The Challenge
With over $800 billion in assets, this bank operated dozens of internal systems — from mainframe core banking to modern CRM platforms, compliance databases, and document repositories. Customer-facing AI agents needed to pull context from 12 different sources to answer a single question.
The existing approach relied on custom API integrations and batch ETL pipelines that introduced 4-6 hour data lag. Agents frequently made decisions on stale data, leading to compliance flags and customer complaints.
The Solution
mantle's zero-copy connectors were deployed across all 12 data sources in under two weeks. The semantic knowledge graph automatically resolved customer entities across systems — matching account numbers, SSNs, and relationship hierarchies in real time.
AI agents now receive quality-scored, compliance-filtered context via a single MCP call. Every response includes provenance tracking, so relationship managers know exactly where each data point originated.
Results
73%
Faster agent response time
94/100
Average context quality score
12
Data sources connected
<100ms
Context delivery latency
“mantle eliminated the data lag that was our biggest pain point. Our AI agents went from guessing to knowing — and our customers can feel the difference.”
— VP of Digital Banking, Top-10 US Bank