Open finance skill
Does the supporting cost detail add up to the reported total?
Foot synthetic cost-detail rows to a control total and show the exact variance without altering either source.
Guided synthetic example
The cost-detail schedule does not equal the total reported in the monthly package.
This is the operating model we use at Level. The percentages are estimated shares of the work in a typical review. Actual effort varies with data quality, complexity, and the issue found.
| Cost line | Detail amount |
|---|---|
| Direct labor | $3,000 |
| Materials | $2,000 |
| Detail total | $5,000 |
| Reported control total | $5,100 |
Sum every numeric detail row and compare the result with the control total using a $1 tolerance.
Needs review
The supporting detail is $100 below the reported control total.
At Level, we record the review logic, known explanations, and questions our finance professionals use in this situation. AI applies that documented human judgment to organize the most plausible reasons, without pretending it knows which reason is true.
Possible reasons to investigate
The variance could be a missing row, a stale report, a classification difference, or a formula-range error.
A Level finance professional validates the source evidence, challenges the AI-assisted analysis, and decides which explanation is supported.
Trace the $100 variance before relying on the schedule for reporting or analysis.
Financial Systems Field Tests
These are repeatable Level Operating Rule tests using invented records only. They identify a control boundary and a refusal condition. They do not establish vendor behavior or replace review of your actual configuration.
FST-001 · LOR-002 · LOR-024 · LOR-033
Test: Compare same-cutoff record identities as well as the aggregate total.
Expected: A matching total without matching detail remains an exception.
Refuse when: Unknown cutoff, Missing record identities.
FST-002 · LOR-014 · LOR-024 · LOR-060
Test: Record source, cutoff, filters, row count, control total, schema, and file fingerprint before use.
Expected: A changed or incomplete export is visible before analysis.
Refuse when: No source definition, Missing fingerprint.
FST-011 · LOR-015 · LOR-061 · LOR-063
Test: Require page identity, expected columns, download completion, and resulting population before transformation.
Expected: A changed page or schema fails visibly.
Refuse when: Missing required column, Unexpected status value.
Use this system
MCP preview
run_synthetic_example {"slug":"cost-detail-footing"}This check organizes evidence and surfaces exceptions. A person with finance domain expertise still needs to interpret the result, validate the source records, and discuss the operating decision.
Open source does not mean open client data. We publish reusable finance tools, checks, templates, and playbooks. Client records, workpapers, credentials, communications, and private configurations remain confidential.
Use it yourself, or bring us the messy part
Tell us where the records stop making sense. We will reply with the first check to run, the records it needs, and the decision that still needs human finance judgment.
Install the free skills yourself