Industrial AI Vision Inspection in Malaysia with 25 Years of Industrial Expertise

PixeVision is a Malaysian industrial AI company, part of the ASPL Group. We build two AI engines on one platform: See.AI™ for AI vision, covering defect inspection, counting, OCR and smart CCTV, and EKE™ for enterprise LLMs, turning documents and data into answers, reports and dashboards.

We start with a consultation to understand your workflows, pain points and the data available, then prioritise the use cases where AI delivers the highest impact and return.

Before full deployment, we run a proof of concept (POC) to show the solution works on your actual products or data, measured against success criteria agreed with you.

Yes, in many cases. We have added AI to existing CCTV cameras, existing X-ray machines and existing AOI equipment, without replacing the hardware.

Where new cameras are needed, we source them, integrate them with our software and deliver everything as one complete solution, so you deal with a single provider.

We connect to your line and systems through PLC signals, APIs and connectors, keeping disruption to a minimum.

Accuracy depends on the application, so we validate it during the POC against success criteria agreed with you.

Once deployed, our solutions come with a Service Level Agreement (SLA), under which we guarantee the agreed AI accuracy for as long as the SLA contract is in effect.

We also design for real-world conditions. Uncertain results can be flagged for human review, and on production lines a reject can stop the conveyor for an operator to check. People stay in control of the final decision.

By default, yes. Your inspection data and records are kept on your own local server, and many of our systems run fully on-site on a local industrial PC.

If your requirements call for it, we can also deploy on a cloud server instead.

Where remote access is needed, it goes through a single port opened and controlled by your IT team. In every setup, we adhere to your organisation's data and network protection requirements.

Access to the relevant data or sample products, time with your process experts, and IT support for integration.

We will also work with your team to define the detection or event criteria: what counts as a defect, a pass or fail, or an event that should trigger an alert, ideally with reference samples or examples.

No AI expertise is required from your side. We handle model development, deployment and support.

Every deployment includes handover training, documentation and ongoing support so your team can use the system confidently.

For wider upskilling, we also run HRDF claimable AI training, from 1-day AI awareness sessions to hands-on technical programmes.

Every project is costed to its scope. To prepare an accurate proposal, these are the typical details we will ask for:

• Use case and goal: what the AI needs to detect, count, read or answer, and what success looks like for you
• Scale: number of inspection points, production lines, cameras or sites
• Production details: line speed or throughput, and the number of product types or SKUs
• Existing equipment: cameras, CCTV, machines or servers we may be able to reuse
• Integration: PLC, reject mechanism, alarms, or links to systems such as MES and ERP
• Deployment preference: fully on-site on a local server or industrial PC, or with remote access
• For EKE™ projects: your data sources, such as documents and databases, and the expected number of users
• Samples: product samples, images or data available for the proof of concept
• Support needs: training, maintenance and ongoing support expectations

Do not worry if you do not have every answer yet. We work through these together during the consultation, and the POC lets you validate results before committing to full deployment.