Excel Intelligence & ControlsA LaurenceYang product

Current test evidence

One synthetic workbook, with clear limits on the claim.

We ran the Browser Analyzer and the saved full workflow on the same four-sheet synthetic accounting workbook. The browser read 58 formulas and flagged a likely overwrite; the saved run caught all four planted issues and produced files for review.

The result shows that the product works on this test case. It is not evidence of adoption by a finance team.

See the saved test run

Recommended first view · No setup required

From workbook to review files

The saved run records each step of the test.

It covers the test path from reading the workbook to saving reports and a dashboard.

  1. 1

    Read

    It opens the workbook

    The browser version read four sheets and 58 formulas on the device. The file was not uploaded.

  2. 2

    Check

    It finds what needs review

    The browser check flagged a likely manual formula overwrite. The saved full test caught all four problems planted in the sample.

  3. 3

    Record

    It keeps the result

    The run saved reports and a dashboard that show what was found and where the finding came from.

Stable local Agent authority

The local product baseline is now 1.34.0.

The website remains a static browser-local product. Separately, the verified local Agent package now contains five Skills for Build, Query, Standardize, Enterprise Management, and Ontology Git collaboration.

Versioned product receipt

  • 1.34.0 stable package · SQLite Schema 34.
  • 5 Skills and 76 shared contracts.
  • A 23-command synthetic product smoke passed across Build, Standardize, governed Query, reversible What-If, handover, and RDF/OWL export.

Open the machine-readable receipt

Runtime boundary

  • The Browser Analyzer on this site runs without a backend or model key.
  • The five Agent Skills run through a separate local installation; they are not simulated by disabled website controls.
  • The receipt proves bounded technical execution on synthetic fixtures—not production adoption, expert approval, savings, or business impact.

Claim boundary

What the test establishes, and what remains unmeasured.

The files document a working test. They do not measure team use, time saved, or business impact.

Established by this test

  • The Analyzer in this website works without a server or model API key.
  • The saved full test caught the problems deliberately placed in the sample.
  • The reports keep findings connected to the test workbook.

Outside this test

  • No accounting or finance team used this test.
  • It did not measure live company use, time saved, or business impact.
  • The product does not approve accounting rules; human review remains required.

Source and implementation detail

Open the saved artifacts and product documentation.

The links below include the generated dashboard, architecture, product guide, method deck, build history, and brand guide.

Technical documentsOpen the saved dashboard, system explanation, product guide, method deck, build history, and brand guide.
Saved report dashboardThe final review page generated by the test run. How the full system worksThe parts of the product and what each one does. Full product guideEvery workflow step, rule, command, and known limit. Method presentationA 24-slide explanation of the product method. Build historyHow the product changed over time. Brand guideLogo, color, type, and writing rules.