← Back to the app

AI Use Policy

version 1.0 · In force since

1. Why this document exists

Boardeon uses artificial intelligence. We want you to know exactly where, how it works, what to expect from it and what not to do. This document forms part of the Terms of Service and fulfils the transparency obligations arising from the AI Act (Regulation (EU) 2024/1689).

Boardeon is operated by JG INVEST prosta spółka akcyjna, ul. Jana Matejki 2, 32-600 Oświęcim, KRS 0000915798.

2. Where Boardeon uses AI

Responses produced by a language model appear in:

  • the Advisor — a conversation about your company,
  • AI Boardroom — a meeting at which several AI directors assess your question from different perspectives and a chair summarises the discussion,
  • strategic, marketing, growth, competition and SEO analyses — commentary on the results of analytical frameworks,
  • the 90-day plan and summaries — commentary on priorities and results,
  • document search — the model converts document text into vectors to find the fragments that match the question.

You always know when you are talking to AI: model responses are labelled in Boardeon, and the directors in AI Boardroom are roles played by the model, not people.

3. What AI in Boardeon does not do

  • It does not calculate your numbers. Business Score, margins, forecasts, simulations and the valuation are calculated by ordinary, verifiable code from your data. The model comments on them but does not produce them. When data is missing, Boardeon says so plainly instead of guessing.
  • It does not make decisions for you. Recommendations are proposals. A human decides what to do.
  • It does not act on your behalf outside Boardeon. AI does not send messages to your clients, does not change settings in other systems and does not spend money.
  • It does not assess people. Boardeon analyses the company, not employees, candidates or clients as individuals.

4. Models and providers

Boardeon selects the model for the task: a faster one for short answers, a more powerful one for strategic analyses and board meetings. We route requests through OpenRouter to model providers, currently including OpenAI and Google. The list of entities processing data is in the Data Processing Agreement.

You may connect your own OpenRouter key. You then settle your Organisation's requests directly with OpenRouter, under your agreement with it, and you have access to more powerful models.

5. Your data and AI

  • Only the fragments of data needed for a specific request are sent to the model.
  • The providers we use do not use data submitted via the API to train models. Nor do we train models on your data.
  • Providers may briefly retain requests to detect abuse, in accordance with their terms.
  • The details, including transfers of data outside the EEA, are described in the Privacy Policy.

Do not enter data into Boardeon that you do not need for the analysis — especially special categories of data, such as health data.

6. Limitations you must keep in mind

Language models work on probability. They can:

  • state false information with complete confidence,
  • misread data or the context of a question,
  • rely on knowledge that is out of date on the day of the response,
  • reproduce biases present in the data they were trained on,
  • give different answers to the same question.

That is why you should check any material information before you base a decision on it — especially in legal, tax, financial and HR matters. Boardeon's results are not professional advice. Whoever makes a decision based on AI responses is responsible for it.

7. Prohibited uses

You may not use Boardeon's AI Features for:

  • practices prohibited by Article 5 of the AI Act, including manipulating people, exploiting their vulnerabilities and social scoring,
  • taking decisions about specific individuals based solely on AI responses — on hiring, dismissal, remuneration, employee evaluation, granting credit or contract terms,
  • creating content that is unlawful, misleading or infringes the rights of others,
  • presenting AI responses to third parties as the opinion of a human expert if they have not been checked by a human,
  • attempting to bypass model safeguards, or to extract system instructions or other customers' data.

A breach of these rules is a breach of the Terms of Service.

8. Costs and limits

Each Organisation has a monthly model-cost budget and limits on the number of requests. When the budget is used up, AI Features wait for it to renew, while the rest of Boardeon keeps working. More powerful models are available in higher Plans or with your own OpenRouter key.

9. Reporting errors

If an AI response is wrong, harmful or infringes someone's rights, write to support@boardeon.com. Reports are read by a human. We use them to improve instructions and safeguards — without passing your data on for model training.

10. Changes

We change models and providers when better or safer ones become available. Material changes — for example, a new provider processing data — are announced in accordance with the Terms of Service and the Data Processing Agreement. Previous versions of this document remain available.