Customer support · Production

Santalucía Seguros builds enterprise-level RAG Virtual Assistant for insurance agents on Databricks

The problem

Insurance agents at Santalucía Seguros needed to access large volumes of documentation spread across multiple locations and formats to answer customer queries quickly, while customers demanded personalized, fast, and efficient service.

Workflow diagram · grounded in source
1
Agent queries VA in Teams
trigger
“The VA is accessed within Microsoft Teams and is able to answer agent questions in natural language on any mobile device, tablet, or computer, in real-time, with 24/7 availability”
2
RAG retrieves relevant docs
ai_action
“This architecture enables the continuous ingestion of up-to-date documentation into embedding-based vector stores, which provide the ability to index information for rapid search and retrieval”
3
GenAI VA generates response
ai_action
“implemented a GenAI-based Virtual Assistant (VA) capable of supporting agents' queries about products, coverages, procedures and more”
4
LLM-as-judge validates quality
validation
“The LLM-as-a-judge consists of natural-language-based criteria for measuring accuracy, relevance, and coherence between expected answers and those provided by the VA”
5
Agent receives real-time answer
output
“whenever a customer asks about coverage they can get an answer in seconds”
Reported outcome

The GenAI-based Virtual Assistant exceeded user expectations, enabling agents to answer customer coverage queries in seconds, positively impacting customer satisfaction, and accelerating product sales.

Reported metrics
Coverage query response timeanswer in seconds
Customer satisfaction impactpositively impacts customer satisfaction
Product sales accelerationaccelerates the sale of products
Agent daily work easemuch easier
Show all 5 reported metrics
coverage query response timeanswer in seconds
customer satisfaction impactpositively impacts customer satisfaction
product sales accelerationaccelerates the sale of products
agent daily work easemuch easier
user expectations metexceeded the expectations of users
Reported stack
DatabricksMicrosoft TeamsRAGMLflowDatabricks Model ServingGPT-4Azure OpenAI APIAzure Key vaultDatabricks MarketplaceMicrosoft Azure
Source
https://www.databricks.com/blog/santalucia-seguros-enterprise-level-rag
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The GenAI-based Virtual Assistant exceeded user expectations, enabling agents to answer customer coverage queries in seconds, positively impacting customer satisfaction, and accelerating product sales.

What tools did this team use?

Databricks, Microsoft Teams, RAG, MLflow, Databricks Model Serving, GPT-4, Azure OpenAI API, Azure Key vault, Databricks Marketplace, Microsoft Azure.

What results were reported?

Coverage query response time: answer in seconds; Customer satisfaction impact: positively impacts customer satisfaction; Product sales acceleration: accelerates the sale of products; Agent daily work ease: much easier (source-reported, not independently verified).

How is this customer support AI workflow structured?

Agent queries VA in Teams → RAG retrieves relevant docs → GenAI VA generates response → LLM-as-judge validates quality → Agent receives real-time answer.

Related customer support cases