Finance ops · Production

Vanguard builds Virtual Analyst to reduce time-to-insight from days to minutes using AWS and Amazon Bedrock

The problem

Vanguard's financial analysts and business stakeholders needed faster access to complex financial data, but the existing workflow required SQL expertise and data team support, with typical requests taking several days to fulfill.

Workflow diagram · grounded in source
1
Analyst asks in natural language
trigger
“business analysts can express their understanding of data relationships in natural language that can be interpreted and translated into structured SQL queries”
2
Foundation model interprets query
ai_action
“Amazon Bedrock for foundation models that power natural language understanding”
3
Semantic layer generates SQL
ai_action
“translates business definitions, rules, and ontologies into executable logic that standardizes how your organization defines key metrics and the relationships between different data elements”
4
Automated data quality checks
validation
“Vanguard set up observability tools to monitor data reliability through automated checks”
5
Query executes against Redshift
integration
“Amazon Redshift for centralized data warehousing”
6
Insights delivered to analyst
output
“Reduced time-to-insight from days to minutes for complex financial queries with the use of the Virtual Analyst”
7
Feedback refines the model
feedback_loop
“user feedback loops to evolve examples and semantics, with automated alerts for model degradation”
Reported outcome

Virtual Analyst reduced time-to-insight from days to minutes, enabled business users to access data independently without SQL knowledge, achieved high accuracy in AI-generated SQL queries, decreased data team workload for routine requests, and established a reusable framework now adopted across multiple Vanguard business units.

Reported metrics
Time-to-insight for complex financial queriesdays to minutes
typical request fulfillment time (pre-Virtual Analyst baseline)several days
AI-generated SQL query accuracyhigh accuracy
Data team workload for routine analytical requestsDecreased
Show all 5 reported metrics
time-to-insight for complex financial queriesdays to minutes
typical request fulfillment time (pre-Virtual Analyst baseline)several days
AI-generated SQL query accuracyhigh accuracy
data team workload for routine analytical requestsDecreased
ground truth exemplars builtover 50
Source
https://aws.amazon.com/blogs/machine-learning/building-ai-ready-data-vanguards-virtual-analyst-journey/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Virtual Analyst reduced time-to-insight from days to minutes, enabled business users to access data independently without SQL knowledge, achieved high accuracy in AI-generated SQL queries, decreased data team workload…

What results were reported?

Time-to-insight for complex financial queries: days to minutes; typical request fulfillment time (pre-Virtual Analyst baseline): several days; AI-generated SQL query accuracy: high accuracy; Data team workload for routine analytical requests: Decreased (source-reported, not independently verified).

How is this finance ops AI workflow structured?

Analyst asks in natural language → Foundation model interprets query → Semantic layer generates SQL → Automated data quality checks → Query executes against Redshift → Insights delivered to analyst → Feedback refines the model.

Related finance ops cases