Bank AI Agent for Banking Process Automation and Support

Problem

The goal was not just to connect multiple services, but to build a unified AI automation system that synchronizes banking data, reduces manual work, and helps answer user questions based on up-to-date information.
The main challenges included:

• combining data from two bank APIs into one workflow
• synchronizing information with Google Sheets and Airtable
• automating processing, filtering, and batch operations
• ensuring real-time data consistency across systems
• providing users with an AI agent for information search and policy Q&A
• implementing smart retrieval with embeddings and a vector store
• adding validation, error handling, and resilience to complex workflows

Process

The project was built around data orchestration, AI retrieval, and reliable automation logic.

• designed n8n workflows for integrating two banking APIs
• implemented synchronization with Google Sheets and Airtable
• added HTTP requests, data merging, country-based splitting, and filtering logic
• introduced batch processing and real-time sync
• developed an OpenAI-based AI agent for information search, calculations, and policy question answering
• connected embeddings generation and a vector store for smart retrieval
• configured validation, error handling, and data consistency control

Solution

The final result was an AI-driven banking automation system that combines workflow automation, data synchronization, and conversational support in one solution.

Key solutions delivered:

• dual bank API integration
• n8n workflow orchestration
• Google Sheets + Airtable synchronization
• OpenAI conversational agent
• policy Q&A and information search
• average amount calculations
• batch processing and real-time sync
• validation and error handling
• embeddings generation and vector store
• infrastructure for smarter retrieval and consistent operations