Client name
PKO BP
Client type
Bank
Industry
Retail and corporate banking
Context
Banking procedures require that a property (house or apartment) be covered by appropriate property insurance. Therefore, the client is obligated to present a valid insurance policy that must meet the bank’s specific formal and financial requirements.
In practice, each of these policies was manually reviewed by PKO BP employees to ensure it complied with loan conditions. Key elements subject to verification included:
the insured amount (whether it matches the property’s value),
the scope of insurance coverage (e.g., fire, flooding, mechanical damage),
the property owner’s information,
policy number and validity dates,
insurer details,
mortgage loan number compliance,
clauses assigning rights from the policy to the bank.
In total, 15 data fields had to be correctly transcribed and verified against the data in the bank’s systems and the loan agreement. This process was manual, error-prone, and time-consuming.
The bank processed over 2,000 policies daily, requiring the involvement of more than 60 operations staff. This extended the time needed for document evaluation, negatively affected customer service time, and generated significant operational costs. Moreover, verification errors could result in a breach of loan agreement terms and expose the bank to legal risk.
Effects and results obtained
Reduction of 60 verification team members previously required for document processing
Analysis and interpretation of over 2,000 policies per day
Document processing time shortened by 75x – from over 10 minutes per document to just a few seconds
Elimination of manual data entry – 15 defined fields are automatically extracted
Automated recognition, interpretation, and classification of documents
Client goals
Automate data verification from documents (insurance policies and various types of correspondence)
Eliminate human errors and reduce processing time
Lower operational costs in the credit department
Starting point
Manual verification of over 15 fields from each home insurance policy
Scalability challenges due to manual data entry
High operational labor costs (over 60 employees involved in this specific process)
Frequent errors in manual verification (potentially leading to breaches of loan agreements and legal risk for the bank)
Document types covered by the implementation
Home and property insurance policies
Format: PDF, JPG, PNG
Scope and course
Why Nu Delta:
Comprehensive banking outsourcing combined with proven experience in implementing the Nu Delta docAI platform across various banking processes, delivering high data extraction accuracy (>95%).
What we delivered:
Data extraction, document classification, validation, and export to banking systems.
Technology:
Proprietary AI platform + NLP + OCR; SaaS model.
Security:
Full GDPR compliance, data processed exclusively within the EU, data encryption, and storage in the client’s private cloud.
Integrations and how it works
Integration with PKO credit systems
Automatic document import and data export to the bank’s workflow
……