The Corrective Action Nobody Closes

Every manufacturer has a corrective action log. Very few have a corrective action log that reflects reality.

Open one at a mid-market plant and the pattern is consistent. There are items raised eleven months ago against a supplier who has since changed account managers twice. There are items marked "pending supplier response" where the last email went out in March. There are duplicates — the same defect raised by two different inspectors under two different part numbers. And there are items that were genuinely resolved on the floor months ago, fixed properly, verified informally, and never closed out on paper.

The log is not a record of quality problems. It is a record of quality problems that someone had time to write down.

Why the backlog forms

Nobody decides to let corrective actions rot. The backlog is a byproduct of how the work is distributed.

Raising a corrective action is fast and it is somebody's job. An inspector finds a non-conformance, documents it, assigns it to a supplier, and moves on. That part of the process has a clear owner and a clear trigger.

Closing one is slow and it is nobody's job in particular. Closure requires chasing the supplier for a root cause submission, reading that submission critically enough to tell a real fix from a reworded apology, verifying that the containment actually held across subsequent lots, updating the supplier scorecard, and filing the evidence in a form an auditor will accept. Five or six discrete tasks, spread across weeks, each depending on someone else's response time.

In a large manufacturer, a supplier quality engineer owns that sequence full-time. In a mid-market plant running two hundred suppliers with a quality team of four, the sequence belongs to whoever has a slow afternoon. Slow afternoons are rare. The backlog grows at exactly the rate at which the team is busy.

What the backlog actually costs

The temptation is to treat an open corrective action log as an administrative untidiness — a housekeeping item to be cleared before the next audit. That underrates it in three specific ways.

Repeat defects go unpriced. A corrective action that is never verified is a fix you are trusting on the supplier's word. When the same defect reappears two quarters later, it arrives as a fresh problem rather than a pattern, and it gets negotiated as a one-off instead of as grounds for re-sourcing. The plant absorbs the cost of the same failure repeatedly and never assembles the evidence that would justify changing suppliers.

Scorecards lose authority. Supplier scorecards are the main commercial lever quality has. If the scorecard is built on a log that everyone knows is stale, the supplier discovers this quickly. The first time a supplier successfully argues that a scored item was resolved months ago, the whole instrument weakens — and the next negotiation happens on price alone.

Audits become projects. A current corrective action log is a two-hour retrieval exercise. A twelve-month backlog is a three-week reconstruction effort involving people who should be doing other work, usually right when the plant is also trying to close a quarter. The cost does not vanish; it clusters, and it clusters at the worst possible moment.

None of these show up as a line item. They show up as margin that quietly failed to materialise.

The mid-market squeeze

The obvious fix is to hire a supplier quality engineer whose only job is corrective action closure. Mid-market operators run into two walls.

The first is availability. A credentialed supplier quality engineer with real automotive or food-safety experience is a scarce, expensive hire, and the search runs three to six months before a further two to four months of ramp. The backlog does not pause during that window.

The second is shape. Most of the closure sequence is not senior work. Chasing a supplier for an overdue root cause submission, checking whether containment held across the last twelve lots, formatting evidence into an audit-ready packet, updating a scorecard field — this is structured, repeatable, deadline-driven work. It consumes the majority of the hours and requires very little judgement.

The judgement is concentrated in a small number of moments: deciding whether a submitted root cause is credible or cosmetic, deciding whether a recurring defect warrants escalation to re-sourcing, deciding what an auditor will accept as sufficient evidence. Those decisions are genuinely senior. They are also a fraction of the total effort.

Hiring for the judgement means paying senior rates for a role that spends most of its week on follow-up. Not hiring means the judgement calls never get made at all, because nobody has assembled the information they depend on.

Splitting the work

This is the split h.work is built around. AI Specialists handle the throughput; credentialed human experts hold the judgement.

A Supplier Quality Documentation Specialist deploys into the channels a plant already runs — email, Teams, the existing quality system — and takes ownership of the closure sequence. It tracks every open corrective action against its due date, chases suppliers on schedule rather than when someone remembers, cross-references new non-conformances against the historical log to surface repeats, assembles evidence packets in audit-ready form, and keeps scorecards current rather than quarterly.

When a root cause submission arrives, the Specialist prepares it for review and routes it to a senior quality expert from the consortium — someone with real supplier quality credentials, identity-verified and credentialed through Humanity — who decides whether the fix is real. Recurring defect patterns get escalated with the history attached, so the re-sourcing conversation starts from evidence. Routine follow-up runs continuously and is monitored. Every action is logged.

The economics work because oversight scales in a way headcount does not. One consortium expert can supervise Specialists serving ten to thirty companies, which is why senior judgement lands at 20–40% of the fully loaded cost of the hire you would otherwise make. For a Manager-tier deployment, that is $3K–$6K a month against a mid-level supplier quality hire at $90K–$160K loaded — on a monthly contract, cancellable with thirty days' notice, deployed inside a day rather than a quarter.

The test worth running

Before evaluating any of this, run one exercise. Pull your corrective action log and sort by age. Count the items older than ninety days. For each of the ten oldest, answer two questions: was this actually fixed, and can you prove it to an auditor?

Most mid-market plants find that the answers diverge — a good number were fixed, and almost none can be proven. That gap is the backlog's real shape. It is not a paperwork problem. It is a record of judgement that was never applied, because the follow-up work required to make it possible never got done.

The throughput is what is missing. The expertise is available. h.work exists to connect the two.