We publish the most important updates about what is happening at allclouds.pl and around SAIE. New deployments, products, features, projects, partnerships and recruitment — everything that shows how we develop our solutions and where we use AI in practice.
This is the place for short updates on what we have just launched, published, deployed or started working on. No marketing noise — specific information about next steps, outcomes and development directions.
Join our free 90-minute webinar
The second episode of “Four questions before choosing an AI environment” separates the word “security” into three distinct issues: data location, employees’ working materials and ownership of developed processes. Each calls for a different solution, and each carries a different cost if handled incorrectly.
Eight AI training sessions for public offices. Up to 5 people from each institution can participate free of charge in September and October
The first episode of “Four questions before choosing an AI environment” shows the full cost calculation: model, infrastructure, maintenance and skills.
The year 2026 brought lawyers three signals at once: the Polish Bar Council added Section 23e on technological tools to its Code of Ethics, the Act on Artificial Intelligence Systems entered into force, and recommendations issued by the judges’ association Iustitia ruled out processing case data in commercial chatbots. The question about AI in a law firm stopped being “whether” and became “how”.
Deterministic municipal financial analytics through natural-language conversation
What we will release in the coming quarters
Secure AI deployment in critical infrastructure
Python, integrations, language models — hybrid or remote
A comparison of three approaches
Workshops, AI adoption, prompt engineering — hybrid or remote
AI system testing, scenarios, response quality — hybrid or remote
Can the CPO and CISO speak with one voice?
Your AI supplier audit just became shorter
We will show you with concrete examples how to safely use AI in government offices, protect data, and prepare your organization for an orderly implementation.
In conversations about AI, “security” usually means three entirely different things that should be separated before a decision is made. An organisation’s most valuable knowledge is often hidden in working materials that employees do not want to send anywhere, while its real advantage lies in process logic that can easily be lost when changing suppliers.
The “AI in Public Administration” programme is based on two documents that currently set the direction for public offices: NASK’s report “AI in e-public administration – the perspective of officials and institutions” and the Ministry of Digital Affairs’ “Guide to Artificial Intelligence for Public Administration”. The exercises are grounded in the realities of public offices and built around four concepts from the report: a chat service for residents, content creation, document search, and AI agents. The programme develops three competencies indicated in the Ministry’s guide — understanding the limitations of AI, critically assessing outputs, and formulating prompts — as well as the obligation to ensure AI literacy under Article 4 of the AI Act.
This is the topic of the first episode in the series “Four questions before choosing an AI environment”. In four short materials, we look at the decisions worth making before an organization commits to a specific technology or supplier:
The year 2026 brought lawyers three signals at once: the Polish Bar Council added § 23e on technological tools to the ethics code, the Act on artificial intelligence systems entered into force, and the judges’ association Iustitia recommended excluding case data from processing in commercial chatbots. The question about AI in a law firm is no longer “whether”, but “how”.
The year 2026 brought lawyers three signals at once: the Polish Bar Council added Section 23e on technological tools to its Code of Ethics, the Act on Artificial Intelligence Systems entered into force, and recommendations issued by the judges’ association Iustitia ruled out processing case data in commercial chatbots. The question about AI in a law firm stopped being “whether” and became “how”.
In the new article, we describe a prototype built in TWIN:DESK that separates language understanding from calculations: the AI model recognizes the treasurer's question, a deterministic engine calculates over 1.85 million reporting records from the Ministry of Finance, and every result is returned with its source, formula and limitations. Comparisons with 66 cities with county rights, multi-year financial forecast analysis and cost-pressure scenarios, all based on public data.
The solution, based on TWIN:DESK, PROXY:AI and CDF Platform, runs fully on-premises, ensuring secure knowledge processing, semantic document search and answers with source citations - without sending data outside the organization.
We are publishing it for a simple reason: organizations in regulated sectors plan deployments over a one-year horizon and beyond. Choosing an AI platform is a long-term commitment — architectural, budgetary and compliance-related. A public roadmap allows client-side teams to plan their own work based on what will actually appear in the product, not on sales declarations
You will build SAIE — TWIN:DESK, PROXY:AI and CDF Platform. You will design AI system components, develop backend services and integrate them with client environments. You need at least two years of experience in Python, integrations or backend development — AI project experience is an advantage, not a requirement
Open source is limited by accountability. The code is transparent, but maintenance, security, compliance and development fall entirely on the organization's own team. There is no service, no guarantee and no other party to the contract. For a company, this is not a platform — it is a project it has to run on its own, indefinitely
You will deploy SAIE — TWIN:DESK, PROXY:AI and CDF Platform. You will run specialist workshops, configure AI applications and knowledge bases, and build dedicated assistants. You need at least two years of experience in change management, business analysis or software implementation, plus fluency in working with language models — at power-user level, not as a programmer
You will test SAIE — TWIN:DESK, PROXY:AI and CDF Platform. You will create scenarios for typical and atypical use cases, catch errors and describe observations in a way the team can work with. We do not require extensive professional experience — curiosity about technology, attention to detail and confidence in working with AI tools matter most
In the new article, we show why neither a ban nor blanket permission resolves this conflict — and how architecture does: an AI gateway that separates two operating regimes. Market research goes to the best external models, procurement-process data stays in the local environment, and the boundary is enforced by technology, not by policy. Add to that an evidence package for every interaction — compliance with public procurement rules and the AI Act that can be shown to an auditor, not merely declared
Bureau Veritas Polska awarded us an ISO/IEC 42001:2023 certificate for our Artificial Intelligence Management System — the first certificate of its kind issued by this body in Poland. The certification covers TWIN:DESK, PROXY:AI and the CDF methodology.