SHIP & SCOPE
RELIABLE AGENTIC AI
WITH ROI
Two one-day intensives in Singapore, each followed by two weeks on your own real work. Build agentic AI that holds up in production - or learn to scope it so organisations keep using it.
AI Consulting
Scope agentic work that organisations keep using and can measure - the bridge between what the business needs and what the product does.
WHAT YOU LEARN
- Interview a practitioner and find where the judgment really sits
- Decompose a workflow: what an agent owns, what stays human
- Design and validate an agentic workflow end to end
- Size it with numbers that survive being questioned
WHAT IT GETS YOU
- Agentic work that moves numbers the business cares about
- Workflows teams still use months later, not pilots that die
- Less patching - it keeps running after you hand it over
- You become the interface between business and product
$1,200 · 20 seats only · scholarships available
Prerequisites: you own or shape AI work - product owners, consultants, transformation leads. Not an introduction to AI.
Agentic Building
Build agentic AI that holds up in production and pays for itself - instead of workflows you spend every week patching.
WHAT YOU LEARN
- Read an unfamiliar, messy codebase and locate the change
- Build an agentic workflow on a real system, not a tutorial
- Put logic in code, the model only where judgment is needed
- Gate anything the agent cannot undo
WHAT IT GETS YOU
- Workflows that hold up without constant babysitting
- Stop rebuilding - the fix is architecture, not more prompting
- Ship something that survives production, not a demo
- Be the engineer who can actually deliver this
$1,200 · 20 seats only · scholarships available
Prerequisites: you should already know some AI or software. Engineers and career-changers welcome.
Why it exists
In a controlled study, doctors were given AI advice that was confident and wrong. Their accuracy got worse. The most experienced clinicians fell the hardest - because years had taught them that a well-articulated claim in their field is usually correct. They had everything they needed to catch the error. They just did not engage it.
That is the whole problem. Not that AI is too clever. That we stop asking.
Why it is harder than it looks
Building an agent that actually works is a multi-disciplinary job. You need design thinking, to work out what the process should even be. Psychology and UX, because you are modelling how someone really works, not how they describe it. Software engineering, because it has to hold up when nobody is watching. And enough understanding of how the model works to know where it will drift.
Most people are strong in one or two. That gap is why so many agent projects quietly die - not because the demo failed, but because the thing needed patching every week until someone shelved it.
The aim is helping people get there, if they can. I will give you everything I have on the day and for the two weeks after. Whether it lands is partly on you, and I would rather say that now than sell you a certificate. Either way you will leave knowing something real about how these tools work and where they will burn you.
, our training partner.Apply
Places are limited and I read every one of these myself. If it is a fit I will send you the brief with the two setup steps.