In a team, AI can speed up preparation — and multiply competing versions of the same document. What makes the difference isn’t the tool, it’s clear roles: who prepares, who reviews, who decides. A team workspace makes those roles visible.
IN SHORT
- Name three roles: prepare, review, decide.
- Share a common context, but limit it to what each use case needs.
- Keep a record of what was done and by whom.
The risk: more versions, fewer decisions
When everyone asks their own assistant for a version of the same text, the team ends up with three proposals and no decision. The time saved on preparation is lost in arbitration.
The answer isn’t to ban it, but to agree on a process: one place for the context, one reference version, one person who decides.
Three roles to name
For each use case, put it in black and white:
- Prepare: who writes the request and chooses the sources;
- Review: who checks the content, the figures and the tone;
- Decide: who authorises sending, publishing or uploading.
One person can hold several roles. What matters is that the team knows who gave the final approval.
A shared context, but a clearly defined one
A common space saves you from sending the same documents with every request. That doesn’t mean it should contain everything: decide which documents serve the use case and which stay out of the preparation.
Shared templates — minutes, summaries, proposals — help everyone start the same way. That is the role of a shared playbook.
A record to find your way back
Once an action has taken place, the team must be able to find what was sent, to whom, and who authorised it. This traceability prevents duplicate sends and “I thought it had gone out” moments.
In Inkore
An Inkore team workspace brings together members, projects and documents with explicit roles: View only, Editor, Administrator and owner. External actions wait for human approval from an authorised person, and “What happened” keeps the status of each action. See Inkore for teams.
Frequently asked questions
How do you stop everyone using AI on their own?
Give the team a shared space with its reference sources, shared templates and one simple rule: the reference version lives in the team workspace, and final approval belongs to a named person. That way, preparation stays fast and the team never loses track of which version counts.
Who should approve actions prepared by AI?
Someone who has the right to authorise them and who has seen the exact content. In a team, this role should be named in advance for each use case, so that no one assumes someone else has already checked. Preparing an action and approving it can be done by different people.