From my perspective, leaver risk is the AI risk most UK SME practices haven’t thought through. The usual offboarding process closes email, revokes system access, retrieves hardware. Specially for AI tools, this isn’t enough. The leaver may have used AI for months on client work — their prompt history contains client data — and closing their account doesn’t retract the data they’ve already seen.
What the kill switch does
Every user in a SCRS-protected platform has their own per-user encryption keys. All the AI content they’ve interacted with is encrypted with those keys. When a leaver is offboarded, a single click revokes the keys — not just access. Their historical AI prompts, retrieved client data, generated drafts — all become un-decryptable the moment the keys are gone.
Un-decryptable means unreadable. Even to us, the platform vendor. Even to the leaver themselves if they’d somehow saved cached data. The mathematical operation to read the content requires the keys, and the keys are gone.
Apart from this, the audit chain remains intact — it records that the leaver interacted, on which matters, when. But the content of what they retrieved is no longer reconstructable. That’s exactly the property you want: audit trail preserved for compliance, but data exposure closed.
Why this matters for UK SMEs
Specially for regulated practices, leaver risk is an ongoing compliance exposure. Under UK GDPR, if a leaver retains client data, the practice is still accountable — the data subject’s rights extend to data held anywhere, not just in your active systems. A kill switch closes this exposure structurally.
I believe this is the single feature most under-rated in AI tool selection. Practices evaluate AI on workflow speed, accuracy, integrations — not on what happens when a staff member leaves. Then the first leaver triggers the question, and only then is the gap visible. Specially for small practices where staff turnover happens regularly, structural leaver handling matters.
Learn more
More on Features and Security. SCRS is self-serve — get started whenever you’re ready.