How Vedomia uses AI

Last reviewed: August 2026

We help organisations disclose their use of AI honestly — so we hold ourselves to the same standard.

Where we use AI

We use AI language tools (large language models) to help draft and refine some of our website copy, our blog articles, our Slovak/English translations, and the wording inside our free tools. AI helps us with drafting and phrasing — not with judging your organisation.

When you ask for it, AI also reads the document or process description you provide and drafts a process map or a letter from it. That draft is about your material, and you are the one who checks it.

We also use AI image tools to generate the illustrative photographs on this website. They depict scenes we composed — not real people, not our clients, and not real events. Each one carries a visible label saying so.

Where we do NOT use AI

Our rule: process first

We believe AI cannot be made transparent in isolation, because AI never works in isolation — it always lives inside a process. So we treat our own AI use the way we advise our clients to treat theirs: map the process, mark where AI enters, make that step visible, and keep the evidence. Every entry on this page is the result of that exercise applied to Vedomia itself.

Human editorial responsibility

Every AI-assisted text published under the Vedomia name is reviewed, edited and approved by Sandra Ahmidat Fedakova before it goes live. She holds editorial responsibility for all published content, in the sense set out in Article 50(4) of the EU AI Act (Regulation (EU) 2024/1689). When you read something here, a named human has stood behind it.

Our editorial process, step by step

Because oversight only counts if it is a real step in a real process — not a sentence in a policy — this is the sequence every AI-assisted piece of content goes through:

This is not bureaucracy for its own sake. It is the practical answer to a simple question: if a client, funder or regulator asked tomorrow, could we show — not just say — how AI is used here? With a mapped process, the answer is yes.

Your data

Our free tools run in your browser — while you are simply filling them in, your answers stay on your device (see our Privacy Notice). One step is different, and only if you choose it: in the process-map builder and on /ask/ you can ask us to read a document or a description you provide. That text is sent once, through our server, to our AI provider (Anthropic) and is processed in the United States; we do not store it and it is not used to train models. What comes back is an AI draft, labelled as one — no person at Vedomia reads it before you do, so please check it before you rely on it.

Why we publish this

Article 50 of the EU AI Act introduces transparency duties in specific situations, with the relevant obligations applying from 2 August 2026 (subject to the scope and exceptions in the Act). We build tools that help charities and NGOs work out where those situations may be relevant and prepare for them — so publishing our own statement is simply us using our own method. If you'd like to explore what might be relevant for your organisation, our free AI Transparency Relevance Check will screen the situations and draft wording for your review in a few minutes.

We will update this statement as our use of AI changes.

Contact

Questions about this statement? Write to contact@vedomia.com.

This is the same method we would use on your process.

Map one of your own processes free, in your browser — and see where AI enters it, who reviews it, and what you could show if someone asked.