the ai os install · one day

Your team doesn't need another AI course.
You need an operating system.

One afternoon, hands-on. We turn how your team works into a system that runs itself — playbooks encoded, busywork automated, the first workflow live before we walk out.

See What Changes ↓
No slides. No theory. Your work, made to run itself.
why you're stuck

It was never a prompting problem.

Your sharpest people still do it by hand.

The same report, the same numbers, the same email — every day. Ten AI tabs open, none of them connected, nothing that compounds.

Everyone "uses AI." Nothing changed.

A prompt doc no one opens, output that depends on who's typing. It never learned your business — so no one trusts it.

And the gap is opening.

A team your size already runs this on autopilot. Every week without a system, the distance compounds.

It's not your team. It's the missing layer between them and the AI. We install it in one day.
the part you should be skeptical about

Does the team still use it after we leave?

friday 14:00install day starts — the team watches one workflow run end to end
friday 17:27they onboard a real client into the system. unprompted.
friday 18:29they build automation #12 — themselves, with the generator we left behind
monday 09:00the weekly client report — data, analysis, branded PDF, draft email — ships itself. nobody from our side in the building.

Latest install: a six-person Portuguese marketing agency. Three growth signals in 72 hours — all of them after we left.

That's the only test that matters — not demo day, the Monday after.

The same architecture runs in production across content engines, BD pipelines, finance operations and market intelligence. See the work

what gets installed

An operating system.
Not a toolkit.

Like the OS on a computer — you never look at it, everything runs on it.

OS_01·

Speaks your language

"Run the weekly report for client X." Done.

the router
OS_02;

Knows your business

Your clients, your voice, your way of deciding.

context + skills
OS_03=

Does the repeatable work

Scheduled in the cloud. Laptops closed.

agentic workflows
OS_04@

Grows itself

Builds new skills by interviewing your team.

the generators
the install day

One afternoon. Live by 18:00.

14:00 — see it runthe brain loads18:00 — first workflow live
14:00

See it run

One finished workflow, start to finish, on real data. Then we build yours.

first half

Load the brain

Your clients, voice and playbooks go in. It answers like your team, not the internet.

a context pack, per person
second half

Encode the judgmentthe hero block

Your best operator's decision tree becomes a skill anyone can run — calibrated live, in the room.

your hero skill + the generators to build more
18:00

Live

Wired to your stack, scheduled in the cloud, tested in front of everyone. Write actions ship paused.

one agentic workflow, in production
life after the install

Mondays look different now.

The report is already in the channel when your team logs in. Data pulled at 08:00, analysis written, client email drafted for review. The work that used to eat Monday morning is just — done. And it compounds: every correction teaches it, every week it knows your business a little better.

·
today

One workflow live — built in the room, on your real data, running on a schedule.

· = ~
30 days

Three. Your team built the new ones — that's what the generators are for.

=+*#%@
90 days

Ten. The predictable work runs itself. Your people are on clients, strategy, judgment — work that needs a person.

The workshop is the install, not the upsell — going deeper is a conversation for after it's run for a week.

who it's for

Teams of 3–8. Any industry.

Founders, operations leads and account teams who already feel the pull of AI but don't have a system. Delivered in English or Portuguese, in person by preference — your office, your data, your clients.

We gate properly: a discovery call and a pre-flight checklist five days out (seats, connectors, data access cleared). And a safety default worth saying out loud: write actions ship paused — nothing sends email or touches budgets without a human switching it on.

Marketing AgenciesExecutive HeadhuntingLegal & Professional ServicesFinance & InvestmentReal EstateHospitality & HotelsE-commerce & RetailWeb3 & CryptoYours — the OS is the same, the skills encode your playbooks
questions

Asked before.

Is this prompt training?

No. Prompting is maybe 10% of the day. The other 90% is the system: your context loaded, your playbooks encoded, plain-language operation, and agentic workflows wired to your real stack. Training changes what people know; an install changes what runs on Monday.

What do we need before the day?

A discovery call, AI seats for each participant, the desktop app installed, and connectors authorized — we walk you through all of it with a pre-flight checklist five days out. Total team effort: about 30 minutes.

Does it work outside marketing?

The architecture is industry-agnostic. What changes is the playbook we encode, and that comes from your team in discovery. If your operation has repeatable decisions, it has skills waiting to be written.

What about our data?

Connectors are scoped to what the workflows need, and write actions ship paused by default — nothing sends to clients or touches spend without a human enabling it. The first thing we configure is the seam between automation and judgment.

How many people can join?

Three to eight per session — it's hands-on, and everyone builds. Larger teams split into two sessions on the same day.

English or Portuguese?

Both, natively. The system itself is bilingual — your team talks to it in whichever language they work in.

How much does it cost?

Scoped on the discovery call — it depends on team size, your stack, and how much pre-work your use case needs. The call is 30 minutes and you'll leave it knowing exactly what one day can install.

ready to install

Install the engine.

A 30-minute discovery call. We map where the leverage hides in your operation — and you leave knowing exactly what one day can change.

The gap is compounding either way. No commitment, no pitch deck — just an honest look at it.

See Deployed Systems →
DONE — Forward Deployed AI That Executes, Not Advises