Also atCBL Analytics, where most of my current work lives: AI-assisted analytics, currently automating DCF financial models and construction-risk QRA/SRA.
Brisbane, Australia · Corporate finance · Analytics · Risk
Independent analytics for governments, infrastructure sponsors and banks. The models are increasingly generated from code, and no serious analysis happens without AI now, which is a good thing as it leaves more time for the design and actual analysis. And the value is in this analysis on top and knowing which questions to ask, not in whatever tools might be used.
Expressways · Ports · Prisons · Ferries · Rail bids · Basel II · IFRS 9 · APRA ICAAP · Treasury FRR
Exhibit 1 — Cumulative NPV percentile paths over 12 years, 10,000 trials, default assumptions. The live model is Exhibit 2, below.
Business cases, delivery-options analyses, tolling and patronage models, bid models, hurdle-rate reviews. Investment-grade DCF models with assumption books, documented so a board, lender or auditor can open a cell and see how a number was built. Mergers, acquisitions and restructures.
See the workData and risk analytics in R and Python: classification, regression, time series, Monte Carlo simulation. Delivered through Quarto, Shiny and Power BI over SQL. The modelling is now heavily AI-assisted, under a gated process: specify and verify before any code is written, test before anything ships.
See the workPPP and project finance models. Simulation-based risk analysis, QRA and SRA cost-risk analytics. Financial-institution risk: regulatory capital and credit systems under the Basel framework, IFRS 9 provisioning, ICAAP stress testing, liquidity and capital.
See the workSelected engagements, 1990s to now.†
| Field | Engagement | Output |
|---|---|---|
| Infrastructure | Brisbane expressway, bridge and water projects Financial models for preliminary evaluations and business cases. | Model |
| Government | Government financial statements (FRR) model Financial Statements delivered through MS Word from an Excel and VBA core. | Model |
| Banking | IFRS 9 provisioning system Built as an Azure and R web application for an overseas credit agency. | System |
| Transport | Ferry operations, Brisbane and overseas Operating and financial analyses for stakeholders. | Analysis |
| Banking | APRA CPS 226 swaps margining Initial-margin compliance for non-centrally cleared derivatives. | System |
| Infrastructure | Correctional facility evaluation Delivery options and financial evaluation. | Evaluation |
| Corporate | Finance Director, medical-device startup Information Memoranda; finance function built out, Xero to NetSuite. | FD role |
| Transport | Rail bid models and sourcing strategies Procurement-side models and sourcing evaluation. | Model |
| Banking | Basel II credit risk system APRA-compliant PD, LGD and expected-loss framework. | System |
| Energy | Sugar mill cogeneration project evaluations | Model |
| Infrastructure | Commercial port facilities options evaluations | Analysis |
| Banking | APRA ICAAP stress testing Capital stress testing for an Australian mutual. | Analysis |
† a selection of a much longer list of engagements.
Live model — Monte Carlo NPV · 10,000 trials · Client-side, no data leaves the page
Exhibit 2 — NPV distribution under correlated cost and revenue uncertainty. Method notes below.
This is a small version of how projects get tested before anyone signs anything: put distributions on what you don't know, run the model ten thousand times, and read the spread of outcomes it produces. Method — revenue and cost draw correlated normal shocks each year (Cholesky, ρ = 0.45; cost volatility 18%); both grow at the revenue growth rate over 12 years; cash flows discount at the chosen rate. Shock multipliers floor at zero, so revenue and cost cannot go negative. Seeded PRNG (Mulberry32, seed 42): the default run reproduces exactly, every time.
The AI side of the practice: automating DCF financial models, construction-risk QRA/SRA analytics, and SEC filing analysis, source-verified throughout.
cblanalytics.com
I'm Steve Parton, a chartered accountant and MBA with thirty years in finance, banking and analytics. I've built the models and business cases behind big infrastructure decisions: expressways, bridges and water projects, ports, prisons, ferries in Brisbane and Dubai Creek, rail bids, a sugar mill cogeneration PPP. Later the work moved into banking risk and public-sector reporting: Basel II credit systems, IFRS 9 provisioning, ICAAP stress testing, Queensland Treasury financial statements.
The toolkit grew with the work. Excel and VBA first, then R, Python, SQL and Power BI, now Quarto and Shiny, and for the last year or so a lot of AI (Claude mostly, in my case). The focus is still the client's objectives; AI has changed how I get there, all to the good. What hasn't changed is the need for transparency, a clear audit trail and results that reproduce.
These days the modelling team includes several Claudes. The judgement, and the review standards, are still mine.
If the decision matters, the analysis should hold up.
Correspondence
stephen.parton@sjpconsulting.comBrisbane, Australia
SJP Consulting · Brisbane, Australia · ABN 97 543 127 136
Set in Newsreader and IBM Plex Mono. Static exhibits generated by the same seeded engine that runs live. No framework, no analytics, readable without JavaScript. © 2026 SJP Consulting.