Operating Studio OS — your day as the operator¶

We've followed Sofia from a cold email to a delivered, paid project. Now let's zoom out from the story to you — the person running the company on Studio OS. This chapter is about what your day actually looks like, how you stay in control, and how you bring the rest of your team into the work.

The short version: you spend your time deciding, not doing. Studio OS is built so that a company can run itself in the background and still be completely in your hands.

Your day in three surfaces¶

Almost everything you do as an operator happens across three simple views.

1. The morning brief¶

You start your day with Uku's board report — a short, plain-language read on the state of the company. Not a wall of numbers, but a judgement: what's healthy, what changed since yesterday, what to worry about, and the two or three things worth your attention today. It's the difference between "let me go dig through everything and figure out how we're doing" and "here's how we're doing, and here's what needs you."

Uku has already looked at the pipeline, the cash, the active projects, and the risks — and compared them to yesterday — so you don't have to.

2. The approval inbox — where you actually work¶

This is the one screen you'll use every day, and it's the beating heart of Studio OS: a queue of decisions waiting for you. Each is a card, and each card is the same shape you've seen throughout Sofia's story:

┌──────────────────────────────────────────────────────────────┐
│  [ Who proposed it ]   [ What it wants to do ]     ● priority │
│  Why: the plain-language reason this is being proposed         │
│  Preview: exactly what will happen / be sent (editable)        │
│           [ Edit ]      [ Reject ]        [ Approve ⏎ ]        │
└──────────────────────────────────────────────────────────────┘

Whether it's Kairi's reply to a prospect, Madis's kickoff email, or Liisa's invoice, it arrives here in the same form: the reasoning, the preview, and three choices. You can move through the queue from the keyboard, approve with a keystroke, edit before you send, or reject with a reason the agent will learn from. Low-risk items can be approved in a batch.

The design goal is glance, understand, decide. On a normal day, clearing your queue takes minutes, not hours — because the work is already done; you're just making the calls.

3. The live feed — the heartbeat¶

Running alongside your screens is a live activity feed: a quiet, continuous stream of what the company is doing right now — an email read, a lead qualified, a project moved, a hand-off passed. You don't have to watch it, but it's there when you want to feel the system working or trace how something came to be. Click any line and you jump to the thing it's about.

Everything on that feed is permanent and searchable. If you ever need to answer "why did we do that?", the answer — with the reasoning and the responsible agent — is on the timeline.

Staying in control¶

Autonomy without control is just a runaway. Studio OS is built the other way around: control first, autonomy where you've earned trust in it.

Nothing consequential happens without you. Anything that touches the outside world — an email, an invoice, a commitment, a change to your files — is a proposal until you approve it. The agents do all the thinking and drafting right up to that line, and then they wait.
You can let trusted work run on its own. Once you've watched a certain kind of low-risk action enough times and you trust it, you can let it flow automatically — for example, having approved emails send the moment you approve them, rather than in a separate step. You choose what graduates to automatic and what always waits for a human. Trust is dialled up deliberately, never assumed.
There's a single pause switch. If you ever want the company to stop acting on its own — during a sensitive week, a migration, or just to catch your breath — one control pauses all the background work at once. The agents go quiet until you turn them back on. It's the seatbelt you hope never to need and are glad is there.
Rejections make the team better. Every time you reject or edit a draft, that feedback is remembered, so the next one is closer to what you wanted. The system tunes itself to you.

The result is a company that runs continuously in the background but never gets ahead of you. You are always the one who decides how much rope to give — and you can take it back instantly.

Bringing your team into Studio OS¶

You are probably not the only human in the company. Studio OS is built for a team of people as well as a team of agents — and it routes work to whoever should actually do it.

People have their own logins and roles¶

Each person on your team can have their own Studio OS account. Roles decide what they can see and do: an operator like you approves the consequential actions and sees the whole company; a teammate can be given tasks and works from their own inbox without needing the keys to everything. You decide who's who.

Work goes to the right person — not always to you¶

This is the part that keeps Studio OS from dumping everything on the owner. When a task needs a specific human, it goes to that human, not to your queue:

The agents can message your teammates. Just as Madis drafts emails to customers, he can draft a message to a colleague — "Jaan, we need Marren's accounting connected by Friday; here's what you need." It's a draft that gets approved and sent, so your people are looped in through the same reliable path as everyone else. Yes — Madis (and the other agents) can absolutely reach your employees by email or message.
The agents can assign tasks to your teammates inside Studio OS. If the person has a login, the work can land directly in their Studio OS inbox — with the context and the reason attached. They do it, mark it done, and the agent that was waiting picks the thread back up. The task never touches your queue unless it needs your decision.
Notifications can be addressed to a person or a role. "This needs a finance approver" can go to whoever holds that role; "Jaan, your task is ready" goes to Jaan. Studio OS knows who your people are — it can even read your staffing system, through Piret, to stay aware of who's around and who's away.

So a fair way to picture it: the AI team and your human team share one workspace. The agents do the continuous work and route each task to the right owner — a colleague for the hands-on jobs, a specialist agent for research or finance, and you for the decisions that are genuinely yours.

                 Studio OS
        ┌───────────────────────────┐
        │  AI team        Human team│
        │  Uku            You (op)   │
        │  Kairi          Jaan       │
        │  Madis          …          │
        │  Liisa                     │
        │  …          one shared     │
        │             workspace,     │
        │             one timeline   │
        └───────────────────────────┘
   Work is routed to whoever should do it —
   and only your decisions come to you.

In one page¶

Your day is a morning brief, an approval inbox, and a live feed. You decide; the company does.
Nothing consequential happens without your approval — and you can let trusted, low-risk work run automatically, or pause everything at once.
Your teammates live in Studio OS too, with their own logins and roles. Agents can email them, message them, and assign them tasks — so work reaches the right person, and only real decisions reach you.
Everything is on a readable timeline, so the company is never a black box.

That's Studio OS: an autonomous company you can actually see, trust, and steer — proven, in this handbook, by watching one customer go from a cold email all the way to a finished, paid project.

Want to see it on your own business? That's the natural next step — the same journey you just read, with your inbox, your customers, and your team.