What FAT and SAT Should Actually Prove Before Go-Live
Start with acceptance objects, not ceremony dates
Before you schedule rooms and flights, list what must be true: safety functions behave as specified; cycle and throughput sit inside an agreed band under a defined load model; quality outputs meet the sampling plan; error handling and recovery work under realistic faults; data and MES handshakes pass agreed messages; documentation and training let operators run standard work. Anything not listed will not be tested—it will be argued about later at higher cost.
What a serious FAT produces
You should leave FAT with traceable test records mapped to requirement IDs, a punch list with owners and dates before shipment, explicit notes on what was simulated versus exercised for real, and frozen identifiers for software and firmware builds. Weak FATs trade on subjective “looks good,” moving targets (“we will tune on site”), and quiet substitutions in tooling, parts, or builds. Manufacturers should refuse that ambiguity.
What a serious SAT produces
SAT confirms plant-specific assumptions, closes gaps with a bounded stabilization window and measurable exit criteria, and produces a handoff that states what is supported on day one versus what is a later-phase improvement. Weak SATs sign acceptance while interlocks are bypassed “temporarily,” optimization is infinitely deferred, or training is sacrificed to production pressure.
Plant-side reality: “small gaps” are not small
Under fatigue and schedule pressure, unresolved issues get rebranded as startup noise. If a gap touches safety, ownership, repeatability, or recovery behavior, it is not noise—it is unclosed risk waiting for the first real production week.
A three-question gate (use at FAT and SAT)
Before you sign any acceptance step, ask: does it meet written criteria with agreed evidence; are known gaps documented with owners, dates, and explicit risk acceptance where required; can operations execute standard work without heroic intervention? If the third answer is no, go-live is a bet, not a decision.
When to pause
Pause when scope changes arrive as casual tweaks without change control, test materials are unrepresentative and undocumented, site staffing does not match the test plan, or internal owners (maintenance, IT, quality) are absent so defects will have no home. Pausing is cheaper than rework on a live line.
How DBR77 Marketplace ties back
Acceptance discipline should trace to what was compared, contracted, and promised before award. That keeps FAT and SAT connected to buying logic instead of floating as detached rituals.
For continuity across contract and execution handoff, see What to Check Before Signing an Automation Contract and What a Clean Handoff From Selection to Delivery Should Look Like.
Acceptance as contract with the floor
FAT and SAT are where abstract scope becomes lived reality. Operators should recognize the tests as their world: real guarding, real materials where applicable, real recovery scenarios, real data paths. If tests are “close enough,” you are not validating production—you are validating a story. That distinction shows up the first time the line runs under customer pressure.
Good acceptance discipline also protects suppliers who did the work correctly. When criteria are explicit, strong performers can prove completion without endless opinion battles. Weak criteria punish everyone by turning completion into negotiation.
Bottom line
FAT proves integrated performance against contract criteria with records. SAT proves the same in your context with operational ownership. Define acceptance early—or pay for ambiguity in the first real production week.
DBR77 Marketplace helps manufacturers keep scope, interfaces, and accountability visible early so acceptance criteria are harder to postpone into the go-live week. Describe your challenge or Start manufacturer demo.