Summary
Zensar worked with Old Mutual to leverage Claude models, including Claude Sonnet and Claude Haiku, through AWS Bedrock to develop its AI-powered financial wellness chatbot. The solution improved the chatbot's ability to understand customer intent, generate contextually relevant responses, and provide conversational access to information on financial products, insurance offerings, policies, and services.
Challenges
Enhancing a customer-facing financial wellness experience:
Old Mutual sought to develop an AI-powered chatbot that serves as a conversational channel for customers seeking information on financial products, insurance offerings, policies, and services.
Improving conversational quality and customer engagement:
As part of development, there was a need to strengthen the chatbot's ability to understand customer intent, provide contextually relevant responses, and support a broader range of customer queries.
How Claude helped:
Zensar integrated Claude Sonnet and Claude Haiku through AWS Bedrock to enhance the chatbot's conversational capabilities, enabling more effective question handling and information delivery across financial wellness and insurance topics.
Solution
Claude models on AWS Bedrock:
Claude Sonnet and Claude Haiku were integrated into the chatbot solution through AWS Bedrock to power customer interactions.
Model optimization:
Claude Sonnet was utilized for complex queries and routing requiring deeper contextual understanding, while Claude Haiku supported efficient handling of routine customer interactions.
Knowledge-grounded responses:
The models were integrated with Old Mutual's enterprise knowledge sources to provide accurate, relevant, and business-aligned responses.
Business outcome
The implementation of Claude Sonnet and Claude Haiku enhanced the chatbot's capabilities by:
Improving understanding of customer intent and query context.
Delivering more relevant and conversational responses.
Expanding support for financial wellness and insurance-related inquiries.
Enhancing customer access to information through natural language interactions.
Strengthening the overall customer experience.
Conclusion
Large language models can significantly enhance customer-facing chatbot experiences when combined with trusted enterprise knowledge sources.
The use of Claude Sonnet and Claude Haiku can improve the chatbot's ability to understand customer questions and deliver relevant responses, demonstrating the value of advanced language models in supporting customer engagement and self-service.