Case study · US Land Software Company
AI-assisted financial reports
Business-critical reports; near-zero documentation; limited access.
Problem
Business-critical reports; near-zero documentation; limited access.
Approach
Built two business-critical financial reports - owners to be paid, and ledger - for a client on SQL Server, working with very limited access and near-zero documentation. Used Python AI/ML techniques to infer structure and relationships the documentation didn't provide, then developed the reports to a standard finance could sign off on. Proof that the modern toolkit pays off even on legacy stacks: the constraint wasn't the technology, it was the missing knowledge, and ML helped recover it.
Architecture
Limited access, full accountability
Consultants rarely get admin rights and a data dictionary. This engagement had neither. The method: profile what's reachable, use ML-assisted analysis to hypothesize relationships, validate against known-good outputs, iterate with the client. The reports shipped; the approach is reusable anywhere the documentation died with an ex-employee.