Smart chatbot for insurance support on Boost.ai platform
The business needs
The client aimed to enhance the performance of the internal service desk by deploying a chatbot that could answer complex customer queries using data from insurance policy documents updated annually. A core objective was to achieve over 90% accuracy in understanding and responding to queries based on dense technical content. However, during initial evaluations, existing market solutions were found to be inadequate for managing such complexity. Not only was the challenge to train a chatbot on long-form, detailed documents, but also to ensure reliable handling of context, memory use and system scalability for potential future deployment to external users.
Our role
We collaborated closely with the client to design and implement a sophisticated, AI-powered chatbot on the Boost.ai platform. Thanks to our strategic configuration and testing approach, the solution was able to understand and respond accurately to complex queries. Our actions included:
- training the chatbot on the client’s comprehensive policy documentation to ensure highly relevant and precise responses,
- configuring advanced context and knowledge settings to support long, multi-layered conversations without compromising accuracy,
- developing AI-driven conversation testing scenarios that mimicked real-world internal queries to enable continuous refinement of the chatbot’s responses,
- collaborating directly with Boost.ai to support improvements in their generative AI models, helping the platform surpass its original accuracy goals.
The impact and benefits
Following targeted optimisations, the chatbot achieved over 95% accuracy in providing answers based on complex policy documents. Internal agents now benefit from faster and more precise access to information, which significantly reduces the time spent searching for details and increases service quality. This project has laid a solid foundation for scaling the solution externally and preparing the client for a broader rollout that could enhance customer service experiences across public-facing channels.
Contact
Let’s talk about your needs
Case Studies
