Warehows.
← Case Studies

Finsights AI

Multi-tenant embedded analytics with sub-second query performance for a finance SaaS serving 80+ tenants.

Headline outcome

80+

Tenants served

FivetranFivetrandbtAirflowSnowflakeSigmaSigma

The challenge

Where they started.

A finance SaaS with no technical team trying to serve 80+ customers with custom Power BI reports — built by hand, per customer, every month. No ingestion pipeline. Excel-driven QuickBooks exports. No multi-tenant data model. Reporting was the product, but the data infrastructure couldn't scale past a few dozen tenants.

The approach

What we actually built.

  1. 01

    Unified ingestion layer

    Fivetran for QuickBooks + bank feeds. Custom Python in Airflow for the edges. Per-tenant isolation by design.

  2. 02

    Full-stack analytics application

    Onboarding flow, multi-tenant data models, and an in-app chat service. Embedded Sigma for per-tenant dashboards.

  3. 03

    dbt transformation

    One codebase, many tenants. Row-level security enforced in the warehouse. Tests on every critical metric.

  4. 04

    LLM-powered insights

    AI-generated daily and weekly commentary on each tenant's numbers — shipped as part of the product.

Results

The numbers that shifted.

80+

Tenants onboarded

<1s

Query response at p95

90%

Manual reporting eliminated

Next case study

Boldspace

Real-time customer feedback analytics with AI-powered insights across 50K+ reviews for a creative and PR agency.

Read next

Not sure which fits?

30 minutes. We'll tell you honestlywhat's broken.

Book a discovery call