CPA with over 15 years across finance, SAP, and enterprise systems. Multiple full-cycle S/4HANA implementations in aviation, mining, pharma, and public sector. Now building multi-agent AI systems that solve real finance problems — same ownership mindset, new layer.
Started as a CPA — financial analyst, external auditor, a decade of practicing finance. That background shapes everything I do in ERP and AI.
Spent over 15 years in finance and transformation: multiple full-cycle S/4HANA implementations across regulated industries. Went deep on Controlling, Project System, financial close — and took ownership of the data migration layer (LTMC, Migration Cockpit, LSMW, LTMOM). A senior functional consultant should own everything in their domain.
Now applying that same mindset to agentic AI. Building autonomous agents, wiring them to ERP systems, owning the intelligence layer the way I owned the data layer.
Agents pass structural summaries, not raw data. Most expensive run across all five assets: $0.46 for a four-report financial statement bundle.
Confidence-based posting thresholds. Human-in-the-loop escalation. Structured audit trails. The agent augments judgment, doesn't replace it.
The intelligence layer belongs to finance, not the ERP vendor. Agent patterns transfer directly across platforms.
Five agentic AI systems. Each solves a real enterprise finance problem — designed, built, tested, and validated independently at near-zero cost.
Profitability analysis required recurring aggregation of high-volume GL data — grouped by multiple keys, reconciled to the penny, with zero tolerance for error.
SAP master data creation from service tickets was fully manual — read, extract, validate, build the upload, update the status. Every record. Every time.
AP invoice processing — matching line items to GL accounts and producing balanced journal entries — relied on manual accounting judgment that didn't scale.
Multiple pipelines for different SAP operations meant users had to know which tool to use. The goal: one entry point that understands intent and routes intelligently.
The full accounting cycle — journal entries, AP and AR invoices, Business Partner creation, financial reporting — required multiple manual steps across the ERP. The goal was a single orchestrated agent handling the complete lifecycle.
The AI work is credible because of what's underneath it.