Problem after problem, solved.

SAP screens didn't fit how the plant actually worked. Cycle counting, receiving, quality inspection, fulfillment, and service each needed interfaces built around the workflow, not the transaction code. And decades of tribal knowledge lived in senior technicians' heads and scattered documents.

A unified operations portal against the live SAP HANA schema — Customer 360, cycle count, receiving, quality, fulfillment, service and executive command centers, with barcode capture where the workflow demanded it. On top: a retrieval-augmented AI knowledge core with role-based filtering, so techs query documentation and service history in plain English and get source-cited answers.

A divestiture put a manufacturing division's Fourth Shift data on a hard deadline to land cleanly in the parent company's Oracle environment: AP invoices, supplier masters, GL extracts — with a corporate integration template of dozens of required fields, and finance teams on both sides who needed the trial balances to tie.

Built and iterated the extract suite against the corporate template, tightening invoice-selection logic to eliminate duplicate-payment exposure. When period batches disagreed between systems, I didn't pick a side — I reconciled to root cause and published the explanation, with query-level proof. Delivered AR visibility screens so finance could see the state of the book throughout.

France's e-invoicing mandate meant a Tier-1 aerospace supplier had to produce and exchange structured Factur-X/CII invoices through a certified platform — a hard government deadline, a multi-party integration spanning the ERP, a compliance platform, and corporate stakeholders on two continents.

Owned the integration architecture and field-level mapping from the ERP through to the certified e-invoicing platform: invoice lifecycle states, error handling, and the unglamorous work of making every mandated field land correctly. Ran the workstream across French operations, corporate, and platform vendors — the coordination is half the job.

Quality management ran on an aging Access database nobody fully understood, alongside a rough shop-floor system. The client wanted modernization but had been burned by big-bang rewrite pitches — and when portal numbers disagreed with the legacy system, "trust the new tool" wasn't an acceptable answer.

Insisted on a Phase 0 gate: fully reverse-engineer the Access database before writing migration code. Then built module by module — customer scorecards with normalized quality/delivery/service/cost scoring, returns management, audit forms, a defect-to-resolution workflow. When scrap numbers disagreed between systems, I reconciled both to root cause and documented exactly why they differed.

High-volume wheel production with barcode-driven material movements — interlayer handling, paint-line conversion, bin logic — where the standard ERP transactions didn't match how material actually flows, and a mis-scan means wrong inventory in a plant that runs around the clock.

Designed and wrote the data-collection scripts that sit between the scanner and the ERP: substitution logic, receiving conversions, guarded prompts that make the right action the easy action. Every change proven in a sandbox environment against production-shaped data before a controlled deployment — because on a live shop floor, "it should work" isn't a deployment plan.

Decades ago I wrote this manufacturer's ERP in FoxPro. It has run their business ever since — a testament to fit, and a growing liability. They needed a path to a modern web platform without betting the company on a rip-and-replace, from a partner who understood every quirk of the system. Fortunately, one was available.

A piece-by-piece conversion to a modern web stack — each module replaced only when its successor is proven, so the business never stops running on software it trusts. AI-accelerated development makes an incremental path economical that would have been unaffordable at consulting rates a few years ago.

A European-headquartered manufacturer's North American operation ran a mature, stable Fourth Shift environment — with reporting that stopped at what the vendor shipped in the nineties. Operations and finance were flying on exports and spreadsheets.

A phased reporting build: procurement intelligence first — vendor scorecards, PO reliability, lead-time drift, open-PO risk — because supplier surprises were the acute pain. Built directly on their real data, structured so each phase funds the case for the next.