-administration—the perspective of officials and institutions”, 17 March 2026
- Ministry of Digital Affairs, “Artificial intelligence guide for public administration”
Recommended course order
Select a course to go to its programme
The recommended sequence moves from practical foundations through knowledge, security and pilots to automation, citizen service, change management and inter-institutional exchange
01 Foundations — AI in public administration work
Why: Organise safe everyday use of AI and produce a controlled task map Audience and outcome are described in the full course programme
02 Knowledge — Documents, knowledge and RAG—how AI can help find information in public administration
Why: Prepare documents and sources for reliable, cited AI answers Audience and outcome are described in the full course programme
03 Security — Safe and responsible use of AI—GDPR, the AI Act and internal rules
Why: Establish a shared standard for safe and responsible AI use Audience and outcome are described in the full course programme
04 Pilot — Implementing AI in public administration—from problem to controlled pilot
Why: Move from a problem to a small, measurable and reversible test Audience and outcome are described in the full course programme
05 Automation — AI agents and process automation in public administration
Why: Choose a process and define safe boundaries for AI-agent support Audience and outcome are described in the full course programme
06 Citizen service — AI in citizen service—faster, simpler and inclusive
Why: Design inclusive information services with a clear human hand-off Audience and outcome are described in the full course programme
07 Change — AI without fear or duplicate work—managing change in a public institution
Why: Reduce fear, disorder and duplicate work during adoption Audience and outcome are described in the full course programme
08 Collaboration — AI together—sharing experience and collaborating across institutions
Why: Share experience without copying solutions blindly Audience and outcome are described in the full course programme
Who it is for
Management and secretaries
Start with rules and accountability, then pilot and lead organisational change
This path establishes accountability, selects a safe first pilot and prepares the organisation for change
Recommended courses: 03, 04, 07, 08
Specialist staff and citizen service
Practical AI, document work and information for citizens
This path covers everyday tasks, documents and citizen contact with human control and clear boundaries
Recommended courses: 01, 02, 06, 08
IT and information security
Knowledge sources, permissions, usage rules and controlled automation
This path focuses on knowledge sources, permissions, the tool register and technical preparation for automation
Recommended courses: 02, 03, 05
Training prices
September–October 2026
Special edition: PLN 0 per person. Up to five participants from one institution on each course with newsletter subscription. Additional places use November rates
November–December 2026
- Single course: 1–5; PLN 200 net; PLN 246 gross; base price
- Team package: 6–10; PLN 170 net; PLN 209.10 gross; 15% less
- Larger-team package: 11–18; PLN 150 net; PLN 184.50 gross; 25% less
- Complete 8-course path: 1 person; PLN 1280 net; PLN 1,574.40 gross; 20% less
Training programme
01 Foundations — AI in public administration work
Workshop • 3 × 80 minutes
Who it is for
Specialist staff and department managers who already use AI or are about to start; no technical background required. This course fulfils the AI literacy obligation (AI Act, Article 4) and covers limitations, critical evaluation of results and effective prompting
What participants gain
- Identifies tasks that are and are not suitable for AI
- Writes precise prompts and knows how to improve them
- Uses AI for documents and plain-language content
- Knows which data must not be entered into AI tools and when to stop
- Can demonstrate that a person verified the AI output
Outcome for the organisation: Staff perform common tasks with AI consistently and under control. The institution reduces informal use of public tools and gains a task map and checklist that can support its internal AI policy
Next step: The task map feeds into Course 03 (policy and register) and selection of the first use case in Course 04. Institutions that want to work on their own documents and infrastructure, including with PLLuM or Bielik, can continue in SAIE
Training programme
A. Where AI helps—and where it does not
- Common AI uses in public administration, based on the four NASK concepts
- Why a model can sound confident when wrong: hallucinations, outdated law and AI slop
- Low-risk, conditional and unacceptable tasks; data that must not enter public tools
Outcome: Task map: use / check conditions / do not use, aligned with the Ministry guide
B. Document work and content creation
- Prompt structure: goal, context, audience, format and constraints
- Summaries, analyses and comparisons; simplifying administrative language
- Drafting a reply or note from named sources without inventing legal grounds
Outcome: A working draft marking what AI generated and what a person verified
C. Human verification and decision
- Hallucinations, omissions and false matches in administrative letters
- Checking source, version, date and completeness; when to ask an expert or DPO
- Decision trail: what was accepted, rejected and changed (human oversight, AI Act Article 14)
Outcome: AI-output verification checklist ready for the institution’s internal policy
Training dates
Term 1: 16.09 • śr • 10:00 - 11:20 | 17.09 • czw • 10:00 - 11:20 | 18.09 • pt • 10:00 - 11:20
Term 2: 20.10 • wt • 10:00 - 11:20 | 21.10 • śr • 10:00 - 11:20 | 22.10 • czw • 10:00 - 11:20
Term 3: 16.11 • pon • 10:00 - 11:20 | 17.11 • wt • 10:00 - 11:20 | 18.11 • śr • 10:00 - 11:20
Term 4: 14.12 • pon • 10:00 - 11:20 | 15.12 • wt • 10:00 - 11:20 | 16.12 • śr • 10:00 - 11:20
02 Knowledge — Documents, knowledge and RAG—how AI can help find information in public administration
Workshop • 3 × 80 minutes
Who it is for
Staff responsible for documentation and procedures, secretaries, IT teams and teams preparing a search engine or knowledge base. 72% of officials see the greatest AI potential in searching and analysing documents (NASK, 2026)
Natural-language document search was the highest-rated NASK concept, but it raises questions about permissions, source currency and ownership. Participants learn RAG logic, organise document status and ownership, and design source-grounded answers that respect access rights
What participants gain
- Understands the difference between a general AI assistant and a RAG knowledge base
- Distinguishes current, archived and working documents
- Knows who owns documents, updates and permissions
- Can respond when reliable information is missing or sources conflict
Outcome for the organisation: The institution gains a structured starting point for a RAG knowledge base: source scope, document statuses, owners, update rules, permissions and answer-quality criteria
Next step: The source catalogue can be used directly in a search pilot (Course 04). In SAIE, the knowledge base runs on the institution’s infrastructure or private cloud, with department permissions and citations in every answer
Training programme
A. From documents to a reliable answer
- Search combined with answer generation
- General-purpose assistant versus documentation-based assistant
- An answer with a citation, location and date as a condition of trust
Outcome: Flow from a question to a source-grounded answer
B. Information governance and source readiness
- Document scope and source hierarchy
- Status: current, archived, draft or approved
- Version, validity period, review date and owner
Outcome: Source catalogue with metadata and owners—the basis for an ingestion procedure
C. Access, testing and quality maintenance
- Roles and least privilege: staff see only what they may access
- Test questions, edge cases and hallucination-resistance tests
- Relevance, completeness and currency; outage procedure
Outcome: Minimum knowledge-base design with acceptance tests and assigned ownership
Training dates
Term 1: 24.09 • czw • 10:00 - 11:20 | 25.09 • pt • 10:00 - 11:20 | 28.09 • pon • 10:00 - 11:20
Term 2: 28.10 • śr • 10:00 - 11:20 | 29.10 • czw • 10:00 - 11:20 | 30.10 • pt • 10:00 - 11:20
Term 3: 23.11 • pon • 10:00 - 11:20 | 24.11 • wt • 10:00 - 11:20 | 25.11 • śr • 10:00 - 11:20
03 Security — Safe and responsible use of AI—GDPR, the AI Act and internal rules
Workshop • 3 × 80 minutes
Who it is for
Management, secretaries, DPOs, information-security officers and department managers. 66% of officials want training on the legal aspects of AI (NASK, 2026)
Unclear regulations, missing regulations and missing internal procedures are the three most cited barriers. This course turns GDPR and AI Act requirements into practical rules: participants classify uses, define human oversight and prepare an internal AI policy
What participants gain
- Recognises risks involving data, content, private accounts and liability
- Knows when AI is allowed, requires approval or must not be used
- Knows when a person must verify an AI result
- Can assess whether an AI tool is approved, secure and compliant with GDPR and the AI Act
Outcome for the organisation: The institution gains a shared vocabulary and a working policy model covering permitted uses, conditions, approval and incident handling
Next step: The policy and register underpin a safe pilot (Course 04). PROXY:AI can enforce approved-model access, personal-data filtering and use logging. allclouds.pl is Poland’s first Bureau Veritas-certified ISO/IEC 42001 company
Training programme
A. Law, data and controlled use cases
- GDPR in AI work: personal and special-category data, minimisation, anonymisation and DPIA
- AI Act: prohibited practices, high-risk public-sector systems and AI literacy (Article 4)
- Approved and unapproved AI environments; what a supplier contract does not guarantee
Outcome: Initial review of AI uses and processed information with risk classification
B. The person in the process and accountability boundaries
- AI as supporting material, not a decision
- Automated actions, proposals for approval and human decisions
- Roles of user, verifier, process owner and expert
Outcome: Human-oversight map for one use case (AI Act, Article 14)
C. Organisational rules, register and incidents
- Register of approved tools and uses
- Minimum logging and log protection
- Reporting errors, breaches or out-of-policy use
Outcome: Draft safe-AI policy ready for adoption by internal order
Training dates
Term 1: 21.09 • pon • 10:00 - 11:20 | 22.09 • wt • 10:00 - 11:20 | 23.09 • śr • 10:00 - 11:20
Term 2: 23.10 • pt • 10:00 - 11:20 | 26.10 • pon • 10:00 - 11:20 | 27.10 • wt • 10:00 - 11:20
Term 3: 19.11 • czw • 10:00 - 11:20 | 20.11 • pt • 10:00 - 11:20 | 23.11 • pon • 10:00 - 11:20
Term 4: 17.12 • czw • 10:00 - 11:20 | 18.12 • pt • 10:00 - 11:20 | 21.12 • pon • 10:00 - 11:20
04 Pilot — Implementing AI in public administration—from problem to controlled pilot
Workshop • 3 × 80 minutes
Who it is for
Secretaries, department managers, process owners, IT and project staff—the team deciding on the first implementation
The Ministry guide starts at Stage 0: is AI needed at all? This course follows the same path from a defined problem to a small, reversible and measurable pilot. Technology is selected only after outcomes, data, risks and controls are clear
What participants gain
- Identifies problems where AI may help
- Assesses benefits, risks and organisational readiness
- Plans a small, safe and reversible AI test
- Defines metrics, operating costs, controls and stop conditions
Outcome for the organisation: Participants finish with a realistic pilot charter rather than a generic idea list. It can be reviewed by legal, security, IT and the process owner
Next step: The charter feeds a CDF pilot: demonstration on institutional data, limited pilot and scale decision. SAIE supports a pilot without buying infrastructure or locking into one supplier
Training programme
A. Problem before technology
- AI versus ordinary automation: AI, a better form or a better procedure?
- Repetition, time cost and verifiability
- User, need and measurable goal: service time, errors and availability
Outcome: Use-case description with a measurable goal and expected benefits
B. Readiness and pilot design
- Data, sources, procedures, skills and infrastructure; GDPR and AI Act risk classification
- Small scope, test group and manual fallback
- Process, data, knowledge and technology owner roles
Outcome: Limited pilot design with assigned roles
C. Measurement and decision
- Tests on real cases: time with and without AI, consistency and unusual queries
- Output quality, errors, corrections and rejections
- Verification, data-update and maintenance cost
Outcome: Pilot charter with metrics and a continue / redesign / stop gate
Training dates
Term 1: 29.09 • wt • 10:00 - 11:20 | 30.09 • śr • 10:00 - 11:20 | 01.10 • czw • 10:00 - 11:20
Term 2: 02.11 • pon • 10:00 - 11:20 | 03.11 • wt • 10:00 - 11:20 | 04.11 • śr • 10:00 - 11:20
Term 3: 30.11 • pon • 10:00 - 11:20 | 01.12 • wt • 10:00 - 11:20 | 02.12 • śr • 10:00 - 11:20
Term 4: 21.12 • pon • 10:00 - 11:20 | 22.12 • wt • 10:00 - 11:20 | 23.12 • śr • 10:00 - 11:20
05 Automation — AI agents and process automation in public administration
Workshop • 3 × 80 minutes
Who it is for
IT, records and document-workflow process owners and secretariat managers, especially institutions considering automated registration or routing
AI agents drew the most scepticism in the NASK study. Participants distinguish agents from general assistants and rule-based automation, then design one bounded process with explicit actions, permissions and human decisions
What participants gain
- Understands assistants, workflows, rule automation and AI agents
- Selects a process suitable for agent support
- Defines what the agent may do and access
- Knows how to reverse an action, fall back to manual handling or stop safely
Outcome for the organisation: Participants can judge whether an agent is justified and design it so people retain decisions and process control. The agent starts with a simple proposal, not autonomy
Next step: The flow can be piloted in SAIE, where automation works alongside knowledge bases and assistants while PROXY:AI enforces permissions and logs every agent step
Training programme
A. Does this process need an agent?
- AI assistant, workflow, rule-based automation and AI agent
- Repeatable, reversible and verifiable tasks
- Risk of duplicating existing document-management functions
Outcome: Reasoned decision: agent / automation / no implementation
B. Designing a safe flow
- Trigger, inputs, knowledge sources and tools
- Least privilege, action limits and no autonomous citizen decisions
- Automated steps, AI steps and human approval
Outcome: One process map with a defined AI-autonomy boundary
C. Testing, monitoring and error response
- Typical, unusual and invalid scenarios tested before launch
- Stop criteria, human escalation and manual fallback
- Logging steps, decisions and tools used
Outcome: Test and oversight plan for the designed flow
Training dates
Term 1: 12.10 • pon • 10:00 - 11:20 | 13.10 • wt • 10:00 - 11:20 | 14.10 • śr • 10:00 - 11:20
Term 2: 05.11 • czw • 10:00 - 11:20 | 06.11 • pt • 10:00 - 11:20 | 09.11 • pon • 10:00 - 11:20
Term 3: 26.11 • czw • 10:00 - 11:20 | 27.11 • pt • 10:00 - 11:20 | 30.11 • pon • 10:00 - 11:20
Term 4: 14.12 • pon • 10:00 - 11:20 | 15.12 • wt • 10:00 - 11:20 | 16.12 • śr • 10:00 - 11:20
06 Citizen service — AI in citizen service—faster, simpler and inclusive
Workshop • 3 × 80 minutes
Who it is for
Citizen-service offices, information desks, call centres, communications teams and departments handling repeated questions. A separate internal-assistant variant is available for regional and central authorities
Officials support citizen chat when it informs rather than advises, clearly identifies AI, offers a human hand-off and creates no new digital barrier. Participants design an informative, bounded and inclusive service
What participants gain
- Distinguishes informing, advising and deciding, and knows where AI is appropriate
- Designs an AI assistant’s knowledge and answer scope
- Ensures plain language and transparency
- Defines when the assistant must offer a staff member or another channel
Outcome for the organisation: The institution gains a safe information-service scenario that reduces simple enquiries without blocking access to a person
Next step: The scenario can run in SAIE on a local knowledge base and a Polish model such as PLLuM or Bielik, with full conversation logging. Small municipalities can share an assistant
Training programme
A. AI’s role in citizen contact
- Repeated questions and first-line needs
- Information guide, not adviser or decision-maker
- Clear disclosure that the user is interacting with AI
Outcome: Assistant scope and a list of matters outside its remit
B. Service path and content design
- User intents and clarifying questions
- Plain-language answer with source, date and legal version
- Link to the correct form, department or instruction
Outcome: Conversation scenario and citizen-service stages
C. Accessibility, currency and trust testing
- Tests with different levels of digital skill
- Hallucinations, unusual cases and missing grounds
- Knowledge-base owner and update cycle after legal changes
Outcome: Service test plan covering quality, accessibility, hand-off and currency
Training dates
Term 1: 07.10 • śr • 10:00 - 11:20 | 08.10 • czw • 10:00 - 11:20 | 09.10 • pt • 10:00 - 11:20
Term 2: 09.11 • pon • 10:00 - 11:20 | 10.11 • wt • 10:00 - 11:20 | 12.11 • czw • 10:00 - 11:20
Term 3: 03.12 • czw • 10:00 - 11:20 | 04.12 • pt • 10:00 - 11:20 | 07.12 • pon • 10:00 - 11:20
Term 4: 17.12 • czw • 10:00 - 11:20 | 18.12 • pt • 10:00 - 11:20 | 21.12 • pon • 10:00 - 11:20
07 Change — AI without fear or duplicate work—managing change in a public institution
Workshop • 3 × 80 minutes
Who it is for
Management, secretaries, HR, department managers and future AI ambassadors—the team leading implementation
Officials’ concerns include error, liability, duplicate work and loss of expertise. Participants identify concerns, involve staff in testing and verify that AI makes work easier without weakening competence
What participants gain
- Understands employee concerns and the real cost of change
- Knows when to explain an implementation better and when to change it
- Can involve staff in testing and evaluating AI
Outcome for the organisation: The institution gains a change plan reflecting staff concerns and knowledge, the work balance and safe adoption instead of forced tool use
Next step: The change plan accompanies the Course 04 pilot and forms part of CDF, where roles, concerns and team capabilities are defined before technology selection
Training programme
A. Diagnosing concerns and stakeholders
- Fear of role loss, control, error and lost expertise
- Fear that overseeing AI costs more than it saves
- Differences between user, manager, IT and control functions
Outcome: List of people involved, their concerns and information needs
B. Co-designing the new way of working
- What disappears, remains and is created—and who verifies
- Staff role in selecting and testing the use case
- Human control at key stages instead of automatic approval
Outcome: Before-and-after comparison of time and competence
C. Communication, learning and sustained change
- On-the-job support when a tool fails
- Short training, practice, consultations and AI ambassadors
- Safe reporting of errors and problems
Outcome: Communication and skills-development plan for the test
Training dates
Term 1: 02.10 • pt • 10:00 - 11:20 | 05.10 • pon • 10:00 - 11:20 | 06.10 • wt • 10:00 - 11:20
Term 2: 12.11 • czw • 10:00 - 11:20 | 13.11 • pt • 10:00 - 11:20 | 16.11 • pon • 10:00 - 11:20
Term 3: 01.12 • wt • 10:00 - 11:20 | 02.12 • śr • 10:00 - 11:20 | 03.12 • czw • 10:00 - 11:20
Term 4: 21.12 • pon • 10:00 - 11:20 | 22.12 • wt • 10:00 - 11:20 | 23.12 • śr • 10:00 - 11:20
08 Collaboration — AI together—sharing experience and collaborating across institutions
Workshop • 3 × 80 minutes
Who it is for
Secretaries and digital coordinators from several institutions, municipal associations, agreements, territorial partnerships, counties and municipalities; delivered as a joint workshop
Small municipalities cite implementation cost and propose shared solutions. Participants learn to exchange AI experience without blindly copying solutions or disclosing protected information, using comparable pilot records and peer review
What participants gain
- Documents an AI test so other institutions understand its process and results
- Separates institution-specific findings from reusable ones
- Structures and discusses conclusions from a test
- Knows what to conceal or anonymise and what must not be shared
Outcome for the organisation: Institutions gain a shared format for comparing experience and a practical collaboration plan that prevents duplicated testing and enables reuse
Next step: A group can launch a shared SAIE pilot: one environment, separate knowledge bases and permissions, with costs shared among participants
Training programme
A. Describing experience so others can use it
- Problem, user, scope, data and environment
- Metrics, result, errors, cost and decision
- Facts, interpretations and recommendations
Outcome: Pilot summary card comparable across institutions
B. Peer review and transferring a solution
- Critical questions without personal judgement
- Conditions for success and causes of failure
- Differences in process, data, local law and infrastructure
Outcome: Assessment of whether other institutions can reuse the solution
C. A lasting collaboration mechanism
- Card repository, webinar, forum or working group linked to AI HUB (ai.gov.pl)
- Owner, moderator and expert roles
- Update, review and retirement cycle
Outcome: 90-day experience-sharing and joint-action plan, including a shared pilot option
Training dates
Term 1: 15.10 • czw • 10:00 - 11:20 | 16.10 • pt • 10:00 - 11:20 | 19.10 • pon • 10:00 - 11:20
Term 2: 17.11 • wt • 10:00 - 11:20 | 18.11 • śr • 10:00 - 11:20 | 19.11 • czw • 10:00 - 11:20
Term 3: 07.12 • pon • 10:00 - 11:20 | 08.12 • wt • 10:00 - 11:20 | 09.12 • śr • 10:00 - 11:20
Training calendar
All 2026 dates in one place
01 Foundations: 16.09 • śr • 10:00 - 11:20 | 17.09 • czw • 10:00 - 11:20 | 18.09 • pt • 10:00 - 11:20 || 20.10 • wt • 10:00 - 11:20 | 21.10 • śr • 10:00 - 11:20 | 22.10 • czw • 10:00 - 11:20 || 16.11 • pon • 10:00 - 11:20 | 17.11 • wt • 10:00 - 11:20 | 18.11 • śr • 10:00 - 11:20 || 14.12 • pon • 10:00 - 11:20 | 15.12 • wt • 10:00 - 11:20 | 16.12 • śr • 10:00 - 11:20
02 Knowledge: 24.09 • czw • 10:00 - 11:20 | 25.09 • pt • 10:00 - 11:20 | 28.09 • pon • 10:00 - 11:20 || 28.10 • śr • 10:00 - 11:20 | 29.10 • czw • 10:00 - 11:20 | 30.10 • pt • 10:00 - 11:20 || 23.11 • pon • 10:00 - 11:20 | 24.11 • wt • 10:00 - 11:20 | 25.11 • śr • 10:00 - 11:20
03 Security: 21.09 • pon • 10:00 - 11:20 | 22.09 • wt • 10:00 - 11:20 | 23.09 • śr • 10:00 - 11:20 || 23.10 • pt • 10:00 - 11:20 | 26.10 • pon • 10:00 - 11:20 | 27.10 • wt • 10:00 - 11:20 || 19.11 • czw • 10:00 - 11:20 | 20.11 • pt • 10:00 - 11:20 | 23.11 • pon • 10:00 - 11:20 || 17.12 • czw • 10:00 - 11:20 | 18.12 • pt • 10:00 - 11:20 | 21.12 • pon • 10:00 - 11:20
04 Pilot: 29.09 • wt • 10:00 - 11:20 | 30.09 • śr • 10:00 - 11:20 | 01.10 • czw • 10:00 - 11:20 || 02.11 • pon • 10:00 - 11:20 | 03.11 • wt • 10:00 - 11:20 | 04.11 • śr • 10:00 - 11:20 || 30.11 • pon • 10:00 - 11:20 | 01.12 • wt • 10:00 - 11:20 | 02.12 • śr • 10:00 - 11:20 || 21.12 • pon • 10:00 - 11:20 | 22.12 • wt • 10:00 - 11:20 | 23.12 • śr • 10:00 - 11:20
05 Automation: 12.10 • pon • 10:00 - 11:20 | 13.10 • wt • 10:00 - 11:20 | 14.10 • śr • 10:00 - 11:20 || 05.11 • czw • 10:00 - 11:20 | 06.11 • pt • 10:00 - 11:20 | 09.11 • pon • 10:00 - 11:20 || 26.11 • czw • 10:00 - 11:20 | 27.11 • pt • 10:00 - 11:20 | 30.11 • pon • 10:00 - 11:20 || 14.12 • pon • 10:00 - 11:20 | 15.12 • wt • 10:00 - 11:20 | 16.12 • śr • 10:00 - 11:20
06 Citizen service: 07.10 • śr • 10:00 - 11:20 | 08.10 • czw • 10:00 - 11:20 | 09.10 • pt • 10:00 - 11:20 || 09.11 • pon • 10:00 - 11:20 | 10.11 • wt • 10:00 - 11:20 | 12.11 • czw • 10:00 - 11:20 || 03.12 • czw • 10:00 - 11:20 | 04.12 • pt • 10:00 - 11:20 | 07.12 • pon • 10:00 - 11:20 || 17.12 • czw • 10:00 - 11:20 | 18.12 • pt • 10:00 - 11:20 | 21.12 • pon • 10:00 - 11:20
07 Change: 02.10 • pt • 10:00 - 11:20 | 05.10 • pon • 10:00 - 11:20 | 06.10 • wt • 10:00 - 11:20 || 12.11 • czw • 10:00 - 11:20 | 13.11 • pt • 10:00 - 11:20 | 16.11 • pon • 10:00 - 11:20 || 01.12 • wt • 10:00 - 11:20 | 02.12 • śr • 10:00 - 11:20 | 03.12 • czw • 10:00 - 11:20 || 21.12 • pon • 10:00 - 11:20 | 22.12 • wt • 10:00 - 11:20 | 23.12 • śr • 10:00 - 11:20
08 Collaboration: 15.10 • czw • 10:00 - 11:20 | 16.10 • pt • 10:00 - 11:20 | 19.10 • pon • 10:00 - 11:20 || 17.11 • wt • 10:00 - 11:20 | 18.11 • śr • 10:00 - 11:20 | 19.11 • czw • 10:00 - 11:20 || 07.12 • pon • 10:00 - 11:20 | 08.12 • wt • 10:00 - 11:20 | 09.12 • śr • 10:00 - 11:20
ALLCLOUDS.PL IS THE FIRST COMPANY IN POLAND CERTIFIED TO ISO/IEC 42001 BY BUREAU VERITAS
We teach AI management in public administration the way we practise it—according to the standard
Our AI management system covers every AI system, the CDF methodology, SAIE ecosystem products, personnel and supply chain from concept to retirement. Training and consulting are covered by the same Integrated Management System alongside ISO 9001, 27001, 27017, 27018, 22301, 14001 and 42001
42001 audit: 0 nonconformities
Instructors
People who build AI systems for the public sector—not just talk about them
Four instructors, each with more than five years of public-institution project experience. Designers of TWIN:DESK and PROXY:AI and co-authors of CDF and an ISO/IEC 42001-certified AI management system
Anna Pawelec — AI systems designer
Specialisms: TWIN:DESK, Systems for central institutions, External-user processes
For more than five years she has designed public-sector IT solutions. A lead TWIN:DESK designer, she developed process systems involving external users for central institutions and now leads the TWIN:DESK roadmap
In the courses: turning institutional documents into a working knowledge base and designing an assistant that always leaves citizens a path to a person
Teaches: 02 KNOWLEDGE, 05 AUTOMATION, 06 CITIZEN SERVICE
Robert Duda — AI security consultant
Specialisms: PROXY:AI, IT security, Critical infrastructure
For more than ten years he has built public-sector IT solutions. A lead PROXY:AI designer, he has worked on critical network and security implementations and specialises in secure AI adoption and training
In the courses: what happens to data entered into AI tools and how to build rules that can be enforced technically
Teaches: 03 SECURITY, 05 AUTOMATION
Rafał Grzegorzewski — AI security and methodology consultant
Specialisms: CDF methodology, ISO/IEC 42001, AI Act, Reporting systems
For more than ten years he has built public-sector IT solutions, including reporting systems for supervisory institutions. He specialises in CDF, ISO/IEC 42001 and AI Act training
In the courses: moving from a problem to a controlled pilot and preparing documentation that stands up to audit
Teaches: 03 SECURITY, 04 PILOT, 08 COLLABORATION
Marek Skrycki — AI systems project manager and adoption lead
Specialisms: Digital transformation, Project management, AI adoption
An experienced project manager specialising in digital transformation, he has managed complex technology projects with multi-million-zloty budgets and now leads practical AI adoption
In the courses: bringing AI into everyday teamwork without fear or duplicate work, and leading lasting change
Teaches: 01 FOUNDATIONS, 04 PILOT, 07 CHANGE
How to order
From order to training in 4 steps
1. Complete the order form
YOU • 5 MINUTES. Choose courses, dates and participant numbers, then enter requester and institution details
2. Receive order confirmation
US • E-MAIL. We e-mail confirmation and, for paid training, a proforma invoice
3. Pay the proforma before training
YOU • PAID TRAINING ONLY. Pay by bank transfer. This step is skipped for the free edition
4. Receive the training link
US • 3 DAYS BEFORE TRAINING. We e-mail it three days before the course. A browser is all you need
Training order form
- Requester: name, business e-mail, phone and role
- Institution: name, street, building number, postal code, city and NIP
- Each course: name, dates, participant count and each participant’s name and e-mail
- Required personal-data processing consent
- Optional newsletter consent determining eligibility for the September–October 2026 free edition
Fields marked with an asterisk are required. The form collects requester, institution, selected-course, price and participant details
Newsletter consent is optional and determines eligibility for the September–October 2026 free edition
Consent to personal-data processing for order handling is required
Training financed entirely from public funds (Article 43(1)(29)(c) of the VAT Act) or at least 70% from public funds (§ 3(1)(14) of the Minister of Finance Regulation of 20 December 2013) is VAT exempt, so the gross price equals the net price. An institution paying from its own budget normally selects 100% funding. The declaration template sent with order confirmation includes both options. If either applies, complete, sign and return it to office@allclouds.pl
What happens after submission
- 1. Order form: Courses, dates, participant numbers and institution details
- 2. E-mail confirmation: For paid training, together with the proforma
- 3. Proforma payment: Before training; skipped in the free edition
- 4. Training link: Sent by e-mail three days before the course
Training FAQ
Registration and free edition
1. How do we register and how long does the process take?
Complete the page form with requester, institution, course, date and participant details. We e-mail a copy and, for paid training, a proforma. Participants receive the training link three days before the first session
2. Who may register participants—must they be authorised to represent the institution?
Any institutional employee may submit the form from a business e-mail address. Paid orders require the proforma to be paid before training
3. How does the free September–October edition work and who must subscribe?
For September and October dates, the first five people from one institution attend each course free when the requester subscribes to the allclouds.pl newsletter. Only the requester needs to consent, and consent may be withdrawn at any time
4. What if we register more than five people for one free-edition course?
Places from the sixth onward use November–December rates. For 6–10 people the rate is PLN 170 net per paid place; eight people in September therefore cost 5 × PLN 0 + 3 × PLN 170 = PLN 510 net. The group maximum is 18
5. Can one institution use free places on all eight courses?
Yes. The five-person limit applies separately to each course, so one institution can receive up to 40 free person-courses, subject to capacity and registration order
Confirmation and changes
6. What happens after submission and when do we receive confirmation?
We e-mail confirmation immediately after submission. A paid order includes a proforma in the same e-mail, and participants receive the training link three days before the first session
7. Can we change a date or participant after confirmation?
Yes, free of charge. Replace a participant from the same institution up to the start date, or change to an available calendar date up to five business days before training by e-mailing office@allclouds.pl
8. When can we cancel without cost?
Cancel up to five business days before the first session. Later, we offer another date or participant substitution instead of charging a cancellation fee
9. Can one order include several courses, dates and participant counts?
Yes. Add any number of items, each with its own date and participant count. Pricing is automatic and paid items appear on one confirmation and invoice
10. How do we order the complete path or a role-based path?
Select the recommended courses one by one in the order form so you can choose suitable dates for each course
Payment and documents
11. When do you issue an invoice and when is payment due?
We issue a proforma after registration and a VAT invoice after payment. On request, the VAT invoice can be split by organisational unit
12. Are prices net or gross, and can training be VAT exempt?
The page shows net and gross prices. Training financed entirely from public funds (Article 43(1)(29)(c) of the VAT Act) or at least 70% from public funds (§ 3(1)(14) of the Minister of Finance Regulation of 20 December 2013) is VAT exempt, so the gross price equals the net price. An institution paying from its own budget normally selects 100% funding. The declaration template sent with order confirmation includes both options. If either applies, complete, sign and return it to office@allclouds.pl
13. Do participant-count and complete-path discounts combine?
Yes. A team of 6–10 on the complete path pays PLN 1,088 net per person; 11–18 pays PLN 960 net. Participant discounts are calculated per course and institution
14. Which procurement documents can we receive?
We can provide an offer, timed programme, order template, public-funding declaration, GDPR notice and company information including KRS, NIP and ISO 9001, 27001 and 42001 certificates
15. Do you sign the institution’s contract and handle orders below the procurement threshold?
Yes. Training is usually delivered under an institution order or contract below the Polish Public Procurement Act threshold (PLN 170,000 net from 1 January 2026), under your procedures
Course delivery
16. Which platform is used and must participants install anything?
Courses run online in MS Teams. A browser and our link are sufficient; no account or installation is required. Tell us if institutional security blocks a service and we will choose a compatible option
17. Why three 80-minute days, and can someone attend only one day?
Three consecutive 80-minute blocks limit disruption and allow practice between sessions. They form one course and are not sold separately; a missed block may be completed on another date free of charge
18. What are the minimum and maximum group sizes?
Groups have 8–18 participants. If fewer than eight are registered three business days before training, we offer the nearest date or a closed session for one institution; we always offer an alternative
19. Are sessions recorded?
No. Participants discuss institutional examples freely, so sessions remain within the group. Instead, participants receive materials, templates and block summaries
20. Are people from different institutions in one group?
Yes, open dates mix institutions to support experience sharing, using anonymised or prepared examples. A closed session is available for one institution
Materials, tools and outcomes
21. What does a participant receive?
Each participant receives PDF materials, editable workshop document templates and a named attendance certificate listing the programme
22. Which AI tools are used, and does the institution need licences?
Exercises use browser tools, our products and several LLMs including PLLuM and Bielik. We provide the exercise environment; no licence purchase is required
23. Is institutional data used in exercises and how is it protected?
No. Exercises use prepared or anonymised material. allclouds infrastructure is in Poland and certified to ISO/IEC 27001, 27017, 27018 and 42001; registration data is used only to deliver training
24. What is the working document, and may the institution use it formally?
Each course produces an editable document such as a task map, checklist, source catalogue, AI policy or pilot charter, designed to support an internal order, procedure or pilot request without licence restrictions
25. Does the course meet the AI Act Article 4 literacy obligation?
The programme covers AI limitations, critical evaluation and prompting. Evidence includes a named certificate, programme and workshop document. The institution remains responsible for judging suitability for each role
Instructors, tailoring and support
26. Who teaches the courses and what public-sector experience do they have?
The instructors have designed and implemented AI solutions in regulated institutions and co-created CDF and an ISO/IEC 42001-certified AI management system
27. Can we order a closed, custom-date or on-site course?
Yes. Closed online training for 8–18 people can use your anonymised processes and documents. On-site delivery and other tailoring are individually priced; ask through the contact form
28. Can the programme be tailored to the institution?
Open-course programmes are fixed but examples reflect the group. Course 06 has an internal-assistant variant; Course 08 supports multi-institution groups. Closed courses offer full process and document tailoring
29. Is consulting or implementation support available afterward?
Yes: 30 days of e-mail consultation is included, with optional paid implementation consulting and a CDF pilot in SAIE. Training does not obligate you to buy either; workshop documents remain yours