Skip to main content
Skip to application content

Trace …

Data Approved sample data

  1. Image analysed
  2. Review action recorded
  3. Batch loaded
  4. Evidence opened
  5. Assistant answered

0 of 5 workflow steps

Guided demo

Evidence Hub

Approved product, technical, policy and evidence-register records that the AI Assistant cites, each carrying a citation identifier, owner and review date.

12 records1 open gaps

Total records

12

Verified

10

Under verification

1

Open evidence gaps

1

Filter records

Search title, citation, summary or keywords.

Batch record schema and validation

CS-SPEC-003 · Technical specification

Verified record
Version
1.8
Owner
Engineering
Review date
2026-08-02
Type
Technical specification

Lists the required and optional fields of a batch record and how invalid rows are reported.

Required fields are record_id, sample_id, status and captured_at. Status must be completed, failed or pending.

Optional fields are failure_reason, output_label, raw_score, review_state, operator, instrument, scenario_version and trace_ref.

Import runs as a preview: rows that fail validation are listed with their row number and reason, and the reviewer confirms before the batch replaces the current view.

Stated limitations

  • Validation is structural. It cannot confirm that a record describes the sample it names.

Dependencies or contradictions

Not supplied

Evidence gaps

Not supplied

batchschemacsvjsonvalidationfieldsimport

Not for diagnosis or clinical decision-making.

Approved sample data only · no patient data · nothing transmitted