Skidata AI - Resorts Go From Over 10,000 Manual Checks To Around 15 Targeted Checks Per Day
SKIDATA is currently conducting an exciting proof of concept for the AI Photo Check in Kitzbühel, Zermatt and Sierra Nevada – and the AI is already learning rapidly.
The challenge
- Today, lift operators often have to visually compare more than 10,000 photos to detect potential ticket fraud.
- This is strenuous, tiring and prone to errors – especially when guests wear helmets and ski goggles, which makes classic facial recognition unusable.
Skidata approach: AI Photo Check
- SKIDATA has developed its own AI for image similarity checks, which marks suspicious passes within fractions of a second.
- Staff only need to check tickets whose probability of fraud exceeds a defined threshold – which in practice means around 10-15 checks per day instead of thousands.
Important: SKIDATA AI Photo Check is not a biometric solution and does not offer facial recognition.
This is a similarity check between ticket photos – specifically developed for ski resorts and their data protection requirements.
- “Ladies Day” tickets used by men – a use case they are currently actively working on.
- Parents who accidentally use their children's tickets
- Ticket transfer: A skier uses a ticket in the morning and passes it on to someone else in the afternoon.
The AI Photo Check already helps ski resorts to detect patterns such as ticket sharing or unintentional misuse – and is continuously improving as we integrate further scenarios such as “Ladies Day” abuse.
Since the beginning of December, over 300,000 runs have already been evaluated in Kitzbühel, Zermatt and Sierra Nevada.
This is just the beginning of how AI can help mountain destinations make access smarter, fairer and easier for everyone.
