Coming soonOperations

Hire an AI Documentation Keeper

Documentation has a universal lifecycle: written in a burst of virtue, consulted for a month, then quietly wrong forever. This employee breaks the cycle. It watches how your systems and processes actually change and keeps the documentation current as they do - updating what drifted, flagging what it cannot verify, and writing what is missing. The docs stop being an archaeology site and start being the answer.

How hiring works

What you get

How it goes, start to finish

  1. Say go

    Your private channel opens the minute you hire. No call, no kickoff meeting, no software to install.

  2. Homework first

    It studies your business and writes your Company Brief, then audits your existing documentation: what exists, what is current, what is quietly wrong, what is missing entirely.

  3. Approve the priorities

    The audit lands as a ranked plan: fix these lies first, fill these gaps next. You approve the order and the docs it may edit versus propose.

  4. Maintenance becomes standing

    It watches for change and updates the affected pages, with edits visible and reversible. Uncertain updates arrive as proposals for a human yes.

  5. The docs stay alive

    New processes get captured as they appear, stale pages get retired deliberately, and the answer to "how do we do X" is current the day anyone asks.

Where the quality comes from

Like every Tenfold employee, this one runs one opinionated method: verify against the system before trusting the page, mark what cannot be verified rather than presenting it as true, and write for the reader at 2am - short, concrete, findable. When the field moves, the method moves, and you do nothing to get the update. Here is the whole method.

Access

It reads your systems and maintains docs through grants you make once, in your portal's Access tab. You choose which docs it may edit directly versus propose changes to, every edit is visible and reversible, and you can revoke access yourself at any time.

Fair questions

How does it know something changed?

It watches the observable record - the systems, the configurations, the activity - and compares it against what the docs claim. Divergence is the trigger; the update or the flag follows.

Can it write docs for things only one person knows?

Yes - it drafts from the observable evidence, then asks that person a few short questions to close the gaps. Twenty minutes of their time turns tribal knowledge into a page that survives their vacation.

What if it updates something incorrectly?

Edits are visible and reversible, uncertain changes arrive as proposals rather than edits, and anything it could not verify is marked as unverified on the page itself. The failure mode is a flagged question, not a confident lie.

Which docs does it manage?

The ones you point it at, with per-doc permissions: edit freely, propose only, or read for context. Most teams start it on operational runbooks - where staleness hurts most - and widen from there.

Will it delete old docs?

It proposes retirement when a page is dead - wrong, duplicated, or describing a process that no longer exists - and archives on your approval. Nothing vanishes silently.

More of the team

Ready when you say go.

Requesting this role moves it up the roster - requests set the order roles open for hire, and you will hear the moment it opens.

See all 51 roles