I use AI to build the finance apps,
dashboards and models leaders trust.
19 years across the US, APAC and Europe — now building with AI. I pair LLMs with Python to design and ship API-driven finance apps, live executive dashboards and self-updating models: from a production AI finance platform (Pecunio) to governed AI reporting inside a global FMCG enterprise. AI does the building and the interpretation; deterministic engines do the maths.
AI does the building. Deterministic engines do the maths.
I use LLMs and AI-driven Python to design, code and ship live finance software — then keep AI on interpretation, with the numbers coming from traceable, deterministic engines that tie out every time.
AI-Built Apps & Dashboards
API-driven finance apps and live executive dashboards, designed and coded with AI — Xero and enterprise data unified into real-time cash, P&L and forecasting views with role-based access and audit lineage.
LLM-Driven Intelligence
Multi-agent workflows and AI executive summaries — a Virtual CFO that reads the ledger, calls deterministic tools and narrates the drivers, grounded in the numbers rather than guessing them.
AI-Automated Models
AI driving Python to generate and refresh whole workbooks and board packs — 3-statement, rate-card and margin models rebuilt from raw exports each month-close, tied out to the cent with embedded checks.
Live, AI-built systems — not slideware
Real, API- and LLM-driven finance apps and dashboards, built with AI. Representative screens shown with illustrative data.
A CFO-trained AI finance platform, built with AI
Xero OAuth and accounting APIs unified into a live executive view of cash, P&L and forecasting. The interface, agents and engine were designed and coded with AI; every number traces to a deterministic engine, and an AI executive summary and multi-agent Virtual CFO sit on top.
| Metric | Jan 26 | Feb 26 | Mar 26 | Apr 26 | May 26 | Jun 26 |
|---|---|---|---|---|---|---|
| Revenue | $1.3M | $1.4M | $1.4M | $1.3M | $1.4M | $1.2M |
| COGS | $988K | $1.0M | $1.1M | $1.0M | $1.1M | $912K |
| Gross Profit | $302K | $334K | $338K | $318K | $346K | $278K |
| GP % | 23.4% | 24.6% | 24.0% | 23.9% | 24.0% | 23.4% |
| OpEx | $225K | $228K | $231K | $226K | $233K | $207K |
| Net Profit | $77K | $106K | $107K | $92K | $113K | $71K |
| NP % | 6.0% | 7.8% | 7.6% | 6.9% | 7.8% | 6.0% |
| Net Cash | $1.36M | $1.42M | $1.47M | $1.41M | $1.52M | $1.44M |
Live CEO dashboard — an AI-generated executive summary over a deterministic engine, with a KPI snapshot, monthly scorecard and 12-month trend, all API-driven from Xero.
| Line ($000) | YTD Actual | Q4 Forecast | FY26 Total |
|---|---|---|---|
| Revenue | 8,420 | 2,880 | 11,300 |
| Cost of Sales | 6,458 | 2,190 | 8,648 |
| Gross Profit | 1,962 | 690 | 2,652 |
| Operating Exp | 1,350 | 452 | 1,802 |
| EBITDA | 612 | 238 | 850 |
| Depreciation | 168 | 56 | 224 |
| Net Profit | 444 | 182 | 626 |
An AI monthly-review app, inside a global FMCG enterprise
A governed, AI-built performance-review application running inside a highly controlled enterprise (M365, TM1/Mosaic, Azure) — reframing AI from a security concern into an approvable capability. Deterministic Python engine, SOX-aligned by design, with AI on the narrative.
| Business Unit | Volume (kt) | BW AOP | Net Revenue | BW AOP | NOPBT | BW AOP |
|---|---|---|---|---|---|---|
| AU Foods | 5.8 | (0.1) | 65.7 | (0.4) | 21.3 | (0.4) |
| NZ Foods | 1.5 | — | 17.4 | — | 5.4 | (0.1) |
| Obela | 0.7 | — | 8.9 | +0.3 | 2.5 | — |
| XC | 0.2 | — | 1.3 | — | 0.5 | — |
| Flex | 0.3 | — | 2.2 | — | 0.7 | — |
| ANZ Foods (Total) | 8.5 | +0.1 | 96.8 | +1.5 | 31.0 | (0.2) |
- ANZ Foods delivered Net Revenue of $96.8M, +$1.5M (+1.6%) ahead of AOP and +$3.6M (+3.9%) above prior year.
- Volume of 8.5 kt ran 0.1 kt above plan on share gains in Impulse/Wholesale and stronger Woolworths NZ sell-through.
- NOPBT of $31.0M was $0.2M (-0.7%) behind plan, driven by a step-up in A&M ahead of the winter promotional window.
FMCG monthly-review app — a governed, AI-assisted enterprise reporting app. Illustrative data; production screens to follow.
Multi-tab board packs that read themselves
Xero data, uploaded knowledge documents and LLM commentary combined into a rich, self-contained HTML report — the same governed-engine pattern, generated with AI and delivered as a CEO-ready briefing.
- KPI headline, P&L / balance sheet / cash tabs, and written analysis in one file
- AI commentary grounded in the deterministic numbers, with data-gap flags surfaced honestly
- Exports to PDF and PowerPoint for the board pack, live-linked to the source model
Northwind Trading — FY26 Performance Review
Revenue by Quarter (A$000)
✦ AI Commentary
When the answer is a spreadsheet, AI builds that too
Not everything should be an app. Where clients live in Excel, I have AI drive Python to generate and refresh whole workbooks from raw accounting and payroll exports — actuals plus a driver-based forecast, tied out to the cent, with 40 embedded checks. Two examples, with illustrative data.
| Segment | FY26 Actual | FY27 Forecast | $ Change | % Chg | % of FY27 |
|---|---|---|---|---|---|
| Health/Aged Care | $6,574,000 | $7,100,000 | $526,000 | +8.0% | 46.4% |
| Education | $2,140,000 | $2,311,000 | $171,000 | +8.0% | 15.1% |
| Catering / Events | $1,286,000 | $1,389,000 | $103,000 | +8.0% | 9.1% |
| Sporting Venues | $942,000 | $1,018,000 | $76,000 | +8.0% | 6.6% |
| Hotels / Motels | $820,000 | $886,000 | $66,000 | +8.0% | 5.8% |
| Clubs | $744,000 | $803,000 | $59,000 | +8.0% | 5.2% |
| Government | $610,000 | $659,000 | $49,000 | +8.0% | 4.3% |
| Corporations | $512,000 | $553,000 | $41,000 | +8.0% | 3.6% |
| Other Revenue | $552,000 | $601,000 | $49,000 | +8.9% | 3.9% |
| Total Revenue | $14,180,000 | $15,320,000 | $1,140,000 | +8.0% | 100.0% |
Budget model — driver-based FY27 forecast off FY26 actuals, with a live model-status gate and segment drill-down.
| FY Week | Revenue | Profit | Profit margin % | Rev 4-wk avg | Margin 4-wk avg |
|---|---|---|---|---|---|
| 2026W40 | $212,814 | $55,430 | 26.0% | $248,555 | 26.4% |
| 2026W41 | $211,586 | $56,541 | 26.7% | $232,538 | 26.4% |
| 2026W42 | $220,022 | $59,577 | 27.1% | $222,691 | 26.6% |
| 2026W43 | $256,158 | $67,393 | 26.3% | $225,145 | 26.5% |
| 2026W44 | $281,242 | $75,047 | 26.7% | $242,252 | 26.7% |
| 2026W45 | $298,068 | $79,720 | 26.7% | $263,872 | 26.7% |
| 2026W46 | $281,185 | $75,108 | 26.7% | $279,163 | 26.6% |
| 2026W47 | $263,921 | $70,053 | 26.5% | $281,104 | 26.7% |
| 2026W48 | $244,748 | $64,332 | 26.3% | $271,981 | 26.6% |
| 2026W49 | $208,475 | $55,113 | 26.4% | $249,582 | 26.5% |
| 2026W50 | $224,679 | $59,375 | 26.4% | $235,456 | 26.4% |
| 2026W51 | $210,293 | $55,990 | 26.6% | $222,049 | 26.4% |
| 2026W52 | $198,502 | $52,965 | 26.7% | $210,487 | 26.5% |
Bottom 10 clients by net margin % (min $10k revenue)
| Client | Net margin % | Revenue |
|---|---|---|
| Marigold Events Pty | 8.1% | $11,053 |
| Compass Facilities Group | 21.3% | $330,320 |
| Metro Turf Club | 22.1% | $162,141 |
| Vanilla Blue Events | 23.3% | $81,733 |
| Riverside Club NSW | 24.0% | $69,401 |
| Hunter Country Club | 24.3% | $13,369 |
Top 10 profit-drag sites (lowest absolute profit)
| Site | Profit | Net margin % |
|---|---|---|
| Corporate Adjustments | ($727) | -41.0% |
| CC Training | ($711) | -28.7% |
| Trinity Grammar Kitchen | ($141) | -18.0% |
| Central Allocations | ($140) | -18.1% |
| LSL Provision | – | |
| Payroll Clearing | – |
MAPPING GAPS
Margin model — shift-level revenue, wages and on-costs rolled to weekly net-margin trends, client/site league tables and mapping-gap controls.
19 years, four continents, one throughline
Turning transformation goals into P&L results — under covenant pressure, post-merger, and inside the most governed enterprises.
PepsiCo
Embedded in ANZ Foods FP&A to modernize performance reporting and build AI-driven financial tooling inside a highly governed enterprise (M365, TM1/Mosaic, Azure). Reframed AI adoption from a security concern into a controlled, approvable capability — securing CFO, IT and enterprise-architecture sponsorship.
Pecunio
Built and shipped a proprietary, CFO-trained AI finance platform integrating Xero and accounting APIs into a unified executive view — automating up to 80% of manual client reporting. Full stack: Django/DRF, React/TypeScript, ingestion pipelines for messy PDF/DOCX/XLSX/CSV, and export to PDF/Excel/PPTX.
Equifax, Inc.
Partnered with GMs of a $150M+ B2B portfolio on pricing, product profitability and go-to-market. Rebuilt APAC forecasting to 99%+ accuracy and recovered $2.5M of lost revenue through billing-process reviews and controls.
Arconic, Inc. (formerly Alcoa)
Strategic finance lead for a $6B group. Key member of the PMI team for a $3B acquisition, owning synergy tracking and integration execution. Built a group-wide project-ROI model recognized as an enterprise-wide “Best Practice” at the global controllers’ conference.
AI in Finance
Governed LLM tooling, deterministic calculation and forecasting engines, multi-agent workflow design, and AI adoption inside security- and audit-constrained enterprises.
Finance Transformation
FP&A modernization, ERP & BI implementation, multi-entity planning and allocation redesign, shared-services design, cost optimization, and financial governance.
Executive & Board Engagement
CFO and C-suite partnering, business cases through IT, security and enterprise-architecture approval gates, and board-level decision support under pressure.
Technical Stack
Python, Django/DRF, React/TypeScript, LangGraph, PostgreSQL, Xero & Stripe APIs, TM1/Mosaic, Hyperion, Power BI — plus GitHub Copilot Enterprise & Microsoft Foundry.