메모렌스 오퍼에이전트

Manual assembly and inspection tasks in manufacturing sites heavily rely on worker experience and visual inspection, and AI model retraining is delayed when work environment changes.

Existing Automated Optical Inspection (AOI) systems require strict lighting and camera settings, making rapid response difficult. Unstructured task support tools are inefficient because they require repetitive manual input. There is a need for technology that reduces result deviations and inefficiencies caused by varying levels of work proficiency.

Memorence's Operagents provides an interface for workers to directly teach AI to "learn instantly." When a worker corrects a mistake, the AI immediately learns and applies the correction. AI models can be instantly adjusted and improved on the line without a separate developer.

The key is the 'instant learning (AutoML)' function, which enables real-time adaptation through human-machine collaboration, rather than just a simple NO-CODE vision tool. On-site deployment is possible with only a standard PC and UVC camera, and comprehensive work support is provided in conjunction with the knowledge base of the Memorence Suite.

Global manufacturers (companies operating production lines for electronics, automobiles, etc.) and automation equipment companies (B2B) are the main customers. Demand is especially high in medium-to-large-sized manufacturing sites with many unstructured assembly and inspection tasks. The company is also entering the market through collaboration with AI technology partners (ODM/OEM) and industrial PC manufacturers.

It can be applied to the entire manufacturing industry (semiconductors, automobiles, home appliances, medical devices, etc.) and can be expanded to fields that require various vision inspections.

It can also be used as an educational tool, and can be applied to the automation of other industry domains by adding language and knowledge bases. It is an industry with high global demand, and technical certification (UI review, etc.) is more important than national regulations.

Technical media such as Digitimes introduced it as a "major advancement in the on-site application of AI," and it has proven its performance by reducing inspection error rates to less than 0.1% in a single customer case. At CES, it was praised for the innovation of human-in-the-loop learning.

In terms of completeness, it is at a proven level with pilot application cases in large overseas companies. On the other hand, data security, integration costs, and worker acceptance, which are uncertain factors when introducing AI, are evaluation issues. There is not much concern about overestimation, but many analysts say that the market growth rate depends on the willingness of companies to adopt it.