Part 5 SMS reaches design
and manufacturing.

The full answer is in the highlighted block above the contents list. The rest of this article covers exactly which Part 21 holders the rule names, the plan-then-implement deadlines, what counts as safety data in a design or production shop, and how the four Part 5 components map onto a quality system you already run.

For most of the SMS era, a formal, FAA-accepted Safety Management System was a Part 121 airline requirement. The 2024 final rule changed the perimeter. It revised 14 CFR Part 5 and pulled three new groups inside: all Part 135 certificate holders, commercial air tour operators flying under 91.147 letters of authorization, and certain Part 21 certificate holders. For the design and production world, that third group is the headline. If you want the full picture of the rule across every operator class, start with our Part 5 final rule explainer.

The deadlines follow a plan-then-build pattern. A Part 21 holder that held its certificate on the rule's effective date, May 28, 2024, had to submit an implementation plan to the FAA by November 28, 2024, and must have the SMS fully implemented by May 28, 2027. The plan is not the finish line; it is the commitment that starts a defined build-out window ending in 2027.

Two clarifications save a lot of wasted effort:

• This is not the airport rule.The FAA's airport SMS requirement is a separate regulation with its own citation and its own staggered timeline, and it does not govern design or production organizations. If your work touches a certificated airport, read the Part 139 airport SMS rule on its own terms. Do not assume one SMS deadline covers both.
• Coverage is certain Part 21 holders, not every certificate. The rule names specific approval types rather than sweeping in the whole part. Before you scope a program, confirm your particular certificate is in scope against the current FAA guidance rather than reading a summary table. If your certificate is covered and you held it on the effective date, the May 28, 2027 implementation date is the one that matters.

In an airline, the raw material of an SMS is intuitive: occurrences, hazard reports, flight data, line observations. In a design or production organization the safety signal is just as real but harder to see, because most of it already lives inside the quality system under names that sound like paperwork rather than safety. Four sources matter most:

• Design escapes. A tolerance, assumption or interface decision that looks fine on the drawing but shows up later as an in-service problem. In SMS terms this is a hazard whose consequence is realized far downstream, on an aircraft flown by someone else, months or years after the decision.
• Production nonconformances. Every nonconformance, every deviation, every material review board disposition is a data point about how the as-built article can diverge from the as-designed intent. Closed as quality paperwork, it vanishes. Read as safety data, it feeds hazard identification and trending.
• Escapes and field returns. Parts that left the facility non-conforming, warranty returns, and service difficulty signals coming back from operators are direct evidence that a control did not hold.
• Supplier and sub-tier signals. A nonconformance at a supplier is your hazard, because the article ends up in your product with your name on the approval.

The shift the rule asks for is not “go collect new data.” It is “stop discarding the safety signal already in your nonconformance and escape records.” A production organization that dispositions nonconformances well but never trends them for safety is throwing away its best leading indicator, the early sign that a process is drifting before it produces a genuinely unsafe article.

Part 5 rests on four components: Safety Policy, Safety Risk Management, Safety Assurance and Safety Promotion. The 2024 revisions also require a confidential employee reporting system and a declaration of compliance submitted to the FAA. A design or production organization does not build these from a blank page. A mature quality management system already carries most of the ingredients; the work is mapping them across, then closing the gaps SMS opens up. If the four components are new to you, the four components of SMS are a good primer. Here is how they land in a manufacturing context:

• Safety Policy. A named accountable executive who owns safety risk, defined safety accountabilities, and a stated commitment. Your quality policy and management structure are the starting point. SMS adds the accountable executive and the confidential channel that lets an inspector or engineer raise a safety concern without routing it through the manager whose area it touches.
• Safety Risk Management.The structured identification of hazards and assessment of risk before a change goes ahead. This overlaps your change-control, first-article and configuration processes, but it is broader than conformity. It asks not only “does the article meet the drawing” but “what could this design, process or supplier change do to safety, and is the residual risk acceptable to a named decision-maker.” A risk matrix and a recorded acceptance decision are the visible outputs, and they sit naturally alongside the rest of your quality management records.
• Safety Assurance. The continuing check that controls still work. Internal audit, evaluation, nonconformance trending and corrective action already live here in a quality system. SMS assurance adds a watch on whether the safety risk controls, not just the quality requirements, remain effective, and it prefers indicators over a single pass or fail audit verdict.
• Safety Promotion. Training, competence and communication that make the system lived rather than filed. The confidential reporting culture belongs here too: people report only when they trust the loop closes without blame.

The confidential employee reporting system and the declaration of compliance are the two additions most easily overlooked, because a quality system has no direct equivalent. The reporting system is a governed intake with a closed feedback loop, not a suggestion box, and the declaration is a formal statement to the FAA that your SMS meets Part 5. Neither is a document you write once and shelve; both have to keep running as living parts of the system.

A design or production organization is rarely one site turning raw stock into a finished article. It is a network of sub-tier suppliers, special-process vendors and distributors, and that network is exactly where safety signal gets lost, because the boundaries between organizations are where responsibility turns ambiguous.

Under Part 5, a supplier's nonconformance that reaches your product is your safety data, and the interface itself is a hazard source. A process change at a supplier, an undisclosed sub-tier, a counterfeit-part risk, a special process drifting out of control: each is a hazard that originates outside your walls and lands inside your approval. Managing it means identifying the boundary hazards and holding assurance that the controls at the boundary actually work, not just that a certificate of conformity arrived with the shipment.

The usual failure is structural, not careless. The supplier nonconformance lives in a receiving-inspection record. The field return lives in a warranty system. The internal escape lives in a material review board case. The audit finding against the supplier lives in an audit tracker. Four systems, four owners, no link. So the one question that matters, “is this supplier's risk trending in the wrong direction,” cannot be answered, because the records that would answer it never meet.

This is the gap eAviora is built to close. Instead of a quality database, a document controller, an audit tracker and a spreadsheet of hazards that never talk to each other, eAviora runs occurrences, hazards, investigations, corrective actions, audits, findings, documents, training, compliance and indicators as one operational graph. A production nonconformance, the hazard it reveals, the investigation, the corrective action that fixes it, the audit finding it maps to and the indicator it moves are linked records, not four disconnected copies of the same event.

That linkage buys three things a binder cannot:

• Nonconformances become safety data without re-keying. Because an escape or nonconformance is a record in the same graph as your hazards and indicators, its trend is visible to the safety function the moment it is logged, not after someone transcribes it into a separate safety report.
• Closure gates stop optimistic sign-off. eAviora enforces that a record cannot close over open risk: a degraded barrier requires a linked corrective action, proven effective, before the parent record can close. A nonconformance tied to a live safety risk cannot be quietly dispositioned and closed while the risk stays open, a property proven by a 13-of-13 live scenario suite. See how the enforced closure gates work.
• The Safety Risk Profile is computed, not narrated. eAviora fuses eight components into a single ICAO Doc 9859 aligned four-level score, and runs real statistical process control on indicators using Western Electric rules, so a drifting supplier or a rising escape rate surfaces as a genuine signal rather than a colored arrow someone chose to draw.

None of this starts from a blank hazard log. eAviora ships a pre-built aviation risk model: 101 bow-tie models, 804 named barriers, 210 scenarios and a 610-indicator SPI library, which a design or production organization scopes down to what fits rather than inventing from scratch. Specialist AI agents can draft a hazard analysis or a bow-tie for review, but every AI proposal is human-gated (accept, modify or reject) and audit-logged, AI consumption is metered in plain credits with hard caps, and your data never trains any model. Tenant isolation is enforced by the database itself, not only by application code, which is the answer a supplier or airline security review actually wants to hear. The SMS module is where this lives day to day.