IRR > 200%
Vendor unbundling & insourcing
- 3 applications insourced
- 5-year savings > $5MM
In-house software that displaces expensive legacy and high-risk vendors.
AI
Advising boards, executives, and builders on AI — value creation, governance, training, and agentic systems — grounded in hands-on building.
I help organizations turn AI capability into governed, measurable enterprise value — working across the board's oversight questions, the executive's value and allocation decisions, and the practitioner's reality of shipping AI. I don't only advise on AI; I build with it, from adaptive learning tools to the AI Engineering Value Model below.
How I operate with AI
I don't only advise on AI — I run my own work on the same discipline I recommend. The story behind it →
Allocation over uplift
I don't count hours saved; I re-allocate freed capacity to the highest-leverage work — the discipline at the core of the value model.
Meter the budget
I run AI like a P&L line: concentrated on the top use cases and reviewed like any other spend, not sprayed across everything.
Execute and deliver
I've delivered $45M+ in AI-enabled net revenue — with zero regulatory findings — and applied the model across the enterprise use cases below. I still build myself, prototyping real products end to end (ReadinessIQ, a chess trainer for my kids, this site). Execution is what makes the advice real.
Multi-model by design
I work across the frontier — Claude, Devin, Codex, Gemini, and more — matching the model to the task and orchestrating them in agentic workflows. That keeps my advice vendor-neutral and current as the tools change month to month.
Human where it matters
Judgment, governance, and taste stay in the loop; agents do the heavy lifting.
Advisory areas
Featured model
AI Value Creation
A rigorous model for turning AI engineering uplift into enterprise value — capacity, constraints, and the allocation decision that separates a modeled gain from booked value.
Advisory
Board & AI Governance
AI oversight, responsible-AI frameworks, and the questions boards should be asking — informed by standards-body work at FDX and the Data & Trust Alliance.
Advisory
Executive & Board Training
AI literacy and working sessions for boards and leadership teams — building the fluency to govern AI and turn it into advantage.
Advisory
Agentic & Autonomous AI
Assurance, controls, and deployment as agentic software and reduced human-in-the-loop reshape how work gets done.
Featured model — AI Engineering Value Model
AI creates capacity. Leadership determines whether it becomes value.
Most AI-ROI math stops at an uplift number. The value is in what comes next — the capacity created, the constraints that gate it, and the allocation decision that turns freed hours into enterprise value.
The model in five slides · View the full PDF →
Applied use cases
Where the model has been applied — from a single pod to the enterprise. Representative outcomes. Scroll for more →
$2.5MM / yr
Developer capacity re-allocation
- Capacity re-allocated to growth
- New features & revenue
Redirecting AI-freed capacity to high-value initiatives, not backfill.
−$180M tail risk
Security risk reduction
- Remediation time −90%+
- −$10MM annual expected loss
- P99 event risk −$180M
AI detection, prevention, and hardening in high-value targets.
> $10MM saved
Legal contract remediation
- 100% of contracts risk-tiered
- 98% of changes automated
Automating large-scale contract matters (e.g., LIBOR → SOFR).
> $20MM saved
Risk & legal population testing
- Full-population testing
- Manual review < 0.02% of content
- No regulatory penalties
NLP and computer vision to defensibly test full populations.
> $5MM saved
Client problem resolution
- Faster resolution, fewer complaints
- Real-time incident feeds
- Programmatic compliance
Smart, compliant self-service that scales cheaper than staff.
Representative outcomes from applied use cases; results depend on scope, constraints, and the allocation of released capacity.
Applied AI Systems Lab
Working prototypes. Visible architecture. Honest limitations.
The evidence beneath the advice. Each prototype leads with the business question and the control point — then opens into a builder view with architecture, models, evaluation, human-in-the-loop controls, and what it can't do yet.
AI Engineering Value Diagnostic
Calibrate the model to your operating economics
I open a small number of executive calibration sessions for leaders who own AI investment, product growth, transformation, operating performance, or technology. The diagnostic calibrates all 15 scenarios to your organization, identifies the binding constraint, and develops a defensible, risk-adjusted value range.