The part doesn’t care about your procedures

A conventional QMS excels at governance (approvals, training records, CAPA workflows, audit trails around documentation). It’s designed to ensure an organization behaves consistently. But AM has a habit of turning “consistency” into a data problem. Because AM quality evidence is not a single document. It’s a chain of linked facts such as powder state, build parameters, machine events, post-processing route, inspection outcomes, and (crucially) what changed since the last accepted build.

If that chain lives in spreadsheets and shared drives, you can still ship parts. You can even pass audits. But you do it the hard way, by reconstructing reality under pressure. That’s not quality management. That’s quality theater.

Auditors don’t want a presentation

If you’ve ever sat in a serious audit, you know the feeling. The auditor isn’t asking for your “best” slide deck. They’re asking for the one thing a document-centric system can’t reliably produce on demand, namely a coherent, part-level story that survives cross-examination.

In production AM, the questions that matter rarely live inside one document. They cross the entire chain. Which powder batch and reuse state was used? Which parameter set was active? What post-processing route was followed? Which inspection results apply to this exact serialized part? What changed since the last accepted run?

If answering these questions requires manual hunting, you don’t have an audit trail, you have an investigation.

From paper evidence to process evidence

This is why AM is forcing a quiet reinvention of quality management. Not because quality suddenly matters more (aerospace and medical have always cared). But because AM changes what “quality evidence” is. In powder-bed AM, evidence is more granular, more interconnected, and more sensitive to drift over time. So the center of gravity moves from document control to process evidence.

Process evidence is not a folder. It is structured, queryable, and connected at part level. It makes it possible to answer hard questions quickly and consistently because the relationships between powder, build, post-processing, and inspection are enforced, not implied.

This doesn’t replace your QMS. It complements it. The answer isn’t “throw out your QMS”, it’s stop forcing your QMS (or your MES) to become something it was never designed to be.

All process evidence, from powder to final part, is linked on part-level. Source: amsight

Stop bolting “quality” onto MES and calling it solved

I’ll say it plainly, because it will save CTOs and operations leaders a lot of pain. MES is not a quality system for AM. It’s an execution system. It schedules, routes, timestamps, and records sign-offs. It is the heartbeat of workflow, not the brain of AM evidence.

What happens in many organizations is predictable, they try to “solve AM quality” by adding custom fields, attachments, and bespoke workflows into MES. Then they wonder why their reporting still ends up in Excel. It’s because they’re trying to make a workflow tool do the job of an evidence tool.

The missing layer is not another platform

For CTOs and integration partners, it helps to stop thinking in product names and start thinking in types of truth. The production stack contains different truths that should not be pushed together:

• ERP holds commercial truth (orders, finance, inventory).
• MES holds execution truth (workflow, routing, status).
• A document-centric QMS holds governance truth (procedures, approvals, compliance framework).
• And a production-level AM quality layer holds process-evidence truth: powder → build → post-processing → inspection, linked at part level.

That last one is the gap most organizations feel but struggle to name. They have “systems,” yet when a customer asks a pointed question about a part, the answer still lives in spreadsheets.

Why AM-specific modelling matters

AM fleets are heterogeneous. Different OEMs, different generations, different log formats, different inspection flows. Traditional systems can store attachments and notes, but they rarely normalize AM data into a consistent model that supports cross-machine comparison, drift detection, or repeatable reporting. And that modelling is not a “nice to have.” It changes what you can do operationally:

You can move beyond “prove it passed” to “prove it’s stable.” You can build repeatable reporting templates instead of rebuilding evidence packs per customer. You can run SPC and trend analysis without starting every month by cleaning spreadsheets.

That’s not just compliance improvement. That’s cost reduction. That’s fewer scrapped builds. That’s fewer weeks lost to non-conformity investigations that begin with “where is the data?”

Qualification is the moment the truth arrives

If you want to know whether your AM quality approach is built for production, don’t look at your best prototype. Look at your qualification package. IQ/OQ/PQ-style thinking forces a process-level view. You’re not only proving the part, you’re proving the process can produce the part repeatedly.

This is why “Process & Machine Qualification” naturally sits next to “Prove Part Conformity” in production AM. One is evidence for this part, the other is evidence that the process is stable enough to trust the next hundred.

If your qualification evidence lives in spreadsheets and PowerPoints, you can still qualify. But sustaining and scaling that qualification becomes painful. Regulated customers don’t just want a one-time dossier, they want confidence that stability holds over time.

The simplest test for leaders

If you want to know whether you’re document-heavy but evidence-light, try this exercise. Pick one shipped part. Now ask whether you can retrieve (within minutes) a coherent chain of evidence such as powder history, build job, post-processing route, inspection results, and a clear record of what changed since the last accepted run.

If the answer involves multiple files, multiple people, and manual collation, you haven’t failed. You’ve simply found the gap. And that gap is exactly what we want to address, making AM process evidence usable at scale, so “audit-ready proof” becomes the default output of production rather than an emergency project the night before an audit.

Conclusion

AM isn’t killing the QMS, it’s evolving it. In a world where production is data-rich and regulated customers demand speed, consistency, and proof, quality management can’t stop at documents. It must include process evidence that behaves like a system.

If that sounds like “more software,” I’d argue the opposite. It’s less chaos. It’s fewer spreadsheets. Fewer bespoke workarounds. Fewer heroic individuals keeping the quality story alive. And in production AM, that is what maturity looks like, not more paperwork, but better evidence.