Revising a draft before you approve

You are not stuck with the first thing your AI drafts. Compliance writing is iterative: you read a draft, you want the scope narrower, the wording tighter, a different control behind a requirement. Kanonik is built for that back-and-forth, and it guarantees the one thing that matters for your record: you only ever approve a version that has been checked, and the check always matches exactly what lands.

The loop

When your AI drafts something, it runs the draft through the safety check and shows you both the draft and the result, before anything is recorded. You read it. If you want a change, you say so in plain language. Your AI revises and runs the new version through the safety check again. You see the updated draft and the updated result. You can go around this loop as many times as you need. Nothing is recorded, and no approval link is created, until you are satisfied.

So the safety check is not a one-time gate you race against. It runs again on every revision. The version you finally approve is a version that was checked in the exact state you approved it.

Why a revision is always re-checked

When you approve a change, the approval link is bound to the exact content in front of you. If the content changes, the link no longer matches it. That is deliberate. A revised draft cannot ride an old approval, and a check run on an earlier version cannot be reused for a different one. A revision produces a fresh check and a fresh link. You can never approve one version and have a different version land.

What "checked once" actually means

You may hear that Kanonik checks a change once rather than twice. That refers to a narrow, safe step, not to skipping anything you changed. Between the moment your approval link is created and the moment you click it, the content is locked: it cannot change. Re-running the identical check in that frozen window would cost time and money for no new information, so Kanonik reuses the result it already has. This is what lets a large batch seal quickly instead of timing out. The only thing it ever skips is repeating the same check on content that is provably the same.

One detail worth knowing

For a control-to-requirement mapping, the safety check is judging one thing: whether this control actually satisfies this requirement. If you only reword the note you attached to the mapping, that judgment has not changed, so the same result carries over. If you change the control or the requirement, it is a different mapping and gets a fresh check. For a policy or a control, the check is about the text itself, so any edit to the body is re-checked.

Why this matters for your audit

When an auditor asks how a change was vetted, the answer stays clean: every recorded change carries the result of a safety check run on that exact content, approved by a named person. There is no drift between what was checked, what was approved, and what landed. You revised freely; the record stayed precise.

More help