A shared AI workflow is useful when teammates can repeat the result without guessing which context, destination, or approval rule another person intended. Sharing everything is not the goal. Preserving the accepted parts of one working process is.
Strawberry can keep team guidance, shared artifacts, ownership, and recurring work together while personal or client-confidential context stays within its approved boundary. Start with one proven workflow and verify one real handoff.
Start with one proven workflow
Choose a process that already creates a useful result for one person. Clarify its inputs, owners, destination, field meanings, review points, exceptions, and what a successful handoff means before trying to standardize it.
Use an existing example or template when one is available. The aim is to preserve the team's judgment, not replace it with a generic process document.
Make ownership and handoffs visible
Map approved inputs to a reviewable output, keep source links where they matter, and make responsibility and approval boundaries explicit. Pilot the process with the people who own the work and verify permissions, field meanings, handoffs, and review behavior.
Ask before sending sensitive communication, changing access, or making irreversible updates. A shared process should make control clearer, not bury it.
Add recurrence only when the team trusts the process
Preserve the accepted process as a team-visible custom skill after the first handoff works. Add a Routine only when the team explicitly wants recurring execution and has agreed on ownership, review, exceptions, and notification behavior.
When the workflow is part of joining or leaving the team, connect it to a reviewed onboarding or offboarding checklist so access and responsibilities do not drift.