AI Transformation Lab
A hands-on leadership cohort where you put an AI agent to work on your own real task — no coding, no theory. By the end of the afternoon you'll have the judgment to direct an agent well, not just watched one.
You've sat through the decks, nodded at the demos, and told yourself you'll properly get into it once things calm down. But every week a peer posts what their team is now doing with AI agents, and the quiet question gets louder: am I already behind? Here's what no deck tells you: this is a people-and-operating-model shift before it's a technical one. It's yours to lead, not IT's. And you can't lead from the passenger seat.
Leaders accountable for their company's AI transformation who haven't gone hands-on. Managers, Directors, and people leaders under pressure to drive AI who haven't had the time to get genuinely hands-on. A curated, application-only cross-company cohort of 8 to 12 leaders, in person in Singapore.
Four hours in a curated, cross-company room of leaders. You leave with three things no tutorial can give you:
You drive an agent on your own tasks and lead from direct experience. Not theory, not a demo you watched.
Most people expect AI to just perform. The leaders who get value treat an agent like a new hire. You'll onboard it with the right context, so you can tell good output from bad.
You leave with a sharper sense of where agents could pay off in your own work and your team's workflows — a starting point, not a vague "AI matters."
Open on a live agent. It reads your calendar, catches conflicts across months, books the meeting — because it has your context. Then you build your own.
Context is everything. AI knows nothing about your world until you hand it the context — get that right and it's transformative. Then the leap that matters: an agent doesn't just answer, it does the work.
The hands-on heart of the day — in the Claude app. First you put an AI agent to work on your own mail and calendar. Then you build working software — with no code — just by describing it. Give it the right context, watch it act — you approve.
What you built runs on your laptop, so it stops when you close the lid. On company infrastructure the same agent runs always-on, scheduled, and shared with your team. We cover where your data goes, plainly, along the way.
Leave with a private opportunity worksheet — your ranked shortlist of where agents pay off in your org.
Plus a round-the-room: the one task you'd each most like to hand off.
The one mental model, then a tour of the three Claude products — Chat, Cowork, Code.
Two builds back to back — an agent on your own mail and calendar, then working software with no code.
Where a laptop agent hits its ceiling, and how the always-on version runs.
Fill in your private worksheet, recap, and agree the one next step.
We set every leader up with a month of Claude Pro. The momentum doesn't stop when the room empties: you keep working on real tasks, not a trial that locks you out a week later.
You walk out with first-hand command of an AI agent — the judgment to direct it well and the confidence to lead your org's transformation, not delegate it.
By the end of the afternoon, you'll have put an AI agent to work on a real task of your own — or the day's on us.
Ready to put it into production? Meet Glimmer — no pressure.

I'm YJ. I co-founded Glints and spent the last decade building it — from the first lines of code to scaling the platform to 15M+ users, first as Chief Software Architect, then Head of Product.
Like every team, we've had to do more with less. AI turned out to be the unlock — it let our people punch far above their weight and took the grind out of the work, instead of letting it burn them out.
Here's the honest part: the potential was obvious to me early, but I watched too many sharp people try AI once, have a clunky first experience, and quietly write it off. That's the real waste — set it up right and it genuinely makes you superhuman. This cohort is me making sure your first real experience is the good one, and passing on what I wish everyone had. It should lift all of us, no one left behind.
Applying is the low-commitment first step, not a checkout — you tell us a little about your role and what you're after, we review it for fit, and confirm your seat.
A short form for a cohort seat: five minutes, no prep. It's an application, not a payment.
We read every application to keep the room small and well-matched — usually within a few hours.
We confirm your spot and send a payment link. Prefer to talk it through first? We'll set up a quick call, but it's not required.
This is a leadership-cohort engagement, not a commodity seminar — priced for a curated room of directors.
Bring a peer and it's 1-for-1 — two seats for the price of one, S$389 a seat. The best rooms have two leaders from the same company comparing notes.
Already know you want a deployed agent instead? talk to us directly.
No. It's built for leaders, not engineers. You direct the agent in plain language — the point is the judgment to command an agent well, not writing code.
The open cohort is built for leaders and decision-makers setting their company's AI direction, so the room stays peer-level and you leave with an org-level read on where agents pay off. If you're keen to get hands-on as an individual contributor, the best route is a private session for your team — point your leader our way.
One afternoon — a half-day cohort, 13:30–17:30. You leave the same day having put an agent to work on a real task of your own.
In person in Singapore — we share the exact venue once your seat is confirmed. We run a cohort every Friday afternoon; apply and we'll confirm your date.
You work on your own task in your own Claude environment — nothing you bring into the room is shared. Bring a real task you're comfortable working on in a small peer group.
The half-day cohort, hands-on time putting an agent to work on your own task, a one-month Claude Pro subscription to keep going, and a private opportunity worksheet to take back to your team. Seats from S$779.
Yes — and it's the best-value way in. Two leaders who sign up together get the 1-for-1: two seats for the price of one, S$389 a seat. Apply together, or mention your peer in your application, and we'll apply it at confirmation. Both seats include the one-month Claude Pro subscription.
No — it's deliberately not a SkillsFuture-accredited course. Accreditation pushes a programme toward standardised, commodity content; we've kept this a small, hands-on leadership cohort instead. It's privately paid, and typically expensed as professional development.
No. The cohort is for leaders who want to get hands-on directing agents first. If you already know you want one built, talk to us directly and we'll scope it with you.
Yes. If you'd rather bring your own leadership team or staff into a closed session, we'll tailor the day to your business and run it just for you. Talk to us to scope it.
This is a new programme and we're keeping the early cohorts deliberately small to get it right — so we read every application and hold each room to a tight group. Applying commits you to nothing; if this cohort isn't the right time, we'll point you to the next one.