← Products
01Overview

DutchBoy DEVICE · Advanced Components Manufacturing

Catch the defects hiding inside a spec PASS — before they ship

DutchBoy DEVICE connects measurements, equipment conditions and inspection results from assembly and test lines at the unit (barcode) and lot level. AI anomaly scores that look at hundreds of parameters together find suspect units within spec, and trace which process and parameters are responsible.

Applicable Processes

SMTplacementBondingdispensingActive alignmentassemblyFunctionaloptical testOutgoing QC

Data Analyzed

T

Table

Test measurements · MES history

P

Pulse

Equipment process conditions

V

Vision

Visual inspection images

02Challenges on the Floor

What existing systems tell you — what happened. What's still up to people — why, and what's next.

Challenge 1

Passed spec, still failed at the customer

Pass/fail judgment on each parameter's limits can't catch abnormal combinations that stay within spec.

Challenge 2

Too many parameters

Each process and test step produces hundreds of parameters, and it's hard to tell which ones separate good from bad.

Challenge 3

Tracing takes too long

When a claim comes in, rebuilding a single barcode's process and test history means piecing together several systems.

Challenge 4

Linking data and reporting is manual

Missing or late test files are caught too late, and customer reports are built by hand.

03What Gets Analyzed, What Comes Out

The data you already have, read as one — anomalies, root causes and actions, delivered.

Input · existing systems and data

TTest equipment measurements · by process and test step
TMES process, equipment and · completion history
PEquipment process conditions
VVisual inspection results
Barcodes · lots · customer quality requirements

DUTCHBOY ANALYSIS

1

Collect & clean

Align scattered data on one timeline

2

Detect anomalies

Score deviations from the normal baseline

3

Find root causes

Where, what and why it differed

4

Connect to action

Next steps backed by past cases and evidence

Sensor, waveform, measurement and image data are all converted to time-series and analyzed by one engine — no need to build a new model for each data type.

Output

  • AI anomaly scores by unit and lot, flagging 'in-spec suspects'
  • Ranking of the parameters that separate good from bad
  • Evidence for lot dispositions and differences from a reference lot
  • Process and test history traced per barcode
  • Failure Pareto and distributions of causal parameters
  • Automatic result delivery, reports and model performance management
04How This Differs From Existing Systems

Your existing systems, unchanged — just the 'why' and 'what next' added on top.

Spec judgment on test equipment

Limitation

Checks only each parameter's limits; abnormal combinations that stay within spec pass

What DutchBoy Adds

Flags suspect units and lots with AI anomaly scores that consider hundreds of parameters together

MES · YMS

Limitation

Actual yield and history lookups

What DutchBoy Adds

Shows actual yield next to AI suspects and jumps straight to the problem lot or equipment

SPC

Limitation

Single-parameter control charts

What DutchBoy Adds

Recommends discriminating parameters, multivariate correlation, and PASS/FAIL distribution comparison

History lookups

Limitation

Split across systems and traced by hand

What DutchBoy Adds

One barcode reveals the full process and test route, down to the measurements

Manual data handling

Limitation

Missing or late files are noticed too late

What DutchBoy Adds

Load-integrity checks and automatic result delivery

05Application Scenarios

One approach, different problems — solved in every process.

Shipping

Screening suspect lots before shipment

  1. 1Detect a rise in AI suspects among lots that passed spec
  2. 2Compare key parameters with a reference lot
  3. 3Record the disposition and its evidence
  4. 4Report and deliver results
Quality

Tracing a customer claim

  1. 1Enter the barcode
  2. 2Review the process and test route and any retests
  3. 3Check other units built under the same conditions
  4. 4Determine the scope of impact
Ramp-up

Finding failure causes for a new model

  1. 1Failure-type Pareto
  2. 2Compare distributions of causal parameters
  3. 3Link to process conditions
  4. 4Verify after adjusting conditions
06How the 6 Modules Work Together

Only the modules you need, added on top of what you have — Core to start, automated reporting and conversational AI as Add-ons. Every module running on the plant data Weave designs.

Agent · Query & reasoning
Visualize · Evidence & reporting
Sense
Recall
Watch
Weave · Factory data design

Your existing systems — unchanged

Test equipmentEquipment PLCsMESQMSCustomer quality systems

Weave

Data design

Loads test files and MES history, organizes parameters and manages load integrity

Core

Sense

Detection & RCA

AI anomaly scores, suspect unit and lot detection, causal parameters for failures, model performance management

Core

Recall

Recall & response

Past actions by equipment and defect type, plus shared notes from the floor

Core

Watch

Integration & alarms

MES completion checks, customer result delivery, pipeline and anomaly-rate alerts

Core

Visualize

Evidence & reporting

Barcode history tracing, distribution and correlation exploration, lot reports

Add-on

Agent

Query & reasoning

'Why did suspect lots increase this week?' → answers with evidence

Add-on

Why an LLM alone isn't enough — Time series from just 150 sensors add up to millions of numbers (about 5.4 million in our technology video example), far more than an LLM can read directly. DutchBoy's manufacturing-specific tools condense them into a few lines of key evidence — severity, pattern, scope of impact and past cases — and hand them to the Agent.

Could this fit your process?

From feasibility to expected impact, the AIBIZ team will walk you through it directly.

Book a Consultation
DutchBoy DEVICE | AIBIZ