Infofla Offers On-Premise 'Record & Replay' AI Automation for Enhanced Data Security

As businesses increasingly explore AI for task automation, a South Korean startup, Infofla, is differentiating itself with an on-premise "Record & Replay" solution. This approach allows AI to learn and automate tasks by observing user demonstrations, eliminating the need for intricate prompt writing or coding. This is particularly beneficial for organizations handling sensitive data within their internal systems, such as ERP or groupware, where cloud-based solutions pose security risks.

The "Record & Replay" method, also known as demonstration-based automation, involves users performing a task once on their screen. The AI then learns this workflow and can replicate it autonomously. This contrasts with older macro or RPA tools that merely recorded and replayed clicks and keystrokes; "Record & Replay" aims to understand the 'intent' behind the workflow.

Major tech players like OpenAI with its Codex agent and Anthropic with its Claude desktop app have recently integrated similar cloud-based "Record & Replay" functionalities. This rapid development suggests a significant shift in the AI automation market towards this intuitive, demonstration-driven approach. However, these cloud solutions raise concerns about data privacy and security, as information is transmitted to external servers for processing.

Infofla's solution, named "Selto," operates entirely on-premise, ensuring that sensitive company data and system environments remain within the organization's network. The company claims that Selto learns and adapts to specific organizational workflows over time, becoming a more optimized and accurate automation tool. Infofla CEO Choi In-mook highlighted that their on-premise "Record & Replay" capability, developed before similar features emerged from global tech giants, holds significant value for sectors where security is paramount, such as public institutions and financial services.