Workspace management and data ops

Every month, identify projects with no activity in 90+ days. DM the project owner in Slack: “This project has been inactive since [date]. Archive it in Dovetail if you’re done.” The agent never archives or deletes — the human confirms and acts in Dovetail directly.

Every Monday, compile active projects, stale data, tagging consistency, and missing metadata rate. Compare to trailing 4-week baseline. DM to the ops lead with any trends worth attention.

Every month, DM the workspace admin in Slack with a checklist reminder: “Review project access against your data governance policy this month. Check: (1) any newly created projects, (2) any projects with sensitive data, (3) any projects with external collaborators.” The agent does not read or change project permissions — this is a human-owned review.

Every Monday, scan insights created in the past 7 days. Compare against existing insights across all projects for semantic overlap. Flag potential duplicates for the research ops lead. Never auto-merge — duplicates are sometimes intentional.

Every Monday, compare contact records across Salesforce and Dovetail. Flag mismatches: missing records, name discrepancies, role changes. Notify the workspace admin. Never auto-reconcile — humans decide.

When an Account’s plan tier changes in Salesforce, update the matching tag in Dovetail. Log the change for audit.

When data is uploaded missing required metadata, identify the uploader. Create a Slack follow-up task: “Add [missing fields] to [data item].” Due in 3 business days. Be specific — generic prompts get ignored.

Every Monday, identify data uploaded outside any active project in the past 7 days. DM the uploader in Slack: “This data isn’t in a project. File it in [suggested project]?” Track resolution.