3 weeks · ¥25,000
Process Suitability Assessment
If you are not yet certain which of your processes are suited to automated handling, this assessment produces a ranked report with reasons — before any implementation begins.
See details
Manufacturing Quality Support · 12 weeks · ¥42,000
An additional layer of checking — without displacing the one already working
Image-based defect detection introduced as a supplement to your existing inspection, not a replacement for it. Twelve weeks including a measured trial period, with full retention of your manual check throughout.
What this engagement delivers
A second pair of eyes on your production line — one that is always there
By the end of twelve weeks, you will have image-based checking running alongside your existing inspection, with measured detection rates across each defect category relevant to your line. You will know exactly how the additional layer performs — not in general terms, but against your specific products and your specific defect types.
The manual inspection that was there before remains in place throughout. Nothing is removed during the trial, and the escalation path for uncertain items is defined jointly with your quality staff before the system goes live.
What you receive
-
Sample collection and classification across the defect categories relevant to your line
-
Threshold setting done jointly with your quality staff, not imposed by an external specification
-
Integration into your current reject handling workflow, without redesigning the line around the tool
-
Measured detection rates by defect category at the close of the trial period
-
Defined escalation path for items the system flags as uncertain, agreed with your team beforehand
The situation many production teams face
Manual inspection is reliable, but it is also finite
Experienced quality staff catch things that systems miss, and their judgment on ambiguous items is genuinely difficult to replicate. At the same time, a line running at production speed creates more decisions per hour than a human check can sustain without fatigue becoming a factor — particularly toward the end of a shift, or when throughput increases temporarily.
Image-based checking does not resolve that by replacing the human check. What it can do, in the right conditions, is catch a defined subset of defect types consistently, leaving the human inspector to focus attention where it is most needed. The question is whether those conditions exist on your line, and whether the detection rates achieved in a trial are sufficient to make the addition worthwhile.
This engagement answers that question with measured data from your line, not from a vendor's reference case on a different product type.
How the approach works
Built around your line, not a standard template
What we do, and when
Sample collection across defect categories
We collect samples with your quality staff, covering the defect types that matter on your specific line. The categories are defined by you, not by a default list from the tool provider.
Threshold setting with your team
Detection thresholds are set in conversation with the people who have been running the line — they understand which margins matter and which are acceptable variation. That knowledge shapes the configuration.
Trial against live production
The trial runs against real production output. Your manual inspection continues in parallel throughout, so there is no period during which the line depends on the automated check alone.
What we do not do
-
Remove or suspend your existing manual inspection at any point during the engagement or trial
-
Use detection rate figures from other production environments as a proxy for what your line will achieve
-
Recommend continuation where the trial data does not support it — the numbers are what they are
-
Restructure your reject handling or quality workflow beyond what is necessary to integrate the additional layer
What the twelve weeks look like
Three phases, one continuous engagement
The twelve weeks are structured as three phases. Each has a clear purpose and a defined deliverable before the next begins.
Weeks 1–4
Preparation
Sample collection across defect categories, threshold setting with your quality team, and integration planning with your existing reject handling. The manual check continues uninterrupted.
Your team's time
4–6 hours across sample sessions and configuration review
Weeks 5–10
Trial period
The image-based check runs alongside your manual inspection against live production. Data is collected by defect category. Escalation paths are in place before the trial opens.
Your team's time
2 hours per week for escalation review and data check-in
Weeks 11–12
Measurement and review
Detection rates are compiled by category and reviewed with your team. A written summary states what was measured and what it means for a decision about continuation.
Your team's time
2 hours for the findings review and decision discussion
Pricing
A fixed-scope engagement across twelve weeks
The Manufacturing Quality Support engagement is priced as a fixed fee covering all three phases — preparation, trial, and measurement. There are no additional charges for the sample collection sessions or for escalation reviews during the trial period.
Hardware and tooling costs, if any are required for integration, are scoped and agreed separately before the engagement begins. The fixed fee covers OR Path Land's time and expertise across the twelve weeks.
Investment
¥42,000
Fixed price · 12-week engagement
-
Sample collection across your specific defect categories
-
Threshold configuration done with your quality staff
-
Six-week live trial with escalation path defined in advance
-
Measured detection rates by defect category at trial close
-
Manual inspection retained in full throughout — no removal at any stage
How results are measured
Detection rates from your line, not from a reference environment
The measurement approach is agreed at the start of the engagement. What gets measured, and how, is determined by what is relevant to your defect categories — not by a default reporting template.
What the trial measures
-
Detection rate per defect category — the proportion of known defects of each type that the system identifies
-
Volume of items escalated as uncertain, and how those were resolved by your quality staff
-
False positive rate — items flagged by the system that your team confirmed were within tolerance
What the report does not claim
-
That the rates achieved during the trial will hold exactly in ongoing production — conditions can shift
-
That the system will detect defect types outside the categories covered in sample collection
-
A recommendation to continue if the trial data does not support the addition being genuinely useful
Our commitment
Your existing inspection does not go away
This is the thing we say plainly at the outset: the manual inspection that is already working on your line will not be removed, reduced, or suspended at any point during this engagement. The image-based check is introduced as an additional layer. That is its purpose, and it remains its purpose through the trial period and into any continuation you decide on.
If the trial data shows that the additional layer is not performing at levels that make it worthwhile on your line, we will say so. There is no commercial reason for us to recommend continuation where the evidence does not support it — OR Path Land has no commercial relationship with the tool provider and no stake in whether you proceed.
Suitable for
Production or packing lines where defect categories are defined, volume is sufficient to generate usable sample data, and your existing inspection is already functioning — this engagement adds to what is working, rather than replacing what is not.
Getting started
How the engagement begins
Before the twelve weeks start, there is a short scoping conversation to confirm that your line and your defect categories are a reasonable fit for this engagement.
01
Initial message
Send a brief description of your line, your defect categories, and your current inspection setup. No formal documentation is needed at this stage.
02
Scoping conversation
We confirm whether your line volume and defect categories are suited to the engagement, and agree the scope of sample collection before anything begins.
03
Preparation phase
Weeks one through four cover sample collection and configuration. Your quality team is involved in setting thresholds from the start.
04
Trial and findings
The trial runs for six weeks. At close, measured detection rates are reviewed and you decide whether and how to proceed from there.
Manufacturing Quality Support
Twelve weeks to measured data from your own line
If you are considering whether image-based checking could add something to your inspection process, the most reliable way to find out is to measure it against your actual production — with your existing inspection still in place throughout.
Get in touch about this engagementOther services
Related engagements worth considering
5 weeks · ¥33,000
Vendor Selection Advisory
Independent support for choosing between tool providers — structured trials using your material, contract review, and no commercial relationship with any vendor under consideration.
See details