Closos
The accountability moat

Trust receipts: how Closos signs and proves every autonomous action

Most AI asks for trust. Closos earns it — one receipt at a time. Every action emits a signed Trust Receipt: the evidence used, the policy matched, the approver chain, and what it would take to roll back. Reversible. Auditable. Yours.
01 · The artifact

Open a live receipt.

This is the artifact Closos hands you after every action it takes — annotated. Click each layer to see what the action was, why it was allowed, what evidence it used, how it's signed, and how to take it back. Hand it to legal. Hand it to your board. Archive it for seven years.

Interactive · click to inspect

Walk a receipt, layer by layer.

This isn't a screenshot tour. It's the same artifact your team gets in production — annotated so you can see what accountability looks like when it's built in, not bolted on.

Trust receipt · #cl-08x14-7f
Signed · reversible
Action

Followed up with Theo Renaud · Acme Corp

Executed by
Closos+ · Autopilot
When
Mon · 8:14 AM
Policy match
outbound.email · ASK → APPROVED by you
Approver
h.maeda@yourco.com
Evidence used
  • ·Last contact: 11 days ago (stale)
  • ·Champion replied positive on Apr 14
  • ·Procurement queued security review (17d SLA)
Signature
Rollback
1 click · 0 cost
PDF · downloadableSigned · tamper-evident1-click rollback
02 · The rails

Two rules no admin can unlock.

Most autonomy comes with a kill switch. Ours comes with a contract. These two rails are compiled into the product. Your admin can't loosen them. Our support can't unlock them. They are how Closos can be trusted to run on a Friday at 4:47pm.

Rail 01
Compiled, immutable
RAIL_01 ::

Outbound to humans is always ASK.

Closos can never auto-send an email, message, or notification to a person outside your org without an explicit human approval. Not in any plan. Not on any policy. Not for any role. This is the line.

Rail 02
Compiled, immutable
RAIL_02 ::

Personal overrides can only tighten. Never loosen.

A user can choose to be stricter than their org's autonomy policy. They cannot choose to be looser. Org policy is a ceiling, not a starting point — the math goes one direction only.

03 · The matrix

The autonomy policy, shown not summarised.

ActionOFFASKAUTONotes
outbound.email — to human LockedRail 01
internal.note
crm.update
meeting.propose
asset.send (templated)
spend.approve (< $50)
spend.approve (≥ $50)
user.override (loosen org) LockedRail 02
04 · Live proof

The privacy machinery, running in public.

Retention purges happen on schedule. DSRs get fulfilled. Erasures leave nothing behind. These numbers are aggregated across the platform — no PII, no per-customer detail, no inferred identifiers. The receipts behind them live inside each org's authenticated dashboard.

Last retention purge
DSRs completed · 30d
Erasures · all-time
Orgs running on Closos

That's not a feature.
That's a contract.

Read the enterprise story

Operated by EthicalBrain Technologies · Dubai International Financial Centre (DIFC), Dubai, United Arab Emirates. Privacy and security questions: privacy@closos.com.