Finance ops · Production

Finch: Uber's Conversational AI Data Agent for Real-Time Financial Insights

The problem

Uber's financial analysts faced slow, inefficient data access requiring manual searches across multiple platforms, complex SQL query writing, or data request submissions that could take hours or days — causing delays that impacted real-time decision-making.

Workflow diagram · grounded in source
1
User query in Slack
trigger
“User query input. A finance team member asks Finch a question in Slack.”
2
Supervisor agent routing
routing
“The Supervisor Agent receives the query and sends it to an appropriate sub-agent, such as the SQL Writer Agent. The appropriate agents are selected based on the nature of the query.”
3
Metadata retrieval from OpenSearch
ai_action
“The appropriate agents can query the OpenSearch index to fetch relevant metadata, including natural language aliases for both column names and values. This enhances the LLM's ability to construct precise SQL queries.”
4
SQL query construction and execution
ai_action
“The SQL Writer Agent dynamically builds the query using the retrieved metadata and executes it against the appropriate data source.”
5
Security permissions validation
validation
“Validates Taya's security permissions before executing the query.”
6
Result delivery to Slack
output
“the results are formatted and delivered back to Slack in a clear, actionable format, with details on the executed query and a link to a Google Sheet with the query results.”
Reported outcome

Finch eliminates friction in financial data retrieval for Uber finance teams by enabling conversational natural language queries in Slack, leading to less friction, fewer delays, and faster data-driven decisions.

Reported metrics
Data retrieval friction and delaysless friction, fewer delays, and faster data-driven decisions
Data retrieval complexityreduces the complexity of financial data retrieval
Reported stack
FinchRAGLangChain LanggraphOpenSearchSlackGenerative AI GatewaySlack AI Assistant APIsPrestoIBM Planning AnalyticsOracle EPMGoogle Sheets
Source
https://www.uber.com/en-IN/blog/unlocking-financial-insights-with-finch/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Finch eliminates friction in financial data retrieval for Uber finance teams by enabling conversational natural language queries in Slack, leading to less friction, fewer delays, and faster data-driven decisions.

What tools did this team use?

Finch, RAG, LangChain Langgraph, OpenSearch, Slack, Generative AI Gateway, Slack AI Assistant APIs, Presto, IBM Planning Analytics, Oracle EPM.

What results were reported?

Data retrieval friction and delays: less friction, fewer delays, and faster data-driven decisions; Data retrieval complexity: reduces the complexity of financial data retrieval (source-reported, not independently verified).

How is this finance ops AI workflow structured?

User query in Slack → Supervisor agent routing → Metadata retrieval from OpenSearch → SQL query construction and execution → Security permissions validation → Result delivery to Slack.

Related finance ops cases