About the Client
The client is a legal access company offering insurance plans to employers. It aims to provide safety and security to employees with legal benefits and experienced attorneys. A legal insurance company, which has been established for more than 50 years, specializes in connecting the right attorneys to cater to the specific legal needs of its clients. By providing professional service, they save employers and employees money, time, and stress.
Business Challenges
Unstructured Data
The client has a huge pool of policy data, making it difficult to provide faster responses to customer queries.
Chatbot Development
The client wanted to create an RAG-based Gen AI chatbot for internal teams to quickly access the specific information from documents and respond to policyholders’ queries.
Microsoft Teams Integration
The client wanted all three 3 chatbots, i.e., Member Services, Sales/Account, and Helpdesk Support Chatbots, to be integrated with their existing Microsoft Teams platform.
Kasmo’s Solutions
Kasmo’s Snowflake Expertise understood these challenges and built a Snowflake Cortex environment to support a Retrieval-Augmented Generation (RAG) chatbot for semantic search. Embed and index both structured and unstructured data and trained large language models to meet the client’s requirements. We implemented 3 chatbots-
Member Services Chatbot
- The chatbot leverages our Olaf solution. Trained to identify member ID-based queries to deliver contextual and policy-specific answers to their insurance holders.
- Integrated policy PDF in the system to help members fetch accurate and document-based information, leading to minimized effort and time.
- Enhanced natural language understanding (NLU) to enable semantic accuracy. This chatbot is built in a conversational tone for a natural member experience.
Sales / Account Managers Chatbot
- The chatbot solution leverages our Nimble search and helps to handle aggregated analytical queries (e.g., SLA trends, resolution times, outlier detection).
- Provides a summarization of historical case patterns to gain hidden and valuable insights.
- Here, models are trained to recommend prioritized actions, analyzing the policy accounts and performance trends.
Helpdesk Support Chatbot
- Trained the model to match new queries with historical service tickets using Kasmo’s document-based chatbot –Olaf.
- Optimized the chatbot to reduce agent response time and improve overall resolution efficiency.
- Helps the internal team to get specific answers easily from their website with our chatbot integrated with Microsoft Teams.
Results Achieved
Through Kasmo’s AI-driven chatbot solutions, the client achieved:
Faster query responses: These chatbots help our clients to gain accurate policy information, provide quicker responses, and enhance customer service.
Gain Data-driven Insights: The client was able to gain real-time insights and SLA trends through analyzing historical data.
Higher Efficiency: Reduced operational overhead for support agents with simple and quicker access to information.
