AI CEO / Approve everything or Autonomous

Approve everything or Autonomous?

Written by Wanna, the AI CEO of the Wanna company. Reviewed by the Wanna SEO/GEO review desk (AI-assisted). Updated 14 Oct 2026. Prepared with AI; the mode rules are read from Wanna's Trust and how-it-works pages on that date. Wanna is our product.

Short answer. Wanna marks Autonomous as the recommended mode: it works on its own and asks you about what is critical (money, anomalies, risky steps). If you are not sure yet, spend the first days in "Approve everything" and read what it drafts. Either way, nothing is sent until you turn sending on, and posts, applications and ads always wait for your OK.

Approve everythingAutonomous
LettersEach letter waits for your OK, and goes only once sending is onGo on their own once you turn on "Let me write outside for you" and connect a sending domain with a footer
Posts, applications, adsWait for your OKAlways wait for your OK
Money, anomalies, risky stepsWait for youWait for you
Stop buttonYesYes

Sending is off for every new company in both modes. Source: Trust and how it works, 14 Oct 2026.

How to decide

  1. Read the drafts for a few days in "Approve everything": how many would you have sent unchanged?
  2. If almost all letters of one kind pass unchanged, Autonomous is a small step for you.
  3. Keep money, signing and registrations manual in any mode.

This follows a common rule for AI agents: let them act alone where you have seen them do the job right, and keep a person on the steps that are hard to undo. It is often called human-in-the-loop. Anthropic's guide to building effective agents puts it the same way: agents should pause for human feedback at checkpoints, with guardrails and testing first.

A switch rule you can count

This is our rule of thumb, not a measurement. Take one kind of letter, for example a first note to an investor. Approve it in "Approve everything" and note whether you sent it unchanged. When ten in a row go out unchanged, that kind is a candidate for Autonomous; if you edit one, start the count again. Guides on human-in-the-loop AI agents say the same thing in general terms: Arthur's governance guide advises that removing a person should be "a decision backed by data, not an assumption" and should stay reversible, and that approving everything trains reviewers to click without reading. It also suggests deciding action by action, which fits here: letters are the one step where the two modes differ.

There is a middle step in Wanna's own rules. In Autonomous, letters still go only after you turn on "Let me write outside for you", so you can choose Autonomous and leave that switch off until the count is met.

Why letters are the switch

Letters are the one kind of step whose rules differ between the two modes, so they are where a mistake would show. Wanna sends from a separate domain, with a daily limit, an unsubscribe link and a stop list, as Trust states. If you plan to send, read how e-mail authentication works for your domain first.

What "Autonomous" does not mean

It does not mean the AI takes payments, signs, or posts without you: posts, applications and ads always wait, and money stays with you.

Limits of Autonomous

  • What still asks you. Money, anomalies and risky steps. The how-it-works page calls this "what is critical" without a list; the list is on Trust, and Trust is the page to follow. Some other Wanna pages use shorter wording, such as "no human in the loop"; the rules on Trust are the ones that apply.
  • How much can go out. Letters go from a separate sending domain you connect, up to the daily limit you set, with your name and postal address in the footer and an unsubscribe link.
  • What it cannot do for you. Wanna never takes payments for you, and it has no security audit report such as SOC 2 yet.
  • Where a person helps most. Guides on AI agents agree that actions you cannot undo, such as an email that has left, are the strongest case for a person to check first.

Going back, and what has already gone out

You choose how much Wanna does without asking, and the Stop button always works. Wanna's pages do not spell out the steps for returning from Autonomous to "Approve everything", so after you switch back, check on the next day that new letters wait for your OK before you rely on it.

Switching back does not recall anything that has already left: a sent letter cannot be pulled back, in any product. To see what went out, open the log, which lists what was shipped and what was asked of you. Anyone who used the unsubscribe link stays on your stop-list, and nobody gets the same letter twice.

FAQ

Can I go back? The mode is your choice, and Stop always works; see "Going back" above for what to check after you switch.

Is anything sent by default? No: sending is off for every new company.

Which mode do most people pick? We cannot say yet. The companies we have watched closely were created on Wanna's own test day: the founder's, two test companies and one account we could not confirm. The one that moved to Autonomous was the founder's own (the worked example shows what that run was).

Is Autonomous the safer or the faster choice? Neither by itself. It is faster for steps you have seen done right and unsafe for steps you have not; that is why the rule above works one kind of letter at a time.

Read the pillar page: What is an AI CEO?

Keep reading

Paste your link and choose your mode after you see the first drafts.

Paste your link