The situation
This global financial services firm is known for its personal approach to investing and financial
planning for individual investors and small business owners. They operate through a network of branch offices across the U.S. and Canada, each typically staffed by a financial advisor and a branch office administrator. This structure allows the firm to provide highly personalized services tailored to individual financial goals and circumstances.
Complication
In-store advisors were spending excessive time searching for the information they needed in order to answer customer questions during store interactions. Often they had to reach out to a central team of domain experts who would sift through a knowledge base of internal documents to provide the personalized advice to be shared with customers. This led to poor customer experience and loss in staff productivity.
The Solution
The company decided to leverage GenAI technology to quickly retrieve relevant information from internal PDF documents and combine that with customer information in user profiles to generate succinct responses to customer queries. The key challenges were avoiding hallucinations and providing accurate responses that could be traced back to the original sources.
With a solution powered by Dataworkz RAG builder and MongoDB Atlas Vector Search, the company was able to use advanced RAG techniques to transform raw data (Slack, JIRA and PDFs) and rapidly build and deploy a production-ready GenAI application fueled by this data.
The application enabled frontline service advisors to instantly tap into their internal knowledge base without the need to manually look for information across multiple documents. This helped improve in-store customer experience and led to faster responses and higher productivity by reducing the number of requests sent to a central team of domain experts.
The Impact
3 months vs 9 months to get to production
<1 day vs 3-6 months for new use case
1 FTE vs 9 FTEs for app development