Training
Pillar oneAI training for teams that have to use it on Monday
Most AI training teaches a tool. A month later the tool has changed and nobody has changed how they work. We train on your own work, with your own files, so what people learn survives contact with the job.
What the sessions actually cover
Every programme is built around work your team already does. We ask for three or four real tasks — the weekly report nobody wants to write, the inbox that never empties, the deck that gets rebuilt every quarter — and the session is spent doing those tasks with AI, in the room, on your files.
That covers the tooling, obviously: how to prompt properly, how to structure a request so the answer is usable, how to chain steps together, and which of the current crop of tools is worth your licence budget. But the part that sticks is the judgement — knowing when the output is good enough to send, when it needs checking, and when the whole approach is wrong.
Where not to trust it
We spend real time on failure. Teams that have only seen AI work well are the ones that get burned by it, because nobody taught them what a confident wrong answer looks like.
So we cover hallucination and how to catch it, what happens to anything you paste into a consumer tool, why AI is bad at arithmetic and worse at citations, and where your PDPA obligations bite. People leave able to say “not for this one” and explain why, which is worth more than another prompt template.
Who it is for
Whole teams rather than individuals. The value shows up when a department shares the same working habits, not when one enthusiast gets faster and nobody else changes. Operations, marketing, finance, HR and production teams have all run this; no technical background is assumed.
For leadership we run a shorter session on where AI is worth spending money and where it is not, which is a different conversation and usually a shorter one.
How a programme runs
Four stepsScope
A short call to find the three or four tasks worth building the session around.
Session
Half a day or a full day, hands on, your files, your tools, in your office or ours.
Playbook
You keep a written playbook of the prompts and workflows that worked, so it survives staff turnover.
Follow-up
A session a few weeks later on what stuck, what did not, and what to automate properly.
Formats and what they cost
A single half-day session for one team is the usual starting point. From there it runs to a full day, a multi-week programme across several departments, or a train-the-trainer arrangement where we build the material and your own people deliver it.
Cost follows the number of people, the number of sessions, and how much of the material has to be built specifically for you. We quote after the scoping call, never before. Singapore SMEs should also read our note on EDG and PSG grants, which can apply to work of this kind.
Why us
Two of the people who run these sessions lecture part-time — Davier at Nanyang Polytechnic, Joshua at NTU — so teaching is not something we picked up to sell training. And the same team builds and runs the AI systems day to day, so the material comes from doing the work rather than from reading about it.
Start with one team
A half-day session on one real workflow tells you more than any proposal. Tell us what your team does all day.