Where AI actually pays off
I spent years building strategy and analytics for enterprises — McKinsey, private equity, insurance, automotive, travel. Then AI changed the economics: what used to need a big team and a big budget now needs a clear use case, an API, and someone who can actually build it.
But most of it fails — MIT puts GenAI project failure around 95%*, almost always the wrong use case or no honest ROI. My job is the opposite: cut through the hype, find where AI genuinely pays off in your business, prove the return, then build the working software myself — not another deck.
Want proof? The Portfolio has live, clickable tools I've built — an FP&A revenue forecaster, a document-extraction engine, a private-equity market-scan — plus my own AI products with real users. Don't take my word for it; click one.
What I do
One person across the whole arc — from process analysis to production. See how I work →
-
01
AI & Data Strategy
Where AI genuinely pays off, sequenced into a plan you can act on. The ex-McKinsey lens — business case, priorities, risk — before a line of code gets written.
- AI audit & maturity read
- Prioritised opportunity roadmap
- Exec-level AI advisory
-
02
AI Opportunity Audit
A fast maturity read and prioritised opportunity map — fixed scope, honest verdict. The free 30-minute entry point.
- Data & workflow review
- Prioritised opportunity map
- Clear next steps
-
03
Data & FP&A Analytics
Forecasting, driver and revenue analysis, financial planning that holds up — from data engineering to the model.
- Revenue & driver forecasting
- Planning & scenario models
- Data engineering to model
-
04
AI Build & Automation
I don't just advise. LLM workflows, document and data extraction, custom tools — designed, built and deployed.
- LLM workflows & agents
- Document & data extraction
- Custom tools, in production
You don't have to
- Wait months for a strategy deck while nothing ships
- Stitch together five specialists who blame each other
- Pay for advice, then go find someone to build it
- Bet on a use case nobody proved would pay
Instead, you could
- Get an honest read in a free 30-min audit, a working prototype in days
- Work with one person across the whole arc — process to production
- Have whoever picks the use case also build and ship it
- Only build what the ROI case actually justifies
Tools I've shipped
Not slides — working software you can click. A few here; the rest are in the Portfolio.
- Live demo
FP&A Revenue Forecast
Revenue forecasting and scenario planning (AutoARIMA).
- Live demo
PE Market Scan
Mini commercial due-diligence: public comps, valuation, voice-of-customer.
- Live demo
Doc → Insight
Invoice extraction with anomaly rules and OCR.
- Live demo
Customer Churn Radar
Churn prediction with SHAP explanations.
* MIT NANDA — The GenAI Divide: State of AI in Business (2025). ~95% of enterprise GenAI pilots showed no measurable P&L impact. Source ↗