AI transparency
Effective: 26 September 2026
Draft — legal review and completion of missing details required before going live.
The campaign is supported by the PitchArena system, which uses artificial intelligence in several steps. Everything that goes out — presentations, e-mails — is reviewed and approved by a human first.
Where do we use AI?
| Task | What the AI does | Human control |
|---|---|---|
| Market research | Web search and summaries on sport popularity and sponsorships — with sources | Data is checked, approved or rejected |
| Ideas | Sport-specific VR, AR and AI activation proposals | Every idea needs human approval before use |
| Presentations | Tailoring template copy to the company’s sponsorship | Reviewed, edited and published manually |
| E-mails | Drafting the outreach e-mail | Every e-mail is approved by a human; nothing is sent automatically without approval |
| Images | Generating illustrations; logo recognition on public photos | AI images are labelled “AI illustration” and never depict real, identifiable athletes |
Models
Depending on the task: Anthropic Claude, OpenAI (GPT and image generation), Google Gemini, Perplexity, DeepSeek. The model and prompt version used are logged for every generated item.
Limitations
AI can be wrong. That is why every market data point carries a source, and incorrect data is corrected or removed on request: szbudahazy@arworks.com.
Automated decisions
There is no automated decision-making that produces legal effects concerning you or similarly significantly affects you.