It support · Production

HP builds a RAG knowledge chatbot and self-service analytics on Databricks Mosaic AI, reducing data team support burden and cutting platform costs by 20–30%

The problem

HP's data engineering teams were overwhelmed by partner and customer support requests, spending 20–30% of their time manually writing SQL queries and cross-referencing data—equivalent to a full additional headcount on a five-person team. Knowledge was scattered across internal wikis, SharePoint files, and team channels, hampering onboarding and real-time decision-making.

Workflow diagram · grounded in source
1
User submits support inquiry
trigger
“Whether users needed access to restricted data or wondered how to use platform features or new employee onboarding, the teams were constantly distracted by customer support requests”
2
Crawler indexes internal knowledge
ai_action
“a web crawler that crawls and tokenizes the internal information on various sites and gets them populated into the Vector Search database”
3
Agent parses and retrieves
ai_action
“a back-end agent that essentially parses the input from the user, searches and grabs relevant data”
4
GenAI endpoint generates answer
ai_action
“then sends it to the GenAI endpoint to generate answers”
5
Chatbot delivers response
output
“a functional assistant chatbot that provided accurate answers in real time”
6
Genie natural language query
ai_action
“AI/BI Genie can play a significant role in reducing some of the manual SQL support required from our data engineers today, by providing the ability to converse with the data in natural language and create data visualization easily and ni…”
Reported outcome

HP achieved 20–30% lower operational costs compared with their AWS Redshift warehouse and forecasts significant productivity gains as the AI agent chatbot reduces the need for manual partner support.
An intern implemented the end-to-end RAG chatbot in less than three weeks.

Reported metrics
Data team time on manual SQL queries20–30%
Manual support burden as FTE equivalentequates to an additional full-time employee
Operational cost savings vs AWS Redshift20–30% lower costs
End-to-end RAG chatbot implementation timeless than three weeks
Show all 5 reported metrics
Data team time on manual SQL queries20–30%
Manual support burden as FTE equivalentequates to an additional full-time employee
Operational cost savings vs AWS Redshift20–30% lower costs
End-to-end RAG chatbot implementation timeless than three weeks
Forecasted productivity gains in data teamssignificant productivity gains
Reported stack
Databricks Data Intelligence PlatformDatabricks Mosaic AIDatabricks SQLUnity CatalogDBRXVector SearchAgent Bricks AI PlaygroundAI/BI GenieAWS
Source
https://www.databricks.com/customers/hp/gen-ai
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

HP achieved 20–30% lower operational costs compared with their AWS Redshift warehouse and forecasts significant productivity gains as the AI agent chatbot reduces the need for manual partner support.

What tools did this team use?

Databricks Data Intelligence Platform, Databricks Mosaic AI, Databricks SQL, Unity Catalog, DBRX, Vector Search, Agent Bricks AI Playground, AI/BI Genie, AWS.

What results were reported?

Data team time on manual SQL queries: 20–30%; Manual support burden as FTE equivalent: equates to an additional full-time employee; Operational cost savings vs AWS Redshift: 20–30% lower costs; End-to-end RAG chatbot implementation time: less than three weeks (source-reported, not independently verified).

How is this it support AI workflow structured?

User submits support inquiry → Crawler indexes internal knowledge → Agent parses and retrieves → GenAI endpoint generates answer → Chatbot delivers response → Genie natural language query.

Related it support cases