The clock can start before quality sees the complaint.

A field representative receives a message that a device stopped during use. Customer service later forwards it to quality. The delay is not just queue time: depending on the product, reporter, jurisdiction, and facts, the organization may already possess information that starts a regulatory reporting assessment.

The first job of complaint software is therefore not to open a ticket. It is to preserve what was received, when it was received, by whom, and when it reached a person or function whose awareness matters under the applicable rule.

A single call can describe damaged packaging, a device malfunction, a dosing error, a lack of effect, and a patient injury. Those are not interchangeable labels on one ticket.

Seal creates one source intake and separates its issues into linked quality, medical, vigilance, counterfeit, distribution, and service work. Shared facts stay shared. Each issue keeps its own classification, owner, rule, clock, decision, and closure criteria.

Channel, reporter, contact permissions, receipt timestamp, original language, verbatim narrative, attachments, product identifiers, lot or serial number, use context, patient impact, event location, and distributor path are preserved.

Subsequent summaries do not overwrite the original report. Translation, normalization, coding, and follow-up are versioned transformations with authors and dates.

The system records receipt and awareness events rather than inferring one date from case creation. Role, organizational unit, source channel, forwarding history, business calendar, and applicable rule remain available to the assessment.

If an awareness date changes after review, the previous value, rationale, approver, affected deadline, and recovery action remain in the audit trail.

Structured questions surface death, serious injury, hospitalization, intervention, malfunction during use, contamination, sterility concern, wrong product, mix-up, counterfeit suspicion, pediatric exposure, pregnancy, remedial action, and recurrence risk.

High-risk answers route immediately to qualified reviewers. Triage establishes urgency and workstreams; it does not silently decide causality or reportability from a few keywords.

The complaint record owns product-quality intake, technical investigation, returned product, batch and distribution evidence, customer communication, and complaint closure. A linked safety case owns patient/event assessment, medical coding, expectedness, seriousness, causality where required, ICSR processing, and safety reporting.

The records exchange controlled facts and outcomes. Neither team has to re-key the story, and neither discipline is flattened into the other.

Reportability depends on product type, legal manufacturer or authorization-holder role, jurisdiction, event and outcome, device or medicinal-product facts, remedial action, authority instruction, awareness, and current rule version.

For example, US device manufacturers generally submit qualifying death, serious-injury, and malfunction MDRs within the applicable 30-calendar-day framework; five-work-day MDRs apply in specified circumstances such as certain remedial actions or an FDA request. The system stores the governed rule and calculation rather than turning an example into a universal clock.

Facts considered, missing information, criterion-by-criterion evaluation, rule version, medical and technical input, precedents, final decision, signer, and signature date form the decision record.

A non-reportable conclusion needs as much traceability as a reportable one. New facts can reopen the decision without erasing what was known at the earlier point in time.

The investigator defines the allegation, hypotheses, evidence needed, product-return plan, manufacturing-record review, distribution review, comparable events, methods, assignments, and due dates.

The investigation distinguishes observed fact, reporter statement, test result, inference, and conclusion. “No defect found” is not a complete investigation unless the tested conditions and limitations are visible.

Product, UDI or catalog identifier, lot, serial number, batch, component lots, manufacturing site, line, equipment, supplier, release record, shipment, consignee, and destination link through governed identifiers.

If the lot is unknown, the record says so and follow-up attempts remain visible. A guessed lot never becomes master data merely because it was needed for a dashboard.

Return request, authorization, shipping kit, chain of custody, receipt, condition, decontamination, storage, photographs, destructive-test approval, examination, retention, and disposition form the returned-sample lifecycle.

When no product is returned, the reason, attempts, substitute evidence, and investigation limitation are explicit. The workflow adapts without pretending the missing evidence exists.

Method, instrument, software, analyst, raw-data location, sample condition, acceptance criteria, controls, deviations, observations, and approved result remain connected to the question each test addresses.

The complaint may consume a result, but it does not replace LIMS, SDMS, calibration, or engineering evidence systems. References remain durable and reviewable.

Report type, jurisdiction, version, due-date basis, required fields, source facts, narrative, review, electronic validation, dispatch, gateway acknowledgement, authority identifier, follow-up need, amendment, and closure are linked to the assessment.

Submission status reflects technical and authority acknowledgements. “Sent” is not automatically “accepted.” Follow-up information can update a report while preserving prior versions.

Counts alone can mislead when sales, procedures, units in use, reporting channels, or coding practices change. Trend definitions retain the complaint population, exposure denominator, product hierarchy, code set, time window, exclusions, baseline, alert method, and version.

A statistical alert is a review candidate, not proof of a defect. Review outcomes may create a focused investigation, CAPA, risk-management update, signal assessment, supplier action, or documented no-action decision.

Controlled issue codes support stable reporting, while the original narrative remains searchable. Synonyms, previous code versions, product generations, failure modes, harms, lots, suppliers, and regions can be examined together.

Machine assistance may suggest codes or similar records. Reviewers see why a match was suggested and confirm or reject it; the model does not manufacture a trend by silently recoding history.

A complaint may initiate nonconformance review, supplier corrective action, CAPA, field action, recall assessment, design change, labeling change, training, or risk-management review. The source complaint and evidence remain linked.

Effectiveness checks use defined outcomes and time windows. Closing the CAPA does not auto-close complaints, and closing individual complaints does not prove a systemic issue is controlled.

Acknowledgement, questions, return instructions, status updates, final response, replacement, credit, and follow-up preferences are controlled communications. Templates help consistency but require case-specific review.

Role-based access, redaction, consent, minimum-necessary views, retention, and jurisdictional privacy controls protect patient and reporter data while technical teams receive the information they need.

Closure verifies classification, required workstreams, follow-up attempts, investigation, product evaluation, reportability decisions, regulatory reports and acknowledgements, customer response, linked quality actions, coding, trend inclusion, and record completeness.

Open follow-up or external action may be dispositioned under a controlled policy rather than hidden. Reopening preserves the previous closure decision and explains the new evidence.

Leaders can see untriaged intakes, awareness reviews, approaching statutory deadlines, overdue follow-up, returned products awaiting evaluation, investigations by risk and age, unresolved report acknowledgements, recurring issue families, affected lots, CAPA links, and response performance.

Every metric drills to source records and denominator logic. The dashboard supports action; it is not a parallel complaint database.

For any complaint, Seal can show the original report, receipt and awareness chronology, triage, issue split, follow-up, product genealogy, return custody, tests, investigation reasoning, safety handoff, reportability decisions, filed reports, acknowledgements, communications, trends, quality actions, signatures, and closure.

That record explains both what the company did and what it knew when each decision was made.