# AI Audit Trail & Decision Records | iKB

> A Decision Record keeps the sources, policy checks, model use and results for each answer. Export JSON or PDF evidence with access and retention controls.

Source: https://kbmy.ai/architecture/decision-records

[iKB](https://kbmy.ai/) / [Architecture](https://kbmy.ai/architecture) / Decision Records

# AI Audit Trail & Decision Records What is recorded behind an AI answer?

A Decision Record keeps the sources, policy checks, model use and results for each answer. Export JSON or PDF evidence with access and retention controls.

[Request a demo](https://kbmy.ai/contact?interest=Decision%20Records)

[_Screenshot: iKB Decision Records with sample answer records across native chat, WebChat, Teams, Slack, API and the omnichannel inbox (OMNI)._](https://kbmy.ai/assets/neo2/decisions.webp) Decision Records iKB / Sample data

How it works

## Three steps to get started.

- 01 ### Sources and context Check the topic, cited documents, transcript references, data records and calculations used for the answer.
- 02 ### Controls and decisions Review prompt versions, privacy checks, connector approvals, model routing and any fallback used.
- 03 ### Evidence exports Export supported JSON or PDF records with integrity checks and access controls. Retention and signed links depend on your configuration.

Features & controls

## What decision records includes.

### Per-answer Decision Records

A Decision Record brings together the sources, policy checks and model use behind an answer. Review records from native chat (iKB's built-in interface), the web widget, Slack, Microsoft Teams, Chatwoot and API conversations. It does not expose a model's hidden reasoning.

### Sealed audit exports

Export Decision Records as JSON or PDF packages. Integrity checks help detect changes. Set access and retention rules. Conversation evidence packages support signed links that expire and audit records for security monitoring systems (SIEM).

### Evidence checks and quality signals

Monitor evidence checks without changing answers, or enable filtering of clearly irrelevant material while keeping protected evidence. A low evidence score does not itself add an automatic refusal rule. Optional native-chat quality scores and checks after an answer inform administrators without rewriting the delivered answer.

### Policy enforcement and audit history

Keep prompt versions and apply access rules, moderation and prompt-injection controls. Check and remove restricted content from outputs. Require permission for source-based inference, external tools and destinations. Linked audit events are append-only and hash-chained to help detect changes. They record configuration, sources, decisions and usage.

Architecture and controls depend on your deployment. Test source quality, permissions, privacy, model failures, retention and exports against your requirements.

Continue exploring

## More from architecture.

[Deployment options](https://kbmy.ai/architecture/deployment) [Data & retrieval](https://kbmy.ai/architecture/data-retrieval)

[View the architecture overview](https://kbmy.ai/architecture)

## Test iKB on your team's questions.

Bring sample sources and clear requirements. Review the answers and controls before rollout.

[Book a demo](https://kbmy.ai/contact)