First Article Inspection: What Designers Need to Prepare Before Mass Production

The first time I sat through a first article inspection review as a junior designer, I mostly watched the quality team argue over measurement details I didn’t fully understand and wondered why it was taking three hours to approve a part that had already passed my drawing review. Twenty years later, having sat through dozens of these reviews as a contract mechanical designer across several Japanese manufacturers, I understand exactly why it takes that long when it’s done properly, and I’ve also seen what happens when it’s rushed. First article inspection, or FAI, is the last real checkpoint before a design commitment becomes thousands of committed parts, and the amount of preparation a designer puts in beforehand determines whether that checkpoint catches real problems or just becomes a paperwork exercise. This article covers what I’ve learned to prepare, and what I’ve learned the hard way not to skip.

Understanding What FAI Is Actually Verifying

First article inspection exists to answer one specific question: does the actual manufacturing process, running under real production conditions, produce a part that matches the design intent, not just a part that happens to measure within tolerance on a good day. That distinction matters because a part can pass individual dimensional checks while still revealing process issues that will cause problems later, like a feature that only holds tolerance because an operator hand-finished it during the first article run, something that won’t happen consistently once production ramps up.

I’ve learned to think of FAI less as a inspection gate and more as a verification that the entire chain — drawing, process plan, tooling, and measurement method — is actually consistent with itself. On a machined housing part early in my career, the first article passed every dimension on the drawing, but during the review the machinist mentioned, almost as an aside, that he’d had to manually deburr a specific internal edge that the CNC program didn’t fully clean up. That comment led us to add an explicit deburring operation to the process plan before production release, something that would have otherwise become an inconsistent, operator-dependent step causing intermittent problems downstream.

Building a Complete and Unambiguous Drawing Package

The single most common cause of a difficult or drawn-out FAI I’ve seen is an incomplete or ambiguous drawing package, not a manufacturing problem. If a critical dimension’s datum structure is unclear, or a tolerance is specified without a clear measurement method implied, the FAI team ends up making judgment calls that should have been made during design. I make it a habit now to review every drawing I release for mass production specifically for measurability — could a quality inspector, without asking me anything, figure out exactly how to verify every toleranced feature.

On a sheet metal bracket project, I had specified a flatness callout without clearly indicating a datum reference frame, assuming it was obvious from context. It wasn’t. The FAI team spent most of a review meeting debating how to fix the part for measurement, and we ultimately had to issue a drawing revision mid-inspection to clarify intent, which delayed the release by nearly two weeks. Since then, I run every GD&T callout through a simple mental test: if I handed this drawing to someone who had never seen the part, could they set it up on a CMM without calling me. If the answer is no, the drawing isn’t ready for FAI.

Preparing the Full Dimensional Characteristic List

For any part of real complexity, I now build a full dimensional characteristic list before the part goes into first article inspection, mapping every toleranced dimension on the drawing to a ballooned number, and cross-referencing that to the inspection report format the supplier or in-house quality team will actually use. This sounds like pure administrative overhead, but I’ve seen enough incomplete FAI reports to know it isn’t. On more than one occasion, a supplier’s report has been missing measurements for several dimensions, sometimes because the ballooning didn’t clearly correspond between the drawing and the report, sometimes because a dimension was genuinely overlooked.

I also flag which dimensions are truly function-critical versus which are more nominal or manufacturing-convenience dimensions, because that distinction changes how much scrutiny I apply to each measurement during review. On a gearbox housing project, I flagged the bearing bore positions and the shaft center distances as critical-priority dimensions requiring CMM verification with full GD&T evaluation, while cosmetic features and non-mating edges got a simpler go/no-go check. That prioritization kept the review focused on what actually mattered instead of spending equal time on every single dimension regardless of functional importance.

Coordinating Measurement Method With Design Intent

A tolerance is only as good as the measurement method used to verify it, and I’ve been surprised more than once by a mismatch between how I intended a dimension to be measured and how it was actually measured during FAI. A true position tolerance evaluated with a hard gauge versus a CMM can produce meaningfully different pass/fail results near the tolerance boundary, and a surface finish spec measured with a contact profilometer versus an optical method can read differently enough to matter on tight specs.

On a precision-machined shaft, I had specified a runout tolerance without thinking hard about how it would actually be checked. The supplier’s standard method used a V-block setup that introduced its own small measurement uncertainty, different from the rotating-center setup I’d mentally assumed when I wrote the tolerance. The measured runout came out right at the tolerance limit, and we spent real time sorting out whether the part was actually marginal or whether the measurement method was adding noise. Now, for any tolerance where the measurement method could plausibly matter, I specify or at least discuss the intended method with the supplier before the drawing is finalized, not after a borderline result shows up.

What I Watch For During the Actual Review

Material certification review is the one item on this list I see designers skip most often, assuming it’s purely a quality department task, but I’ve caught real problems by looking at it myself. On one project, a certification showed a heat treatment lot that technically met the hardness range specified but sat at the very bottom of the allowed band, which was fine dimensionally and metallurgically compliant but gave me enough concern about fatigue margin that I asked for a tighter hardness range on future lots. Nothing in the dimensional inspection would have surfaced that; only reading the cert directly did.

Common Failure Patterns I’ve Seen Repeat Across Projects

A few patterns show up again and again across different suppliers and part types. Datum shift issues, where a part measures fine against one datum setup but reveals problems against another, usually trace back to an ambiguous datum structure on the drawing rather than a manufacturing defect. Tolerance stack-up surprises, where individually acceptable dimensions combine to create an assembly fit problem, usually mean the FAI was reviewed dimension-by-dimension without anyone checking the assembly-level implication, which is why I now always request a trial assembly fit check as part of FAI for any part that mates closely with other components, not just individual dimensional sign-off.

The other recurring pattern is a rushed FAI driven by schedule pressure, where a part gets approved with known open issues on the promise that “we’ll fix it in the next lot.” I understand the schedule pressure is real, but I’ve tracked enough of these promised fixes to know they get deprioritized once production starts and the immediate crisis passes. My rule now is that any open issue identified during FAI gets a written disposition with an owner and a date, reviewed again before the next production lot, rather than a verbal agreement that quietly disappears.

Turning FAI Into a Design Feedback Loop, Not Just a Gate

The most valuable habit I’ve built around first article inspection is treating it as feedback for future designs, not just a checkpoint for the current one. Every time an FAI review surfaces a drawing ambiguity, a measurement mismatch, or a tolerance that turned out to be harder to achieve than expected, I write it down and check whether the same issue exists on other active drawings I’m responsible for, rather than treating it as a one-off fix for this particular part.

Over enough projects, this habit compounds. The categories of problems I catch during my own drawing review before release now overlap heavily with the categories that used to surprise me during FAI years ago, because I’ve internalized the patterns. First article inspection done well isn’t just quality control checking a supplier’s work — it’s one of the clearest windows a designer gets into whether their own drawings, tolerances, and assumptions actually survive contact with a real manufacturing process, and treating it that way has made me a noticeably better designer over time, not just a more careful one.