Reducing identity verification costs with an AI-powered KYC platform

Challenges

To reduce the high cost of third-party identity verification, an iGaming technology company partnered with Tameshi to build its own AI-powered KYC solution. The serverless platform automatically extracts data from identity documents and performs multi-layer fraud detection with jurisdiction-specific requirements - all while reducing verification costs to just $0.03 - 0.05 per transaction.

The solution

Tameshi designed and built a serverless, AI-powered KYC platform and to maximize both accuracy and efficiency. The solution combines traditional algorithms with AI.

Mathematical validation checks document structure, image quality, data consistency, MRZ checksums and other objective indicators, while AI analyzes contextual signals that are difficult to capture with deterministic rules alone. This includes:

• validating document content against country-specific formats,
• names,
• cultural patterns,
• detecting unusual characters or inconsistencies,
• analyzing security features such as shadow images to identify document type and page,
• performing facial recognition to compare the document photo with the document holder.

Together, these validation layers help identify anomalies that may indicate fraud.

The platform uses Amazon Textract for supported document types and automatically switches to AI-based OCR for documents written in languages not fully covered by traditional OCR engines, enabling reliable verification across multiple alphabets and regions. Documents flagged as suspicious at any stage are automatically routed for manual review together with the complete verification history.