"Streamlining researchers' tedious tasks"...Pluto Labs envisions AI expansion

Categorizes research articles using citation network…the service supports hypothesis formulation and research design

Secured 2 paid contracts in South Korea in 3 months after its launch…adopted by Postech

Expanding into the pharmaceutical sector, analyzing competitiveness and providng an experimental agent service

While a researcher's competitiveness is in formulating new hypotheses, a significant portion of research time is spent searching for published research articles, categorizing prior studies, and verifying experimental methods.  

​Pluto Labs, an artificial intelligence (AI)-driven academic technology startup, aims to address this repetitive workload. The company aims to extend beyond searching articles to supporting research topic identification, hypothesis formulation, and experimental validation, thereby expanding researchers' execution capacity.  

​DailyPharm met with Junseon Yoo (38), CEO of Pluto Labs, to discuss the story behind the company's founding, the key differentiators of its research-support AI, and its strategy for expanding into the pharmaceutical and biotech sectors.

​The startup began with frustrations over research article search

Junseon Yoo, CEO of Pluto Labs

Pluto Labs originated from the personal frustration CEO Yoo experienced while directly engaging in academic research.

​Having majored in Industrial Engineering at POSTECH, Yoo participated in authoring research papers in mechanical engineering before expanding his research focus into AI and computer science. Throughout this process, Yoo realized that searching for relevant papers and data consumed as much time as conducting the actual experiments.  

Yoo began developing the service as a project in 2017 and incorporated the company in 2019.  

​CEO Yoo remarked, "When conducting research, I experienced much greater frustration while trying to find existing information created by others than in formulating hypotheses and running experiments. Pluto Labs is the company founded specifically to solve that problem.”

​Pluto Labs' core technology relies on analyzing citations and categorizing research areas.

​Papers within the same discipline repeatedly utilize similar keywords. Simple keyword searches often retrieve papers with entirely different methodologies or research questions. Pluto Labs overcomes this by analyzing how papers cite one another, clustering them according to specific research topics and methodological approaches.  

Pluto Labs' leading service, Scinapse, analyzes field-specific research trends so that users can search for literature and researchers. Scinapse AI is currently expanding into supporting the process from initial literature reviews to generating and evaluating research hypotheses.  

​Currently, its primary paying customers are universities and research institutions. The platform is utilized in the early planning phase to help researchers identify viable research topics tailored to their existing technologies and equipment, or to check unaddressed research gaps in existing literature.  

​CEO Yoo explained, "The service helps researchers identify which problems will yield the highest impact given their available technologies and equipment, and pinpoints remaining research gaps in a given field," and added, "It functions less like a simple paper search tool and more as a strategic roadmap generator for research direction."

​Research that can be performed immediately over "Nature-Level" hypotheses

​Global corporates are also entering the AI market for research hypotheses and planning. CEO Yoo cited DeepMind's Co-Scientist as an example attempting to solve problems similar to Pluto Labs.  

​However, Pluto Labs focuses less on generating the highest-level hypotheses for top-tier publication and more on presenting options that most researchers can realistically execute within their current resources and timeframe. 

Yoo stated, "No matter how exceptional a proposed hypothesis or research plan may be, if it requires unavailable equipment or experiments that take two to three years, it is unusable for a researcher who needs to publish a paper this year," and added, "Pluto Labs focuses on identifying the best possible choices researchers can execute under their present constraints." 

The company considers securing a leading position advantage in the research institution market as a competitive strategy. Universities and research institutes rarely adopt duplicate software solutions for the same function and instead use established platforms for a long period.

 

Yoo stated, "Institutions and universities are a conservative market that rarely replaces tools once adopted," and added, "Securing domestic paid contracts and overseas pilot projects early on creates a strong barrier to entry."

 

​Following the release of Scinapse AI in October 2025, Pluto Labs secured two paid contracts in South Korea within three months. The company is currently running over six pilot projects, including overseas trials. In South Korea, POSTECH officially adopted the service, while KAIST initiated a pilot implementation.

​Enterprise software targeting universities and research institutes typically involves long sales cycles aligned with institutional fiscal years and budget allocations. Yoo analyzed that securing paid contracts and pilot programs immediately post-launch was a significant milestone proving early market viability.

Yoo said, "Institutional research tools operate in a market where sales cycles are estimated at a minimum of one year, yet we secured domestic paid contracts within three months of launch," and added, "Considering this product targets academic institutions, expansion is progressing at a remarkably fast pace."

​From pipeline analysis to experimental AI agents

Pluto Labs is also actively expanding its university and research institute-centric services into the pharmaceutical and biotech industries.

​The company's AI agent currently under development analyzes information embedded in research literature, such as corporate affiliations, investigators, and funding sources, to map out competitive drug discovery landscapes. The goal is to identify companies researching specific targets or modalities and global pharma firms investing in related fields, thereby supporting business development (BD) strategy formulation.  

​Currently, the company is validating market demand through AI-generated report offerings, with plans to later deliver this functionality as a standalone software tool for direct researcher use.  

​Pluto Labs is also developing an agent designed to minimize trial and error in experimental execution. By analyzing existing literature and quantitative datasets, it evaluates hypothesis relevancy while identifying potential outliers or noise within experimental data.  

The focus of this AI agent is preventing scenarios where researchers belatedly discover flawed experimental conditions or missing variables, forcing redundant trials. The company aims to unveil this service officially this fall. 

Yoo noted, "Identifying a drug target does not automatically yield a therapeutic. One needs an experimental design to validate whether the target is truly viable, and the iterative trial cycles consume the most time," and added, "Ultimately, our goal is to build products that reduce repetitive experimental cycles and compress the overall R&D timeline."

​In the long run, the vision encompasses utilizing not only published literature but also unpublished internal datasets and negative and failed experimental logs accumulated within research institutions.

Yoo observed that current Large Language Models (LLMs) face inherent limitations in comprehending numerical data and domain-specific contexts. Even identical quantitative values hold vastly different implications depending on the research field and experimental setup. Nuanced language-centric models struggle to differentiate this sufficiently.

Yoo remarked, "A difference of 20 may be negligible in one discipline, but massive in another," and added, "Academic research centers around numbers and symbols, but current models struggle to guarantee sufficient intuition or reliability regarding numerical values."

​Yoo concluded by noting that "There are tasks that are important yet tedious, and tasks that are minor yet tedious," and added, "I want Pluto Labs to be remembered as a company that resolves these bottlenecks, expanding the scope of what researchers can transform from imagination into research works."