H6.ai freezes financial test data with Lakebase branches
Understand this story
H6.ai describes its approach to AI financial automation regression testing: it creates an independent Lakebase Postgres database for each customer, and during testing branches a runtime branch from a frozen snapshot, so that invoice extraction, matching, classification, and validation run for real on data in which purchase orders, suppliers, and policy states remain unchanged; the branch is deleted after the run. H6.ai says creating a branch does not copy data; the branch shares storage with the parent database and stores only its own changes. Tests do not post to the customer's ERP, nor do they write to the customer's live database. Results are determined jointly by deterministic assertions and an AI review panel composed of models from multiple vendors, which gives a pass, fail, or uncertain verdict along with reasoning. Branch management uses a dedicated service principal with least privilege, and the branch protection mechanism refuses to touch the customer's live branch. H6.ai also says that in the future it hopes to use Lakehouse//RT to run read-only analytical queries against Lakebase tables without copying data.
AI overview based on source reports · Updated 2 days ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Databricks BlogHow H6.ai builds AI-powered financial automation with Lakebase Postgres
H6.ai automates accounts payable on top of customers' existing ERPs by combining AI-assisted invoice understanding with deterministic, auditable policy execution. To make regression testing reliable despite constantly changing financial data, it gives each customer a dedicated Lakebase Postgres database and uses copy-on-write branches to freeze snapshots, run each test on its own disposable branch, and judge results with deterministic assertions plus a panel of AI judges. The company says branching provides isolation, frozen state, cheap disposable copies and schema-safe replays, and it plans to use Lakehouse//RT queries on Lakebase tables for read-only analytics without copying data.
Activity for this story
- Current comparable activity
- 3
- Peak comparable activity
- 10Nov 13 02:00
- Change in the past 24 hours
- -50%
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.