Gen 2.0 AI Assist Framework
Client Context
Pan-regional bank operating in 32 markets with 8.5M+ customers. Required faster releases and consistent behavior.
Scope
- Web Applications
- API Integration
- 32 Market Variants
Challenges
- Lagging Automation: Scripts consistently 2-3 sprints behind development
- Flaky Regression: Inconsistent behavior between Web UI and API layers
AI-Assist Architecture
IGS Gen 2.0 Platform
Unified architecture using Playwright with MCP as automation core
AI-Driven Generation
AI assist suggests test scenarios and generates skeleton code from requirements
Business Function Modelling
Core journeys modeled as functions with formalized stable locator contract
Hybrid Orchestration
Seamless integration of existing Karate API suites within web flows
Technology Stack
Outcome
60-70%
Drastic reduction in regression execution time across all 32 market variants
Shift Left: In-Sprint
Backlog Eliminated
Mobilization: 4-5 Weeks
To First Pack
Significantly reduced manual regression effort. Automation now runs effectively for feature branches via GitHub Actions