Back office ops · Production

Blend reduces time-to-value by 4 months using dbt Cloud and Monte Carlo

The problem

Blend's data team relied on manually scheduled SQL queries in Airflow that drained engineering resources and made product analytics fully dependent on data engineering for every new workflow request. Insufficient data quality monitoring led to a revenue-skewing incident caused by poor data quality calculation.

First attempt

Blend's in-house POC data quality framework used orchestrated validation queries that completely overwhelmed the Redshift warehouse with CPU-heavy, slow-running queries, making it unviable as a production solution.

Workflow diagram · grounded in source
1
Multi-source data intake
trigger
“we have around 32 sources of data, and 12 different teams utilize the data on top of our platform”
2
dbt Cloud SQL transformation
integration
“the product analytics team uses dbt's modular SQL framework to pre-process complex SQL operations before creating the dashboards that will ultimately be consumed by internal and external users”
3
ML anomaly monitoring
ai_action
“Monte Carlo's ML-powered monitors were instantly deployed across 100% of Blend's production tables out-of-the-box— evaluating freshness, volume, and schema changes automatically without any configuring or thresholding”
4
Real-time Slack alerting
output
“Monte Carlo would be there first with automated alerting through Slack to alert to breakages in real-time”
5
Field-lineage root-cause
validation
“Monte Carlo's automatic field-lineage also meant the team could root-cause incidents at a glance, resulting in faster resolutions and a deep understanding of the downstream impact of anomalies across pipelines”
Reported outcome

Blend reduced time-to-value by 4 months compared to their internal POC framework, gained automated data quality coverage across all production tables, and reduced warehouse compute costs through metadata-based monitoring and pre-processed SQL operations.

Reported metrics
time-to-value reduction vs internal POC4 months
production tables with ML monitoring coverage100%
Speed-to-insights improvementDramatically improved speed-to-insights
Warehouse and compute costsReduced warehouse and compute costs
Reported stack
dbt CloudMonte CarloAirflowSlackSnowflake
Source
https://www.getdbt.com/case-studies/blend
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Blend reduced time-to-value by 4 months compared to their internal POC framework, gained automated data quality coverage across all production tables, and reduced warehouse compute costs through metadata-based monitor…

What tools did this team use?

dbt Cloud, Monte Carlo, Airflow, Slack, Snowflake.

What results were reported?

time-to-value reduction vs internal POC: 4 months; production tables with ML monitoring coverage: 100%; Speed-to-insights improvement: Dramatically improved speed-to-insights; Warehouse and compute costs: Reduced warehouse and compute costs (source-reported, not independently verified).

What failed first in this deployment?

Blend's in-house POC data quality framework used orchestrated validation queries that completely overwhelmed the Redshift warehouse with CPU-heavy, slow-running queries, making it unviable as a production solution.

How is this back office ops AI workflow structured?

Multi-source data intake → dbt Cloud SQL transformation → ML anomaly monitoring → Real-time Slack alerting → Field-lineage root-cause.

Related back office ops cases