AI platform company Dtonic said on Wednesday it has registered two patents that reduce technical bottlenecks that arise during corporate AI transformation, or AX.

The patents cover automatic workflow generation based on ontology and user intent recognition, and parallel query processing that links heterogeneous accelerators based on a hierarchical spatial index. The two technologies focus on addressing limitations in applying AI models in real business environments: complex design and slow processing speed.

The first patent (No. 10-2955453) is a technology in which AI itself understands and structures complex work processes in business settings. When a user inputs the goal to be achieved, it designs data processing steps in real time through an ontology-based knowledge structure.

The second patent (No. 10-2955452) moves beyond conventional CPU-centered methods and boosts computing speed by organically combining various hardware accelerators such as GPUs and FPGAs.

Dtonic CEO Yong-joo Jeon (전용주) said, "True AX depends not only on having an excellent model, but on how easily and quickly it can be run in the field." He added, "Based on these patented technologies that enhance both AI 'planning intelligence' and 'execution capability,' Dtonic will provide differentiated business value to customers."