AI approach

The assistant suggests. A person decides

Every change the assistant proposes waits for a human to approve it. Nothing is written into your company's record on the strength of a model's confidence.

Why this matters

An assistant that can write straight into your structure, goals or policies is not a productivity feature. It is an unreviewed editor.

Confidently wrong reads the same as confidently right

A model has no way to signal the difference between the two. A person reading the proposal usually does.

Records need an author

When a decision is questioned a year later, the AI added it is not an answer anybody can act on.

Context beats fluency

An assistant that has not read your DNA, structure and metrics will write something plausible and generic. The value is in the reading, not the writing.

How it works

  1. It reads your company, not the internet

    The assistant works from your own DNA, strategy, structure, goals and metrics, and its answers cite what is actually in your workspace.

  2. It proposes, in a reviewable form

    A suggestion arrives as a card you can read, edit and reject, rather than as a change that has already happened.

  3. A person approves

    Approved changes are written with their origin recorded, so anything the assistant contributed stays identifiable later.

  4. It refuses when it should

    Asked to set a goal in a cycle that does not exist, or to write a key result that contradicts itself, it says so instead of guessing.

What we hold to

No silent writes

There is no path where the assistant changes your workspace without someone approving it first.

Attributed changes

Anything approved from a suggestion carries that origin in the record, so the trail stays honest.

Scoped to the pillar you are in

The assistant in Strategy is not the assistant in Structure. Each is given only the context its job needs.

Always a manual path

Every action the assistant offers exists as a form you can fill in yourself. Nothing is reachable only through the AI.

Questions

Does the assistant make changes on its own?
No. Every proposal needs a person to approve it before anything is written to your workspace.
What does the assistant know about our company?
Only what is already in your workspace, and only the part relevant to the pillar you are working in. It is not given your whole account for every question.
Do we have to use it?
No. Every action it offers has a manual equivalent, so the product works fully without it.

See how the assistant works

Start with the plan that fits your size, or read how the pieces of Green Organization work together.