Platform FAQ

Aegis SafeForge is an AI-assisted engineering workflow platform for functional safety, cybersecurity, and regulated risk work. It connects project context, source documents, analysis tables, requirements, reviews, approvals, traceability, and evidence in one governed workspace.

ASF Core covers context, requirements (high- and low-level) and baselines, HARA, safety goals, FSC/TSC, FMEA, FMEDA, FTA, ETA, DIA, TCA, FHA, HAZOP, LOPA, verification, railway RAMS and assurance records, and the ISO 26262 safety case, confirmation review, audit, assessment, configuration baseline and safety release (manual entry). Five configurable profiles cover automotive, railway, medical devices, process industry and aviation. TARA and cybersecurity workspaces are beta. ASF does not calculate PMHF; FMEDA shows it as not computable, with the reason.

The CRA Compliance Cockpit is a separate pilot offering for classification, organisation and product gaps, evidence, SBOM, vulnerabilities, VEX, remediation, conformity exports and post-market reporting preparation. ASF Core connects engineering artifacts and review evidence. EU AI Act workflows are offered as a pilot. Qualified people review compliance decisions and authority submissions.

No. AI accelerates drafting, retrieval, and structuring. Qualified people remain responsible for reviewing assumptions, changing generated content, approving engineering decisions, and determining compliance.

Artifacts can be linked across the engineering chain, including item definitions, hazards, threats, goals, requirements, controls, claims, and verification evidence. Aegis SafeForge is designed to build traceability while work happens, so teams do not have to reconstruct spreadsheet relationships right before review. Each record shows upstream and downstream links; a traceability canvas, a V-model gap view and lifecycle models with evidence gates show what is missing before review.

Spreadsheets are flexible but weak on traceability, ownership, and review governance. Generic ALM tools are powerful, but often require heavy setup and are not safety-workflow-first. Generic AI tools can draft text, but they do not manage review state, approval, deterministic checks, source grounding, or downstream impact. Aegis SafeForge connects workflow structure, AI assistance, human approval, traceability, and evidence in one governed environment.

Generation can use the active system and project context, selected standards, approved upstream artifacts, structured schemas, and documents made available to the relevant organisation, project, or system scope.

Yes. Each AI-suggested row has an Explain panel. Where the generation recorded them, it shows the rationale, which fields the model filled and which came from templates, a confidence value, and the sources used, per document and per passage, with relevance score, scoring method, page number and highlighted passage. Anything that was not recorded is marked as not recorded.

No. AI rows start as “AI Suggested”, go to “Needs Review”, and only a person can approve them; regeneration never overwrites approved work. AI can also be switched off entirely.

Yes. Set the AI mode to Off. Every workspace works manually; Conservative, Balanced and Exploratory set how much the AI proposes.

The platform uses authenticated access, organisation and project boundaries, role-based permissions, and protected service-to-service communication. Deployment-specific controls and customer responsibilities are confirmed during onboarding.

In production, browser traffic is intended to go through the authenticated backend. The retrieval and generation service is treated as an internal service and protected with service credentials.

Aegis SafeForge supports managed SaaS, region-specific SaaS, private cloud or VPC, and self-hosted deployment models. EU companies can request EU servers for SaaS. Regional availability, data-residency scope, and architecture are confirmed during onboarding.

No. Aegis SafeForge supports engineering workflows and audit preparation. It does not certify products, guarantee compliance, or replace independent assessment where one is required.

Contact the Aegis SafeForge team to discuss your standards, workflow, deployment constraints, and pilot scope. Early-access evaluations are planned around a defined use case and review process.

Yes. Requirements import via ReqIF 1.0.1 (and file import); imported rows keep their source tool and file as provenance and start in an “Imported” state. Project documents and standards are indexed for search and AI grounding. Migration scope is confirmed per pilot dataset.