Beyond the Checkbox: Interrogating an FAI Report Like a Quality Engineer

The First Article Inspection occupies a strange position in American manufacturing culture. Formally, it is a rigorous discipline — for aerospace work, AS9102 prescribes exactly which features get measured, against which standards, with what documentation. Informally, in many procurement organizations, it is a PDF that gets attached to the supplier file and never opened again. The buyer sees the three-letter acronym, sees the word "complete," and moves on.

This is a missed opportunity bordering on negligence. A well-produced FAI is the single richest data document a supplier will ever hand you before production begins. Read carefully, it tells you not only whether the first part was good, but whether the shop that made it understands its own process. Read carelessly, it will happily tell you nothing at all.

What an FAI Actually Represents

Start with what the document is: a full dimensional, material, and process verification of the first part (or first lot) produced under production conditions — production tooling, production operators, production process parameters. That last clause matters. An FAI run on a machinist's personal pet setup tells you about one machinist. An FAI run on the production cell tells you about the system.

The AS9102 format breaks into three forms: Part Number Accountability, Product Accountability (the balloon drawing and results table), and Characteristic Accountability with process verification. Buyers tend to stare at the results table — pass/fail, pass/fail — and stop. The engineering insight lives elsewhere.

The Measurements That Predict Capability

Not every dimension on an FAI carries equal information. Three categories deserve disproportionate attention:

Datums and datum-adjacent features. How the part is located during measurement determines whether every other number on the report means anything. Check that the FAI references the same datums as the drawing, and that the inspection setup — CMM fixturing, part orientation — is documented. A beautiful column of measurements taken off uncontrolled datums is a work of fiction.

Features spanning multiple operations. A bore machined in one setup is easy. A bore machined in setup two, referenced back to features from setup one, exposes the shop's real process control. These cross-operation characteristics are where capability problems hide, and they are exactly where weak shops report the least data.

The tightest tolerances on the print. Find the two or three most demanding features and look at the actual measured values, not just the pass/fail column. A dimension with a ±0.0003" tolerance reported as "0.5000, 0.5000, 0.5000" from three parts is either a very good process or a very quiet one — you cannot tell which from the FAI alone, and that ambiguity itself is information.

Red Flags: Reading What Isn't There

The most instructive FAI failures are omissions. Watch for these patterns:

• Tight Cpk statistics on non-critical dimensions. If the report brims with process capability data on easy-to-hold features but goes quiet on the hard ones, the quality system is optimizing for the report, not the part. Capability analysis is expensive; selective capability analysis is a choice.
• Missing datum scheme on the ballooned drawing. Balloons that don't map to the drawing's GD&T framework suggest the inspection was performed from a model or a habit rather than the controlled print.
• Generic measurement methods. "Caliper" on a ±0.0005" feature is not a measurement method; it is a shrug. The FAI should state the instrument and, ideally, its calibration reference.
• Copy-paste consistency across revisions. If your FAI for Revision C contains characteristic notes from Revision A, the report was assembled from a template, and you should wonder what else was.
• A perfect report. No first article is perfect. An FAI with zero marginal results and zero commentary across forty characteristics suggests either exceptional process control or an inspection sheet filled out backward — measured to fit, rather than fit to measure.

Using the FAI as a Vetting Instrument

Treat the FAI as a live interview with the supplier's quality system. A few practices convert it from filing-cabinet filler into genuine due diligence:

Request the raw measurement data, not just the summary form. Willingness to share actual readings, CMM programs, and setup documentation signals confidence. Hesitation signals something worth investigating.

Ask one supplier to explain a marginal dimension. Pick the closest-to-limit result and ask how it will be controlled in production. The quality of the answer — SPC plan, tool wear compensation, in-process checks — tells you more about the shop than any audit checklist.

Compare FAIs across suppliers on the same part family. When two shops inspect identical drawings, the differences in datum discipline, measurement methods, and data density become vivid. It is the cheapest supplier benchmarking exercise available.

Re-baseline on revisions. Every drawing revision deserves at minimum a delta-FAI on changed and affected characteristics. Suppliers who volunteer this are protecting your program; suppliers who need reminding are protecting their schedule.

The Report as a Mirror

Here is the uncomfortable truth: the FAI reflects the buyer as much as the supplier. Teams that demand complete, honest first article data — and reward suppliers for surfacing marginal results rather than hiding them — end up with better reports and better parts. Teams that treat the FAI as paperwork to be satisfied teach their supply base exactly that.

A clean FAI report should tell you three things: the part conforms, the process that made it is understood, and the people who made it will tell you the truth about both. If your current FAIs are telling you only the first thing, you are getting a fraction of the value you are paying for — and a fraction of the protection the document was designed to provide.