Holes drilled under agent control
41.6M
▲ +32% increase
Seven agents that do the work, not just advise on it
Each agent owns one part of the structural build. They share one perception layer, one as-built twin and one conformance record — so a decision made at the drill is visible at the join, and provable at the audit.
| Agent | Headline metric (%) |
|---|---|
| Drill & Fasten | 99.4% |
| Hole & Fastener Inspection | 99.1% |
| Align & Shim | 98.4% |
| Seal & Join | 99.4% |
| Line & Takt | 96.8% |
| Robot & Handling | 99.7% |
| Quality & Conformance | 97.9% |
Jump to an agent
Every agent below is deployable on its own. Most factories start with one and expand as autonomy is earned.
Drill & Fasten
Adaptive control of drilling, countersinking and rivet/bolt installation.
Hole & Fastener Inspection
Vision + in-process metrology sensing of hole, countersink, fastener, gap and FOD.
Align & Shim
Metrology-assisted alignment and predictive shimming that removes hand-fit loops.
Seal & Join
Sealant application control and fuselage / wing-body join sequencing.
Line & Takt
Moving- and pulse-line station balancing, takt optimisation and travelled-work control.
Robot & Handling
Crawlers, AGVs, positioners and cranes moving and holding large structures.
Quality & Conformance
Right-first-time, non-conformance, rework and full airworthiness traceability.
As-built twin
Not an agent — the arbiter. Every agent’s plan is simulated here before it reaches a machine.
Drill & Fasten
Adaptive control of drilling, countersinking and rivet/bolt installation.
Closes the loop on feed and speed, one-up assembly, burrless hole quality, countersink depth and fastener interference/torque — per stack-up, per material, per fastener code.
- Holes controlled / shift
- 18,400
- Countersink depth σ
- 0.011 mm
- Adaptive feed decisions/s
- 240
| Depth through stack (mm) | Adaptive feed | Fixed feed | Burr risk index |
|---|---|---|---|
| 0 | 100 | 100 | 4 |
| 2 | 98 | 100 | 6 |
| 4 | 94 | 100 | 14 |
| 6 | 91 | 100 | 27 |
| 8 | 93 | 100 | 22 |
| 10 | 96 | 100 | 11 |
| 12 | 95 | 100 | 9 |
| 14 | 94 | 100 | 12 |
| 16 | 96 | 100 | 8 |
Hole & Fastener Inspection
Vision + in-process metrology sensing of hole, countersink, fastener, gap and FOD.
Predicts hole diameter and roundness, countersink flushness, fastener seating, edge distance, interlaminar gaps and foreign object debris before the structure moves to the next station.
- Detection recall (FOD)
- 99.1%
- Flushness resolution
- ±0.008 mm
- Inference latency p99
- 38 ms
| Characteristic | Rivetira in-process | Sampled manual inspection |
|---|---|---|
| Diameter | 100% | 25% |
| Roundness | 100% | 10% |
| C/sink depth | 100% | 25% |
| Flushness | 100% | 40% |
| Seating | 100% | 100% |
| Edge dist. | 100% | 15% |
| FOD | 100% | 60% |
Align & Shim
Metrology-assisted alignment and predictive shimming that removes hand-fit loops.
Fuses laser tracker and photogrammetry clouds into a live fit model, predicts the gap field across the join, and generates machined shim geometry — replacing iterative hand-fit with one pass.
- Gap prediction MAE
- 0.031 mm
- Shim iterations
- 1.0 (from 3.4)
- Alignment cycle
- 22 min
| Fit-up iteration | Hand-fit baseline | Predictive shimming |
|---|---|---|
| Iteration 1 | 100% | 100% |
| Iteration 2 | 84% | 6% |
| Iteration 3 | 46% | 0% |
| Iteration 4 | 12% | 0% |
Seal & Join
Sealant application control and fuselage / wing-body join sequencing.
Controls bead geometry, fillet profile and cure window, and sequences the join so that mating, fastening and sealing happen in the order that protects tolerance stack-up.
- Bead width CV
- 4.2%
- Sealant waste
- −37%
- Join sequence steps
- 1,180
| Sealant bead | Bead width |
|---|---|
| Bead 1 | 6.12 mm |
| Bead 2 | 6.08 mm |
| Bead 3 | 6.15 mm |
| Bead 4 | 6.11 mm |
| Bead 5 | 6.09 mm |
| Bead 6 | 6.13 mm |
| Bead 7 | 6.10 mm |
| Bead 8 | 6.12 mm |
Line & Takt
Moving- and pulse-line station balancing, takt optimisation and travelled-work control.
Continuously rebalances station content against real cycle times, predicts where a position will breach takt, and proposes the resequencing that keeps the line moving.
- Stations balanced
- 46
- Takt breach warning
- 4.2 h ahead
- Travelled work
- −44%
| Station | Station cycle time |
|---|---|
| FA-01 | 41 h |
| FA-02 | 47 h |
| FA-03 | 38 h |
| FA-04 | 52 h |
| FA-05 | 44 h |
| FA-06 | 39 h |
Robot & Handling
Crawlers, AGVs, positioners and cranes moving and holding large structures.
Drives and supervises the machines that carry the load — flex-track crawlers, mobile platforms, AGVs and cranes — with fail-safe stop and structural-load awareness at every step.
- Fail-safe stop
- <40 ms
- Crawler positioning
- ±0.05 mm
- Handling events/day
- 2,900
| Week since go-live | Unplanned handling stops | Handling events (thousands) |
|---|---|---|
| Wk 1 | 14 | 42 |
| Wk 4 | 11 | 51 |
| Wk 8 | 8 | 58 |
| Wk 12 | 6 | 64 |
| Wk 16 | 5 | 71 |
| Wk 20 | 4 | 78 |
| Wk 24 | 3 | 87 |
Quality & Conformance
Right-first-time, non-conformance, rework and full airworthiness traceability.
Every agent action, sensor reading and human override lands in an immutable, assurance-grade record mapped to AS9100 and the program’s airworthiness requirements.
- Traced actions/day
- 1.4M
- NCR cycle time
- −52%
- Audit export
- < 90 s
| Month | Right-first-time | NCRs raised (per 100 airframes) |
|---|---|---|
| Jan | 88.8 | 142.0 |
| Feb | 89.9 | 131.0 |
| Mar | 91.2 | 118.0 |
| Apr | 92.4 | 104.0 |
| May | 93.5 | 94.0 |
| Jun | 94.4 | 83.0 |
| Jul | 95.2 | 74.0 |
| Aug | 96.0 | 66.0 |
| Sep | 96.6 | 59.0 |
| Oct | 97.1 | 53.0 |
| Nov | 97.6 | 48.0 |
| Dec | 97.9 | 44.0 |
Perceive, plan, control, verify, learn
The same loop runs at every station, from a single drilling cell to a whole moving line. Nothing is actuated that has not first been simulated against the as-built twin.
-
01
Perceive
Vision, in-process metrology, laser tracker and photogrammetry fuse into a live as-built model of the structure, hole, gap and fastener state.
38 msFusion latency p99
-
02
Plan
Align, shim, drill, countersink, fasten, seal and join are sequenced against the predicted gap field and the program tolerance stack-up.
4.1 sTwin sim per join
-
03
Control
Feed and speed, interference, torque, bead geometry and crawler position are driven adaptively, with fail-safe stop at every actuator.
240 HzControl loop
-
04
Verify & learn
Hole, countersink, fastener, gap and FOD are sensed, conformance is logged immutably, and supervised corrections retrain the models.
1.4MTraced actions/day
Shadow first. Autonomy is earned, not switched on.
Every agent starts by watching. It is promoted only when its measured accuracy clears the mechanic-plus-metrology baseline and the twin agrees.
-
Weeks 1–4
Shadow
Agent observes the station, predicts every outcome, actuates nothing. Accuracy measured against what the mechanics actually do.
-
Weeks 4–10
Advisory
Agent recommends feed, shim geometry and sequence. A human accepts or rejects; every rejection becomes training data.
-
Weeks 10–20
Supervised
Agent actuates with a human in the loop and a live fail-safe stop. Airworthiness-critical dispositions still require sign-off.
-
Week 20+
Closed loop
Agent runs the step. Humans handle exceptions and the twin gates any change to the control policy.
| Week | Agent accuracy | Mechanic + metrology baseline |
|---|---|---|
| Wk 2 | 88.2% | 94.1% |
| Wk 4 | 91.4% | 94.0% |
| Wk 6 | 93.6% | 94.2% |
| Wk 8 | 95.2% | 94.1% |
| Wk 10 | 96.4% | 94.3% |
| Wk 12 | 97.1% | 94.2% |
| Wk 16 | 97.8% | 94.1% |
| Wk 20 | 98.3% | 94.2% |
| Wk 24 | 98.6% | 94.3% |
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.
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.
Questions engineers actually ask
Anything else goes to the people who build it. Ask an assembly engineer or read the full FAQ.
Yes — that is the normal path. Cell pricing covers one agent on one assembly cell or station. Most factories start with drilling and fastening control or predictive shimming, where the ROI is clearest, then expand.
They share the perception layer and the as-built twin, but each agent has its own policy and its own promotion gate. An agent that reaches closed loop at one station does not automatically actuate at another.
Supervised corrections. Every time a mechanic or liaison engineer overrides an agent, that override — with the full sensed context — becomes labelled training data. This is the compounding assembly-fit-and-quality-data moat.
They fall back to the program tolerance model and run in advisory until enough as-built observations exist to fit a structure-specific model. New program onboarding typically takes six to ten weeks.
Start with one agent, one cell
A line assessment maps one station, quantifies the rework, shim and rate opportunity, and returns a modelled ROI in three weeks. No production disruption.