Not a report. Not a pilot. A capability-building programme that leaves behind working AI skills your team owns — measured in capacity, cycle time, and dollars.
Your team isn't underperforming. It's maxed. Every pitch, every brief, every reconciliation runs on individual effort — and when the work grows, the only lever you've ever had is another hire.
So here's the arithmetic almost nobody runs. At roughly S$6,000 per head per month fully loaded, unlocking just 15–20% of productive capacity across a 50-person team is the equivalent of 7–10 additional full-time hires — S$45–60k of capacity every month, without a single new hire.
That is what this programme is designed to unlock — not a tool rollout, not a pilot, but a measured capacity gain your P&L can actually see.
It's also how the programme prices itself. The investment is sized off the value — a modelled 20× return after EDG co-funding — not off a day rate. Move the levers; the cost lever follows.
Built for services teams. If most of your headcount is field or production labour, we model only the desk-bound layer — the percentage shrinks, the logic doesn't.
Each workstream delivers a targeted set of AI skills — highest-value use cases first, compounding across functions as the library grows.
Finance goes first — a contained, fast-win environment that proves the methodology before we touch revenue-generating teams. Workstream 01 is our four-phase finance roadmap at programme tempo — and if you've already walked the roadmap with us, your proving ground is built. The programme starts ahead of schedule.
Your functions aren't these five? The Week-1 discovery workshop maps the programme to how your business actually runs.
Traditional consulting delivers a deck that gathers dust. This programme delivers working AI skills embedded in daily operations — owned by you, improving with use.
Every skill ships with a written methodology, prompt library, and usage guide — not tribal knowledge in one person's head.
Validated against your real workflows before sign-off. Nothing enters the library on a demo alone.
Designed for team-wide adoption and new-hire training — capability for the organisation, not dependency on an individual.
We chain skills across functions on purpose — each new capability multiplies the ones before it.
Matches supplier invoices against the bank feed and drafts the month-end reconciliation workpaper. Use this skill whenever new supplier invoices arrive, the bank feed updates, or the user asks to reconcile payables, chase unmatched items, or prepare the close pack. Flags duplicates, part-payments and FX differences for review — it never posts an entry without a named human's sign-off.
Matches supplier invoices to bank transactions and drafts the month-end AP reconciliation workpaper for human review.
2027-03 AP Reconciliation ✓ 142 matched automatically ⚠ 6 exceptions for review (3 part-payments, 2 FX, 1 duplicate) → workpaper: 2027-03-AP-recon_for-review.xlsx
4 hrs per week → 20 min review. Locked at discovery, measured monthly. If the number drifts, the skill gets fixed — not the story.
Specimen from the library — details illustrative.
Every one of the 100 skills ships like this: documented, owned, audited against its baseline.
Each new skill multiplies the ones before it. By Q4 the target isn't 2× faster — it's a different gear.
Tools don't transform a business — people who use them do. We run the change side of this programme as deliberately as the build side.
One mid-programme, one at completion. They accelerate adoption across teams — and surface the organic use cases no roadmap would have found.
The hackathons reveal the people with a gift for this. We recognise their wins in the open and build them into a core of in-house experts who carry the capability long after we leave.
Adoption is built into the rhythm of the business and how wins are recognised — celebrated, never policed — so the skills library becomes the path of least resistance, not an extra chore.
Capacity is not a euphemism for headcount. The programme is designed to redeploy freed hours into growth — more pitches fought properly, faster delivery, better client work — not to hand you a retrenchment list. Your people end the year more valuable than they started it: trained, certified on the skills library, and harder to poach than they were to keep.
Skills are built on commercial-grade AI accounts — primarily Claude (Anthropic) — whose terms exclude training on your data. No client work ever runs through personal or consumer accounts.
Wherever possible, skills run inside accounts you own and control. When the programme ends, nothing needs migrating — it was yours all along.
A mutual NDA is standard before any data leaves your building. Access is scoped per function, logged, and revoked at handover.
Personal data inside your processes is handled in line with the PDPA — minimised, purpose-bound, and never used to train anything.
The full data-handling annex is part of every engagement letter. Ask for it in the first conversation.
AI will run your reconciliations and your renders whether we're here or not. The only question that matters is whether you trust the output enough to decide on it.
Value does not arrive as a flat monthly figure from day one. It ramps as functions come online and compounds as skills interconnect — in the model, steady state arrives around months 7–12. We publish the model and measure against it monthly.
Every figure above follows from the levers you set in the math section — drag them, and this model moves with you. The model values capacity at cost; realised value also counts cycle-time, win-rate, and rework effects — modelled here at 1.1–1.6× the pure capacity figure, which is where the 22–32× comes from. Strip the grant out entirely and the same model still returns 11–16× — the EDG doubles the multiple; it doesn't create it. We'll build your version of this model in the first conversation — your functions, your headcount, your numbers — and you'll hold us to it.
The Enterprise Development Grant, administered by Enterprise Singapore, co-funds up to 50% of qualifying consultancy costs for projects that upgrade, innovate, or transform a business. A properly scoped AI transformation programme is exactly the kind of project the EDG is designed to support — subject to EnterpriseSG's assessment of scope, outcomes, and provider.
Most owners we meet have never applied. Not because they're ineligible — because the application is unfamiliar, the timeline is real, and one procedural misstep voids the whole grant. That's the part we handle with you: we prepare the application together, scope the project the way EnterpriseSG assesses it, and nothing starts until the letter of offer is in hand.
No engagement letters signed, no work performed, no payments made before the grant is approved — or the application is void. We've navigated this; the sequence is built into how we engage.
Fixed fee, scoped to your size. As an anchor: a full programme for a mid-sized services business runs around S$100k gross — S$50k net with EDG co-funding at the maximum 50% — invoiced by milestone, never by the hour. Smaller teams scale down; the capacity model above will give you a realistic range before we ever speak.
Nothing in this programme is designed around retrenchment. Freed capacity is redeployed into growth — and your team ends the programme trained and certified on a skills library they own.
See 'Your numbers stay yours' above: commercial accounts whose terms exclude training on your data, your own accounts wherever possible, NDA before data, PDPA-aligned handling.
Every programme is built around the published model and measured against it monthly — in a report you see. Phases are gated: we agree the checkpoints and the stop conditions in the engagement letter before anything begins. If the model reads red and we can't fix it, stopping is a clause, not a confrontation.
Then no work has started and no fees are owed for work — that's how the no-commencement rule works in your favour. We'll tell you honestly whether to re-scope and reapply, or whether the unsubsidised math still clears (it usually does — see the 10× line above).
You do. The library, the documentation, the accounts, the trained team — all yours at handover. We keep nothing you'd have to rent back.
A line item we model upfront, not a surprise: typically a small fraction of one salary for the whole company. It's in the value model from day one.
Designed for a business running at capacity: short build sessions per function, two or three champions with a few hours a week, leadership in one monthly review. The Week-1 discovery workshop maps the exact cadence to your delivery calendar before we commit.
Skills are documented at the process level, not bolted to one model version — when the underlying tools improve, your skills usually get faster, not broken. Handover includes the maintenance playbook.
Yes — that's the beachhead. The Strategic Finance roadmap builds the proving ground first: a finance function that can audit AI's returns. The full programme follows when the math has proven itself on your own books — and roadmap clients start it ahead of schedule, because their proving ground is already built.
Every skill tested in our own operations first. Our practice runs on the same AI-augmented operating model we build for clients — document production, market intelligence, financial modelling, all of it. We run Nitro on the same system we sell.
Deep financial acumen in every function we touch. Every AI skill is designed with ROI in mind, not just productivity — measured in dollars, not demos.
We are not building a dependency. The library, the methodology, and the upskilling are yours — the engagement ends, the capability doesn't.
Tell us about your business and we'll show you the model behind the multiple — your functions, your headcount, your numbers. We'll be in touch within 48 hours.