Published documents
8
The research behind the claims
Benchmarks with methodology attached, ROI models you can argue with, and buyer guides written for the engineer who has to defend the decision internally.
Holes in benchmark corpus
4.2M
Stations analysed
70
Methodology disclosed
100%
Everything we have published
The assembly bottleneck: why aerospace rate ramps fail in final assembly
A structural analysis of where hours actually go on a narrowbody final assembly line, and why capital spend on more positions rarely fixes it.
Predictive shimming: modelling the gap field before the join
The physics, the data requirements and the measured accuracy of predicting a gap field from as-built part geometry and fixture state.
Hole and fastener inspection: model performance report 2026
Precision, recall and false-negative cost analysis across 4.2 million inspected holes on metallic, composite and mixed stacks.
Autonomy promotion outcomes across 70 stations
How many stations reach each autonomy level, how long it takes, and what causes the rollbacks.
Running a line assessment: a manufacturing engineer’s checklist
What to measure in the three weeks before any vendor touches your station, so the baseline is unarguable.
Writing an outcome-based automation contract
Baseline definition, measurement telemetry, dispute resolution and the clauses that make outcome pricing actually work.
Aerospace assembly labour: the 2026 workforce gap
Retirement curves, apprenticeship throughput and the arithmetic of the skilled-mechanic shortfall through 2032.
ITAR-compliant AI deployment on defense assembly programs
Air-gap architecture, US-person access enforcement and what an accreditation timeline realistically looks like.
Inspection model performance, by material stack
Recall matters more than precision here: a false positive costs a mechanic two minutes, a false negative can cost an airframe.
| Material stack | Recall | Precision |
|---|---|---|
| Aluminium | 99.4% | 97.8% |
| CFRP | 98.6% | 96.2% |
| Titanium | 99.1% | 97.1% |
| Mixed CFRP/Ti | 98.1% | 95.4% |
| Bonded assemblies | 97.4% | 94.1% |
Where the hours actually go on a final assembly line
| Segment | Share |
|---|---|
| Value-added assembly | 34% |
| Fit-up & shimming | 19% |
| Inspection & sign-off | 14% |
| Rework & travelled work | 13% |
| Waiting for parts / crane | 11% |
| Setup & changeover | 9% |
The finding that drives everything else
Two-thirds of an assembly mechanic’s time is not assembly. It is fitting, measuring, inspecting, reworking and waiting — activities that exist because the structure in front of them is not exactly the structure the model predicted.
Adding positions scales all six slices proportionally. Removing the divergence between as-designed and as-built shrinks four of them at once. That is the entire thesis, and it is why Rivetira starts with the as-built twin rather than with a robot.
Value-added share, untreated
34%
Value-added share, deployed
78%
The arithmetic, shown rather than asserted
These are the input assumptions behind the modelled return in a line assessment. Change any of them and the answer changes — which is the point of publishing them.
| Fully loaded mechanic rate ($/h) | 92 | 118 | 145 |
|---|---|---|---|
| Shim hours per join, baseline | 48 | 94 | 160 |
| Shim hour reduction (%) | 35 | 57 | 70 |
| Rework hours per airframe, baseline | 210 | 480 | 820 |
| Rework reduction (%) | 25 | 44 | 58 |
| Escapes per quarter, baseline | 18 | 84 | 190 |
| Escape reduction (%) | 30 | 52 | 71 |
| Cost per escape ($k) | 14 | 38 | 95 |
| Joins per year per cell | 120 | 310 | 640 |
| Modelled annual value per cell ($k) | 286 | 1,108 | 3,410 |
The typical column is the median of 14 deployed cells. Conservative and aggressive are the 10th and 90th percentiles. None of these is a promise; all of them are inspectable.
The gap that is not closing
| Year | Skilled mechanics required | Skilled mechanics available |
|---|---|---|
| 2024 | 100 | 100 |
| 2025 | 108 | 99 |
| 2026 | 117 | 97 |
| 2027 | 126 | 94 |
| 2028 | 134 | 91 |
| 2029 | 141 | 88 |
| 2030 | 148 | 86 |
| 2031 | 154 | 84 |
| 2032 | 160 | 82 |
Written for the person who has to defend the decision
Run the assessment yourself first
Three weeks of honest measurement on one station before any vendor arrives. If your baseline is soft, every subsequent number is arguable.
Insist on outcome-based terms
If a vendor will not tie a portion of fee to a metric they claim to move, ask why. The clause structures that work are published.
Scope the ITAR question early
Air-gap accreditation is the long pole on defense programs. Starting it in month one rather than month six changes the whole timeline.
Where the money actually comes back
Rivetira is priced on the metric it moves. This is the value stack from a representative 14-station wing-box deployment.
| Value source | Annualised value (USD) |
|---|---|
| Shim labour removed | $4.9M |
| Rework hours avoided | $3.6M |
| Rate capacity unlocked | $3.1M |
| Scrap & escapes avoided | $2.2M |
| Overtime reduction | $1.4M |
| Inspection labour | $0.9M |
Annualised value
$16.1M
Platform cost
$2.7M
Payback period
2.4 mo
Net revenue retention
136%
▲ +36% increase
Assembly autonomy, measured on the line
Every figure below is produced by the same telemetry the agents act on — station cycle, hole quality, gap field, fastener state and conformance. Pilot and design-partner aggregate, trailing 12 months.
Holes drilled under agent control
41.6M
▲ +32% increase
Right-first-time, structural joins
97.9%
▲ +9.1% increase
Shim hours removed per join
68%
▼ -68% decrease
Assembly-line uptime
99.94%
▲ +0.3% increase
Aggregate across design-partner lines. Baselines are the same stations before Rivetira, measured over an equivalent period.
What changes on the floor
Plant directors, liaison engineers and airworthiness leads on what autonomy did to their numbers.
“We stopped arguing about whether the gap was 0.4 or 0.6 millimetres. The twin predicted it, the shim came off the machine right, and the join closed in one pass.”
−64% shim hours per join
“The rate ramp was going to cost us four more positions. Instead the line rebalanced itself every shift and we found the capacity inside the stations we already had.”
+38% delivered rate
“Every hole is now inspected, not one in four. My airworthiness record writes itself, and I can hand an auditor a complete trace in ninety seconds.”
100% inspection coverage
Reads and writes back to the machines you already own
Rivetira is not a rip-and-replace. It connects to the drilling machines, crawlers, trackers, shim cells, sealant robots and MES already on your floor over APIs and OT protocols.
Drilling & fastening
- Flex-track crawlers
- Gantry drilling machines
- C-frame riveters
- Automated fastening cells
- Hand-fed ADUs
Metrology
- Laser trackers
- Photogrammetry rigs
- Structured-light scanners
- iGPS / indoor GPS
- Articulated arm CMMs
Shimming & sealing
- 5-axis shim machining
- Liquid shim dispense
- Sealant robots
- Cure monitoring
- Bead inspection
Factory systems
- Assembly MES
- PLM / as-designed models
- ERP work orders
- Quality / NCR systems
- Historians & OPC UA
OPC UA, MTConnect, MQTT, ROS 2, REST and file-drop connectors. Read-only shadow mode first; write-back enabled only after twin validation.
Questions engineers actually ask
Anything else goes to the people who build it. Ask an assembly engineer or read the full FAQ.
Whitepapers and guides are available on request with a work email. Benchmark reports containing customer-derived data require an NDA, because the underlying figures belong to the programs that produced them.
Yes, with attribution, and we would rather you cite the methodology section than the headline number. Figures marked [PLACEHOLDER] are modelled rather than measured and should be cited as such or not at all.
The autonomy promotion benchmark includes rollback rates and the material stacks where recall is weakest. A benchmark that only shows wins is marketing, not a benchmark.
Request the research library
A line assessment maps one station, quantifies the rework, shim and rate opportunity, and returns a modelled ROI in three weeks. No production disruption.