Challenge
A mortgage financing company, suffered from inefficiencies throughout its complex and highly manual workflows and operations.
The company sought a rigorous, AI-enabled, and data-rich approach to process mapping and optimization, with the goal of improving ROI from process improvement initiatives.
Approach
- Collected, transformed, and integrated activity log data from disparate systems to create a holistic dataset of selected processes suitable for process mining
- Visualized end-to-end workflows (happy-paths and all the variants) to identify inefficiencies and opportunities
- Validated and prioritized complexities and opportunities in workshops with business stakeholders
- Developed an AI tool and a model using Anthropic/Claude and Streamlit Python as a POC to demonstrate potential of digital twin modelling
Results
Compressed process mapping from weeks to days, revealing previously hidden complexities
Improved visibility of path variants and dependencies, not only ‘happy path’ mapping
Identified operating cost savings of 20% by optimizing manual workflows with low-cost AI tools
Developed a repeatable, low-touch, scalable approach to be rolled out across the enterprise