AI Visual Inspection for PDI

AI Overview Summary

Ombrulla successfully transformed its Pre-Despatch Inspection (PDI) for industrial vehicles using an AI Visual Inspection system powered by 10 strategically placed cameras. The solution validates 78 inspection checkpoints in seconds - a process that previously required more than 30 minutes per vehicle - while improving consistency, auditability, and dispatch throughput.

• •Process: Pre-Despatch Inspection (PDI) for industrial vehicles
• •Cameras deployed: 10 cameras positioned for full-coverage capture
• •Checkpoints inspected: 78 checkpoints in seconds
• •Prior cycle time: More than 30 minutes per vehicle
• •Current cycle time: Seconds (one-click capture + automated evaluation)
• •Primary outcomes: Higher throughput, consistent quality gate, digital evidence and traceability

Business Context

End-of-line Pre-Despatch Inspection (PDI) - the final quality gate before a vehicle leaves the facility.

• •Customer trust and brand protection: Preventing the dispatch of visible defects or missing fitments.
• •Lower cost of quality: Reduced rework, warranty exposure, and delivery disruptions.
• •Throughput improvement: Removing PDI as a dispatch bottleneck.
• •Audit readiness: Time-stamped inspection evidence tied to vehicle identifiers.

The Challenge

Ombrulla's PDI relied on manual visual checks and checklist-based sign-off. The approach was dependable but constrained by cycle time and variability.

Cycle Time

Manual PDI required more than 30 minutes per vehicle, creating a queue at dispatch.

Consistency Risk

Results varied by inspector experience, fatigue, and shift conditions.

Evidence Gap

Limited visual proof for audits, customer disputes, or internal root-cause analysis.

Scalability Constraints

Volume growth required additional inspection labor rather than improving flow.

Late-stage correction cost: Defects discovered at PDI are more expensive to resolve than upstream fixes.

The Solution

Ombrulla implemented an AI Visual Inspection solution for PDI using 10 cameras positioned at strategic locations around the inspection bay. With a one-click capture, the system ingests synchronized images and automatically evaluates 78 checkpoints in seconds.

What changed operationally:

• •Inspection became a standardised, repeatable digital quality gate rather than a manual bottleneck.
• •Inspectors shifted from repetitive scanning to exception validation and corrective action feedback.
• •Inspection outputs became auditable records, improving governance and dispute resolution.

How It Works

1. Camera Strategy

10 cameras placed to maximise coverage across front, rear, side profiles, and cabin-area views to reduce blind spots.

2. One-click Capture

Vehicle enters PDI bay, operator triggers capture, and system collects images from all 10 cameras simultaneously.

3. AI Evaluation

AI models evaluate 78 checkpoints in seconds (presence/absence, alignment, defects, labelling compliance).

4. Instant Feedback

Dashboard presents pass/fail per checkpoint, highlights areas of interest, and logs evidence against vehicle ID.

5. Human Value

Inspectors focus on confirming exceptions, providing feedback, and supporting root-cause analysis.

Measurable Impact

Core performance gains from the implementation.

• •PDI cycle time reduced from more than 30 minutes to seconds per vehicle.
• •78 checkpoints are inspected automatically per inspection cycle.
• •Higher consistency across shifts with standardized evaluation criteria.
• •Digital traceability: time-stamped evidence retained per vehicle for audits and dispute handling.

Checkpoint Coverage

The 78 checkpoints are managed as a controlled checklist and can be summarized into leadership-friendly groups:

Safety & Compliance

Lights, reflectors, mandatory markings, warning labels.

Fit & Finish

Panel alignment and surface anomalies visible at dispatch.

Assembly Completeness

Presence of required parts, caps, guards, fasteners.

Branding & ID

Decals, VIN/ID plates, serial label readability.

Dispatch readiness: visible damage risks, accessories, packaging conditions (where applicable).

Implementation Approach

Delivery phases used to reach production readiness:

• •PDI mapping: Checkpoint definitions, defect taxonomy, acceptance criteria.
• •Camera layout and lighting validation: Coverage and reflection control.
• •Model configuration per checkpoint: Detection/verification logic and thresholds.
• •Pilot run: Parallel manual + AI inspection to benchmark performance and tune exceptions.
• •Go-live: SOP updates, training, audit logs, and exception workflows.

KPIs to Track Post Go-Live

• •PDI cycle time per vehicle (trend by shift and line).
• •First-pass yield at PDI.
• •Defect escape rate (post-dispatch issues attributable to visual checkpoints).
• •Rework hours prevented and rework cost avoided.
• •Dispatch adherence (on-time dispatch improvement).
• •Audit retrieval time (time to produce inspection evidence).