Concept

What is the AI Accountability Ledger?

The AI Accountability Ledger is the append-only record of every AI action across an organization: what ran, in which department, under whose ownership — and whether it should have. Actions without an owner are surfaced; actions that shouldn't have run can be reversed. It is how growth answers for itself.

In practice

• Every action a worker takes — a refund issued, a payment scheduled, access granted — becomes a ledger row with an owner state.
• Rows with no owner are flagged, not buried: accountability gaps are made visible before they become incidents.
• Each executed action is sealed as a signed, HMAC-chained receipt you can export, replay, and verify.
• Risk, security, and IT approve rollouts because the record exists by construction, not by promise.

Ledger vs. log

A log is written for engineers to debug. The Accountability Ledger is written for an organization to govern: it speaks in business actions and owners, it is append-only and tamper-evident, and its unit is the receipt — a verifiable record of one action — rather than a line of text.

Common questions

Can AI agent actions be audited?

Yes — on the 95% platform every agent action is recorded in an append-only, tamper-evident log that can be exported and verified. Each execution is sealed as a signed, HMAC-chained receipt that can be replayed.

What does a security review need to approve AI agents?

Per-tool permissions and role-based access, human approval gates on high-risk actions, per-team budgets and model allow-lists, and an exportable, tamper-evident audit record — the controls the ledger makes inspectable.