The trust gap
Why generic AI can’t be used on a safety case.
Duty holders can’t put ChatGPT, Claude or Copilot anywhere near this work, for four reasons. Percept is built around every one of them.
Generic AI
Duty-holder security policy blocks public cloud AI outright.
Percept
Deployed where the work is. UK-hosted today, air-gapped tomorrow.
Generic AI
Generic models hallucinate. An ungrounded claim is a regulatory finding.
Percept
Every claim linked to the exact source passage it came from.
Generic AI
No provenance. No way to show a regulator where a sentence came from.
Percept
An immutable audit log: who changed what, when, human or AI.
Generic AI
A general model asked to pretend it understands nuclear.
Percept
A real domain knowledge base, built from your own project documents and standards.
Product Overview
The smartest way to draft, edit, and deliver safety case content.
Draft from your own reference documents, cite every claim, and keep the whole team inside one audited workspace.
Other Features
Project Document Storage
Project Management
Regulation Standards
Secure Cloud Hosting
Plan & Draft Reports
Upload context files
Drag and drop to store files for Percept's AI agents to use when creating the report draft
Expert agentic generation
Percept's agents expertly tailor a draft using our knowledge base for specific report types
Citation engine
Perform due diligence by reviewing auto-referenced excerpts or query Percept AI
AI review
Our agents review the final draft before submitting to regulators, checking compliance
Powerful Features
Comparison
What Sets Percept Apart
CURRENT WORKFLOWS
Manual, offline document drafting
Word/Excel-based collaboration
Static report formatting
Long feedback cycles
Disconnected hazard documentation
No traceability or rationale logs
VS
AI-assisted chapter generation
Structured, in-browser text editor
Regulation-aligned templates
Real-time inline editing + review
Centralised, searchable knowledge base
Explainable AI suggestions
Contact