Case Studies

The Challenge

A leading B2B software provider managed thousands of service contracts across global resellers and distributors. Renewal processes were inconsistent across geographies, risk identification was manual, and frontline teams had no governance mechanism to prioritise high-risk contracts, which drove renewal rates down and operational costs up.

The Solution

Dr. Shylu John led the design of a data-driven governance model, built on an unsupervised ML (K-means) segmentation engine that classified resellers into four renewal-risk clusters: Loyal, High Performer, Medium Performer and Non-Performer. A capacity-aligned intervention strategy directed effort to the highest-risk accounts, delivered through a scalable Shiny R governance platform with real-time dashboards, automated segmentation and drill-down analytics, giving leadership 120-day advance visibility into renewal risk.

Further Outcomes

• Significantly improved renewal forecasting accuracy and governance visibility

The Challenge

Underwriters spent hours manually mining annual reports, ESG disclosures and loss reports for risk signals. The process was vulnerable to incomplete data, shifting regulations and high integration costs for new analytics tools.

The Solution

Dr. Shylu John was engaged to design and deliver a GenAI-powered risk assessment tool on Google Cloud Platform, combining a predefined Q&A module with an interactive chatbot integrated with OpenAI's GPT. It automates analysis across three report types, enabling automated risk identification and precision query handling at scale.

Further Outcomes

• Deeper, faster insight into financial stability, business plans and regulatory exposure

The Challenge

Vehicle registration documents require capturing 40+ data fields (identification, vehicle specifications, ownership and compliance certificates) for the insurance quote and buy process. Manual extraction from thousands of scanned or photographed documents was slow, labour-intensive and error-prone, directly affecting premium accuracy and regulatory compliance.

The Solution

Dr. Shylu John was engaged to design a deep learning-based OCR system trained specifically on German registration documents, built to handle real-world variability in scan quality and formats. Deployed as a scalable API service, it integrates directly into the underwriting workflow, with human oversight, fairness checks and full data-privacy compliance built in throughout.

The Challenge

A UK insurer's Employer's and Public Liability claims teams had 35 business days to propose Stage-2 settlement counteroffers under the MOJ portal. Limited visibility into historical settlement patterns, no structured way to estimate breakeven values and inconsistent approaches across teams were driving up litigation rates, solicitor fees and operational costs.

The Solution

Dr. Shylu John was engaged to build an end-to-end predictive settlement model, from data discovery and feature engineering through champion/challenger testing and cross-validation, deployed via a scored, versioned pipeline into a web-based decision-support tool. The application gives claims handlers model-guided settlement ranges, backed by audit logs and drift monitoring for ongoing governance.

The Challenge

Global supply chain disruption, rising parts and labour costs and shifting regulations across markets drove significant claims inflation, pushing up average claim costs, customer premiums and repair cycle times. The insurer had no unified way to track and act on these trends, relying instead on multiple static Excel reports across regions.

The Solution

Dr. Shylu John partnered with business stakeholders to consolidate inflation analysis, claims transformation and root-cause analysis into a single automated Power BI platform, replacing manual reporting with near real-time dashboards. The solution ingests data automatically, refreshes all filters and calculations in one operation, and gives teams analytical depth across claims costs, vehicle repairs and parts pricing, built as a standardised framework deployable across multiple Operating Entities.

Further Outcomes

• Reduced claims delays and costs through streamlined processing
• Faster claim resolution and more cost-effective repair decisions
• Strengthened competitiveness through more effective inflation management