LCL Bank Automates KYC with AI Document Extraction

Challenge

LCL Bank faced a massive regulatory compliance challenge that threatened to overwhelm their manual processes. As one of the most regulated sectors in the world, banks must maintain strict Know Your Customer (KYC) standards and comply with evolving Corporate Social Responsibility (CSR) requirements.

The compliance requirements were overwhelming.

• Millions of customer profiles and documents to process
• New regulations required new types of documents to process
• Traditional approaches couldn't scale due to the variety of formats, including non-standardized documents
• Regulatory deadlines needed to be met

"We cannot afford any mistakes given the importance of our two missions. The challenge resides in the high expectations from the business units (Compliance Department, CSR Management)."

Axel Cypel
‍AI Expert at LCL

Read their compliance article

Solution

LCL chose to build AI-powered information extraction algorithms using supervised machine learning, partnering with Kili Technology to create the high-quality training datasets essential for success.

Their labeling strategy:

Secure deployment: LCL implemented Kili Technology's labeling platform on-premises, ensuring data security while enabling their team of five dedicated labelers to work efficiently on entity identification, classification, and annotation tasks.

Using Kili Technology's SDK, LCL integrated its existing systems to:

• Extract tens of thousands of images from their Enterprise Data Management system
• Apply custom OCR models to identify areas of interest
• Prepare documents for manual annotation through automated pre-processing
• Eliminate previous pain points like lost work and complex file management across multiple Excel files

Collaborative Approach: The platform's user-friendly interface enabled broader participation across the organization. Business unit experts could easily contribute to the AI training process, creating company-wide engagement around AI initiatives.

Scalable Annotation Process: LCL established a standardized workflow where each annotation campaign typically processed 5,000 documents in 2-3 weeks, supported by dedicated customer success management from Kili Technology.

Outcome

LCL's AI-powered document extraction system delivered transformational results, processing regulatory compliance at unprecedented scale.

What they've achieved

• The trained models successfully processed more than 13 million documents in production batches, automatically checking KYC completeness across LCL's entire customer base.
• LCL met strict regulatory deadlines that would have been impossible with manual processes.
• For extra-financial reporting requirements, LCL's algorithms now automatically extract relevant information from DPE documents during the credit process—a task that would be prohibitively expensive and time-consuming if done manually.
• Eliminated previous workflow inefficiencies, replacing nightmarish file management with a streamlined, secure platform that met LCL's stringent data safety standards.
• LCL now has a scalable AI infrastructure that can be rapidly deployed for new use cases, from processing voice recordings with speech-to-text algorithms to analyzing various customer data formats including images, emails, and contracts.

"Thanks to the fact that our AI infrastructure now includes Kili Technology, we can use the tool for all kinds of projects... LCL teams can accelerate drastically the creation of their training datasets, which means a significant improvement for all the parties involved."

Axel Cypel
‍AI Expert at LCL