Client instructions & approval workflows
Take an instruction, route it, approve it and record it in one sealed thread. Private AI reads the request and drafts the workflow — in an isolated environment operated by Letro, never a third-party model.
One instruction, five recorded moves
The client writes once, in the channel
A structured instruction, end-to-end encrypted, with the files attached to the matter — not to an inbox.
Private AI turns it into a workflow
It summarises the ask, names the approvers your policy requires and flags what is missing. The model runs in an isolated environment operated by Letro — not a third-party service.
- —3 asks extracted from 6 pages
- —Routed to Corporate · partner approval required
- —Missing: signed board resolution
The right people, in the right order
Roles decide who can view, approve or escalate. If an approver goes quiet, the escalation path moves it on.
Confirmed where the decision was made
The client’s counter-approval lands in the same thread, timestamped on delivery and receipt. No printing, no scanning, no third mailbox.
Every move, provable ten years later
Who asked, who approved, what they saw, when they confirmed — one tamper-evident record per matter.
Built for sovereignty, not retrofitted
Built on Matrix, an open standard for decentralised communication. Message content stays end-to-end encrypted while the organisation keeps its controls around the workflow.
End-to-end encrypted content
Instructions, files and confirmations are readable only by the parties on the matter.
Role and policy controls
Who can view, send, approve, escalate or administer — modelled per entity, team and counterpart.
Your deployment model
Cloud, Dedicated or Sovereign — including air-gapped. With a Sovereign deployment, no one else holds your data.
Tamper-evident audit trail
Every interaction is timestamped, traceable and exportable for a regulator or a dispute.
Bring us one workflow that keeps slipping
45 minutes. We map it, build it in your Letro deployment and hand you the audit trail.