Requirement coverage
Verify whether design, procedure or test evidence explicitly covers the requirements in an approved reference corpus.
SAE industrial intelligence
SAE Industrial Retrieval & Intelligence Unified System
Document intelligence for controlled, traceable engineering review—built to help specialists work through complex repositories without turning engineering judgement over to automation.

Why SIRIUS
Regulated projects generate thousands of requirements, procedures, references and revisions. Manual review remains essential, but searching and repeating it at scale consumes scarce expert time.
SIRIUS prepares the repository, retrieves relevant evidence and structures the comparison. Engineers can then concentrate on exceptions, technical meaning and accountable decisions.
AI supports the engineer; it does not replace engineering judgement.
Where to use SIRIUS
SIRIUS is most useful when a technical question must be answered across many controlled documents and the answer must remain traceable to source evidence.
Verify whether design, procedure or test evidence explicitly covers the requirements in an approved reference corpus.
Find contradictions, revision mismatches and inconsistent technical statements across controlled document sets.
Compare setpoints, thresholds, operating limits and measured values with their authoritative source.
Follow equipment identifiers, KKS codes, signals and references across engineering disciplines and suppliers.
Identify missing acceptance criteria, verification methods or evidence before a controlled project milestone.
Give engineering teams a structured view of an unfamiliar technical corpus before detailed review begins.
How it works
Main capabilities
A structural index, repository profile and domain context are prepared before analysis starts.
Keyword precision, semantic similarity and knowledge-graph relationships reinforce one another.
Extraction, cross-reference lookup and reasoning run as separated, reviewable stages.
Document, page and extracted evidence remain connected to every reported finding.
Dry runs, progress tracking, partial results and runtime controls support long analyses.
Qualified specialists confirm scope, challenge results and remain accountable for decisions.
What makes it different
SIRIUS does not begin with a generic chatbot prompt. It first learns the controlled structure and vocabulary of the repository, then combines deterministic task stages with AI reasoning while keeping the evidence chain visible.
Cooperative operating model
Define the technical question and evaluate meaning.
Protect configuration, evidence and review discipline.
Maintain analysis methods, models and indexes.
Confirm scope, challenge findings and own decisions.
How to start using SIRIUS
SIRIUS is introduced through an SAE-led pilot, not as an open autonomous chatbot. The pilot establishes the corpus, analysis rules, evidence expectations and review roles.
Choose one bounded verification or document-analysis objective.
Identify controlled source documents, implementation evidence and confidentiality boundaries.
Run a readiness check and agree the extraction, comparison and acceptance logic.
Evaluate traceable findings with domain specialists before deciding whether to extend the scope.