The specific cost of turnover today
Sarah joined 4 months ago as a CX lead. She's spent two hours a day with Claude building a customer-onboarding playbook — 80+ hours of refined work product locked inside her personal Claude conversation history. She gives notice. The next 4 weeks: she frantically copies notes into Confluence; Tom (her replacement) reads the notes and asks Claude to "rebuild what Sarah was working on"; the model produces 60% of Sarah's work and 40% reinvention. The org loses ~2 months of effective ramp-up on Tom because the AI institutional knowledge didn't transfer.
GenZAgents' agent transfer mechanism
Every employee's AI activity flows through their agent DID. When Sarah hands her notice in, IT initiates a transfer in /agents/[did]/manage → Transfer. The transfer is a cryptographically signed handover: Sarah's old keypair countersigns the transfer; Tom's new keypair accepts it; the agent's receipts, memory snapshots, project tags, and identity migrate atomically. On Tom's first day, he opens Claude Desktop, the MCP server picks up Sarah's (now Tom's) agent, and the org_context_lookup tool surfaces 4 months of relevant past work on his first prompt.
Day-1 ramp speed
Conservative measurement: a typical white-collar replacement role takes 4-8 weeks to reach 80% productivity. With agent transfer enabled, our design-partner data shows 1-2 weeks to 80% in roles where AI is central. The mechanism: Tom's first AI conversation already has the org's past work pre-loaded; he isn't starting from blank-context, he's starting from "here's what your predecessor figured out, build on it".
Why this matters at scale
A 100-person team turns over 15-25 people / year (industry average). At even 1 week of saved ramp per replacement × £100k loaded cost × 20 replacements = £40k/year in recovered productivity. That's before counting "knowledge that walked out the door and we never even noticed it was lost". The Enterprise tier pays back on the first 2 replacements; the rest of the year is upside.
Compliance angle for high-turnover orgs
For roles with regulatory continuity requirements (financial advisors, healthcare providers, legal counsel), the AI knowledge transfer also matters for compliance. The agent transfer is signed end-to-end so the chain of custody is auditable — "Sarah's AI work product was lawfully transferred to Tom on date X" is a receipt your regulator can verify.
What if Sarah leaves on bad terms?
The transfer is initiated by the org owner (typically IT or the People team), not by Sarah. Sarah's personal keypair countersign happens in a forced-exit scenario by org-admin override — see the org-admin recovery flow. Receipts can't be revoked or deleted by a leaving employee; the agent transfer is the only path that preserves continuity, and that path is org-controlled.