Automating insurance claims across two markets with a conversational AI platform
How Softescu built a dual-component claims automation platform for a leading insurance broker — an AI-powered chatbot for customers and a custom middleware connecting it to the client's core insurance systems — deployed across two territories in Central and Eastern Europe.
Technologies: Conversational AI, Drupal, PHP, MySQL, MariaDB, AWS KMS, REST API, OAuth2, JavaScript
The Challenge
A leading insurance broker needed to automate the mobile device insurance claims process for customers across a first market — and later expand the solution to a second territory in Central and Eastern Europe. The existing process required customers to contact agents directly, creating delays and manual overhead across multiple claim types: accidental damage, theft, loss, liquid damage, and hardware failures, each with its own policy rules and documentation requirements.
The solution required two tightly integrated components: a customer-facing conversational interface to guide claimants through the process, and a secure middleware layer to validate policies in real time and synchronise claim data bidirectionally with the client's existing infrastructure. Both components had to meet strict insurance industry security requirements and operate across two distinct market configurations.
Our Solution
Softescu built a dual-system platform consisting of an AI-powered chatbot and a custom middleware application. The chatbot guides customers through dynamic claim submission workflows, automatically adapting conversation paths based on claim type — with distinct flows for theft and loss scenarios requiring police report documentation, and separate paths for accidental damage, liquid contact, and hardware failures. The chatbot integrates with the middleware in real time to validate policy coverage before accepting a claim.
The middleware exposes REST APIs that handle bidirectional data exchange with the client's insurance platform: automated policy imports and claim data exports keep both systems synchronised. The architecture was secured with OAuth2 authentication, AWS KMS encryption at rest, IP access controls, and Shield module protection. When the client expanded to a second market in Central and Eastern Europe, the platform was extended with a market-specific configuration split, a localised admin UI, and separate production environments — without requiring a rebuild of the core system.