Shastic

Overview

Shastic is an AI workflow-automation platform for banks and credit unions, focused on digital lending and loan origination. Its product surface includes a family of named AI agents and services, the Maggie AI assistant, the Elle service family, and a ReCapture module, that automate lending workflows such as application tracking, borrower onboarding, abandoned-application recapture and HMDA compliance, accessible to loan officers through a web app and a browser extension.

The challenge

Loan origination at banks and credit unions is heavily manual: applications stall, abandoned applications are rarely re-engaged, and regulatory reporting like HMDA adds overhead. These institutions also operate under strict security and compliance expectations, so any automation platform has to be built on infrastructure that can stand up to SOC2-level scrutiny while scaling across many lending workflows.

What we built

Shastic runs on an extensive AWS architecture defined as CloudFormation infrastructure-as-code. Compute spans ECS and auto-scaling EC2 for the Elle microservices (ElleChat, ElleAuth, ElleAI) alongside Lambda functions, with Amazon RDS (PostgreSQL) and ElastiCache (Redis) for data, and SageMaker, Step Functions, Glue, Kinesis and Athena powering the ML and data pipelines. API Gateway fronts the services, EventBridge and SNS/SQS handle messaging, and Secrets Manager, CloudTrail and SOC2-oriented IAM roles enforce the security posture. The product layer delivers AI lending agents, application recapture, an application tracker, onboarding automation, an ROI calculator and a browser extension for loan officers.

Results

Shastic reduces the manual work in loan processing through AI-driven lending workflows and re-engages abandoned loan applications via automated recapture. HMDA compliance is built in, and the whole platform sits on a SOC2-aligned AWS architecture designed for the security expectations of banks and credit unions.