NAVER D2SF has invested in Clone Labs, a South Korean AI startup developing technology that predicts user intent and decisions as AI agents become more common in everyday workflows. The investment was announced on May 18, 2026, by NAVER’s corporate venture capital arm, which described Clone Labs as a company building a “User Model” that learns computer and agent usage patterns. The startup is led by CEO Minchan Kim and is focused on reducing the friction users face when repeatedly guiding AI tools.
Addressing a New AI Bottleneck
The core problem Clone Labs is targeting is the growing burden of managing AI agents, especially as more users rely on multiple tools to complete complex tasks. Although AI agents can act quickly, they often pause for clarification, context, approval, or review, which can create a new layer of work for users. NAVER D2SF said it was drawn to the team’s ability to identify this emerging issue and test solutions rapidly.
Clone Labs’ User Model Technology
Clone Labs is developing an AI system designed to understand a user’s context, preferences, and intent without requiring constant manual direction. Its User Model is structured around three layers: recording computer usage patterns, building memory from decision-making context, and predicting the next action based on that accumulated data. The company also uses a confidence-based approach, allowing the system to act automatically when certainty is high while sending uncertain or sensitive decisions back to the user.
Products and Target Users
The company has introduced its first products, Clone Desktop and Clone Plugin, aimed at AI-native builders who regularly work across multiple agents. This target group is important because developers, researchers, and AI builders are often among the earliest adopters of agent-based workflows. By focusing first on users who already experience heavy agent-management demands, Clone Labs can test whether its model improves productivity in real-world use cases.
Research Foundation
Clone Labs was founded by undergraduate students from Seoul National University and has published seven research papers connected to AI agents. Its research background includes work involving computer-use agents, long-term memory systems, and privacy-preserving memory architecture, which are all relevant to systems that must learn from user behavior over time. Media coverage also noted research collaborations involving Stanford University and Carnegie Mellon University.
NAVER D2SF’s Investment Rationale
NAVER D2SF first identified Clone Labs through its 2025 Campus Startup Competition and invested roughly three months after the startup entered its incubation program in January 2026. The investment reflects NAVER D2SF’s broader interest in founders building around new challenges created by the AI shift, rather than only improving existing software categories. Sanghwan Yang, head of NAVER D2SF, said Clone Labs recognized the productivity bottleneck between humans and agents and is addressing it with research depth and execution speed.
The investment positions Clone Labs as part of a growing wave of startups trying to make AI agents more autonomous, useful, and less demanding for end users. Its approach is notable because it focuses not only on making agents faster, but also on reducing the human supervision that can limit their practical value. For NAVER D2SF, the deal reinforces a strategy of backing early teams that are building around the next operational challenges of the AI era.