Three lines of work. All project-based.
Build it, hand it off, it runs without anyone. No retainers, no ongoing dependency.
AI decision systems for trading firms
The problem
Every trading firm has rules. Risk limits, position sizing guidelines, pre-trade checklists, post-trade reviews. Most of them exist in a PDF, a spreadsheet, or someone's head.
Under pressure — when the market gaps against an open position, when a winning streak creates overconfidence, when a loss triggers the urge to recover — rules break. Not because traders are undisciplined, but because rules are static and human psychology is not.
What gets built
A custom AI decision system that integrates into the team's existing workflow and enforces process in real time. Not a chatbot. Not a dashboard. An operational layer that sits between the trader and the market.
How it works
Who this is for
- —Proprietary trading firms with discretionary traders
- —Crypto trading desks and digital asset hedge funds
- —Quantitative firms with a discretionary overlay
- —Family offices with active trading allocations
- —Any firm where the gap between the rules and what actually happens is costing capital
This is not education. Traders don't need to be told what confirmation bias is — they need a system that catches it in real time, in their specific context, before it costs them money. The system is built from the inside out: starting with the firm's existing methodology and reinforcing it.
Decision intelligence for executives and founders
The problem
High-stakes decisions in business share the same structural challenges as trading: they are made under uncertainty, influenced by cognitive biases, rarely evaluated on process quality, and the feedback loops are slow and noisy.
Investment committees approve deals based on groupthink. Founders refuse to kill projects because of sunk cost. Executive teams pivot based on recency bias. Boards evaluate performance by outcomes, not by the process that produced them.
What gets built
Practical decision frameworks and AI systems that integrate into how the organisation already makes decisions — not theoretical models that sit in a slide deck.
Format options
Diagnose existing decision patterns. Introduce the framework. Apply it to a real decision. Leave with practical tools the team can use the next day.
The full framework operationalised as an AI system. Tracks patterns, detects biases, enforces cooling periods, generates review reports. Build and handoff.
Who this is for
- —Venture capital and private equity investment committees
- —C-suite teams making strategic decisions
- —Startup founders navigating pivots, fundraising, and hiring
- —Family offices evaluating deal flow
- —Any leadership team where the cost of a bad decision is measured in millions
This is not generic leadership coaching. The frameworks come from a decade of application in real-time, high-stakes trading — where every decision has an immediate, measurable financial consequence. What works under those conditions works everywhere else.
AI skill development for experts and firms
The problem
Every organisation has expertise that lives in the heads of its most experienced people. When the expert retires, takes a new job, or simply isn't available — the knowledge goes with them.
The traditional solutions — documentation, training, mentorship — help but don't solve the fundamental problem. Documentation goes unread. Training is forgotten within weeks. Mentorship doesn't scale.
What gets built
An AI system that captures domain expertise and applies it operationally — not as a reference document, but as an interactive decision-support layer that guides, checks, and reinforces the expert's methodology in real time.
Who this is for
- —Trading firms — scale the best trader's methodology across the team
- —Professional services (law, audit, consulting) — capture senior partner expertise
- —Educators and coaches — turn methodology into interactive AI systems
- —Any organisation with complex operational processes that live in manuals people consult after something goes wrong
Scope options
One process, one domain. A due diligence checklist, a pre-trade protocol, or an onboarding flow.
3–5 interconnected modules with cross-referencing, routing logic, and shared context.
5–10+ modules with hierarchical structure, automated tracking, review systems, and evolution mechanisms.
Most knowledge management solutions are databases. They store information and hope people search for it at the right time. This is an operational system that applies expertise in real time. The difference is between a library and a copilot.
Related thinking
Every engagement is project-based.
Build, deliver, hand off. No retainers, no ongoing dependency. The system stays and runs independently. If there's a situation that could benefit from a structured approach — start with a conversation.
Start a conversation →