QuickBooks Cleanup | CR Bar Livestock

A good cleanup isn’t just about making the numbers look right—it’s about figuring out why they were wrong in the first place.

I completed a comprehensive QuickBooks Online cleanup for CR Bar Livestock, a small agricultural business with livestock sales, breeding stock purchases, ranch expenses, owner transactions, and historical bookkeeping that needed to be reviewed and organized.

The project included:

• Reconciling cash and bank activity against supporting records • Reviewing and correcting historical transaction classifications • Cleaning up Undeposited Funds and resolving unmatched transactions • Reviewing sales receipts and deposits to ensure income was recorded correctly • Correcting owner transactions and equity classifications • Reviewing the Chart of Accounts for accuracy and organization • Reviewing the Profit & Loss and Balance Sheet for unusual or incorrect balances • Completing a Books Review of the cleanup period • Adding documentation and notes to create a clearer audit trail

One of the biggest parts of cleanup work is tracing transactions back to what actually happened in the business rather than simply changing numbers until they match. By the end of this project, the books were cleaner, reconciled, easier to understand, and much more useful for making business decisions.

This project is also a great example of why I enjoy working with agriculture and livestock businesses—I understand that the bookkeeping has to reflect how the operation actually works.

Agentic Financial Document Intelligence & Transaction Reconciliation Engine

• Role: Lead AI & Systems Architect • Deliverables: Multi-Modal PDF Parser, Hybrid Agent Orchestrator, Deterministic Math Engine, Audit Trail Logger • Timeline: 3 Weeks

OVERVIEW Financial institutions and back-office teams spend thousands of manual hours reconciling unstructured documents (loan contracts, invoices, compliance bylaws) with relational transaction ledgers. Standard LLM wrappers fail catastrophically in this domain due to mathematical hallucinations, version conflicts, and a lack of explainable audit trails.

I engineered an agentic document intelligence system that combines multi-modal PDF extraction with deterministic validation tools to automate complex cross-referencing with 100% computational precision.

THE PROBLEM & DIAGNOSIS LLM Math Hallucinations: Generative models cannot be trusted to perform exact financial calculations, interest compounding, or tax audits.

Unstructured vs. Tabular Asymmetry: Document facts are scattered across messy layouts, whereas transaction logs live in strict relational schemas.

Document Version Discrepancies: Clauses change across revisions; systems must resolve conflicting policies before executing reconciliation.

KEY ARCHITECTURAL DECISIONS • Decision 01: Separation of Reasoning and Execution The LLM acts strictly as a reasoning and extraction layer, identifying entities and formulas. All financial calculations and reconciliations are handed off to deterministic Python execution sandboxes. • Decision 02: Cross-Source Entity Disambiguation Built a cross-referencing engine that maps unstructured text snippets directly to SQL transaction IDs using hybrid vector-keyword scoring. • Decision 03: Fully Auditable Chain-of-Thought Every validation output includes a step-by-step citation trail pointing to the exact page, bounding box, and ledger entry.

MEASURABLE RESULTS & IMPACT • 0% computational hallucination rate achieved via deterministic code execution. • 90%+ precision in entity extraction across multi-page, complex financial agreements. • Production-grade audit trails satisfying compliance requirements.

TECH STACK Python 3.11, FastAPI, Pydantic, LangChain, PyMuPDF, Pandas, PostgreSQL / pgvector, Docker.