Powertasks: “Our Agent Finds What Even Experts Miss, Pursuing It to the End”
Powertasks: “Our Agent Finds What Even Experts Miss, Pursuing It to the End”
Posted October. 30, 2025 14:14,
Updated October. 30, 2025 14:19
- General-purpose AI like ChatGPT is useful but faces limitations in accuracy and
context, especially in specialized fields such as finance and investment.
- Powertasks, led by CEO Chang-hoon Moon, is developing Agentic-AI technology
called Outcode to address these issues by providing structured,
source-grounded research and analysis for professionals.
- The company aims to become a leader in AI-driven, evidence-based research
platforms for specialized industries worldwide.
Diverse general-purpose artificial intelligence (AI) services like ChatGPT, Gemini, and Claude have been released, shaking up daily life. They are useful for answering questions and generating ideas. However, problems like plausible falsehoods (hallucinations) and a lack of context have emerged. While this may be acceptable in daily life, it's a different story in specialized industries like investment and finance. A single, minor error can lead to enormous losses.
As a result, the industry is turning its attention to Vertical-AI, which has a deep understanding of specific fields, and 'Agentic-AI,' which autonomously performs complex tasks. Vertical-AI is noted for its advantages, including high accuracy, lower costs, and enhanced regulatory compliance, complementing the limitations of general-purpose large language models.
Market analysis firm Technavio identified language models optimized for specific industries and multi-agent systems as key AI market trends post-2025. This indicates growing demand for AI solutions equipped with deep industry knowledge, suggesting the market's required value from AI has shifted from novelty to practical utility.
AI technology startup Powertasks is also in the midst of building an Agentic-AI service aligned with this changing industry trend. It is preparing to transition from a no-code-based work automation platform to a research agent platform for professionals. What are Powertasks' AI service strategy and goals? We met with Powertasks CEO Chang-hoon Moon to hear his story.
"We Find It to the End": Building an AI Like Sherlock Holmes
CEO Moon founded Powertasks after discovering the potential of AI tools to support professional research and investment analysis. The company was started with the conviction that AI could expand experts' capabilities by assisting with vast information exploration and sophisticated data structuring. His professional industry experience, spanning from engineering to management consulting and investment firms, became the foundation for the startup. He subsequently developed ‘Outcode’, an Agentic-AI service for professionals.
Outcode focuses not on speeding up simple tasks, but on securing analytical capabilities for experts. The focus is on value creation rather than cost reduction. Moon’s philosophy is that if AI can resolve the cognitive load and anxiety over incompleteness that come from analyzing massive documents, professionals can immerse themselves in insight and judgment.
Powertasks applied its 'Traverse Chain of Reasoning' technology to Outcode. CEO Moon likens this technology to the fictional detective Sherlock Holmes, due to how it collects, analyzes, and organizes all evidence related to solving a case "to the end." Furthermore, mindful of issues plaguing AI like context loss, bias, and hallucinations, the company has worked to ensure output reliability through multi-stage evaluation and evidence-tracking.
When a user uploads documents and inputs a query, Outcode understands the user's intent like a seasoned analyst and meticulously collects all related information. It then connects scattered clues from multiple documents to present a complete inference. It uncovers causal relationships and hidden information that are difficult to find with simple keyword searches or fragmented summaries.
What's noteworthy is Outcode's output. Most generative AI produces long, conversational sentences. In contrast, Outcode provides structured answers. This reflects the fact that professionals don't want long explanations, but clear, systematic information ready for immediate use. "Writing clearly and concisely is harder than writing at length," said Moon. "To conclude 'This is A' instead of giving a vague answer means the AI has perfectly reviewed vast amounts of data."
Outcode can analyze hundreds of pages of business reports or complex financial documents in minutes, regardless of language. It considers the professional user's workflow, offering concise, structured reports, clear evidence, and information source citations.
As it targets professionals, its standards for security and citation are strict. "All data is encrypted at a bank-level during transmission and storage, and we operate on a 'Zero Data Retention' policy, meaning client data is not stored on AI service company servers or used for training," Moon said. "We build agent trust by citing every analysis step and the exact source of the information in the output."
"Professional Research Will Be Handled by Agents"
Powertasks plans to lead work innovation with its Agentic-AI technology specialized for document processing. The core is to overcome the limitations of Large Language Models (LLMs) through technology. Moon explained that the problem isn't the model itself, but how it's utilized.
Various AI tools, such as document summarizers and dictation services, are on the market. But why do they fail to perform adequately in professional work? Moon diagnoses that existing AI tools fail to reflect the work experience and level of completion required in specialized fields. His solution is as follows: thoroughly collect and connect all information, then repeatedly infer from an expert's perspective to produce a structured result. This implies clear limitations to a standard AI (GPT) processing structure without agentic functions.
While people often perceive agents as replacements for simple repetitive tasks, Powertasks refutes this perception. Their stance is that a properly designed agent can perform diverse tasks like research, report generation, and memo writing, much like a human. As AI innovation accelerates and success cases emerge, CEO Moon revealed his ambition to lead the market with AI technology that will change the paradigm of professional research and analysis.
Goal: To Be No. 1 in 'Source-Grounded Research'
"Development of Agentic-AI specialized for specific industries is underway. We are building services in a way that Agentic-AI will replace or supplement complex tasks. Since this service is applicable to global industrial environments, we plan to actively pursue global market entry."
Powertasks successfully secured Pre-Series A funding, providing the momentum to advance its Agentic-AI technology and enter overseas markets. Although still in pre-launch testing, Moon explains that it is a source-based research agent platform technically on par with global companies.
Powertasks was also selected for the AWS Jungle Program, co-hosted by Seoul National University of Science and Technology (SeoulTech) and AWS, receiving technical support necessary for operating its AI service. Support also included various programs like professional investment mentoring and workshops. "Thanks to AWS's support, the completeness of Powertasks' service has improved," Moon said. "We expect it to be a great help in our future global market entry."
Powertasks' goal is to become the No. 1 company in the field of 'source-grounded research,' where information is analyzed based on clear provenance. This is an expression of confidence that it will set a new standard in the AI service market for professionals based on Outcode's technology. Furthermore, the company plans to expand the Outcode service into a platform that covers the entire workflow of specialized fields such as investment screening, legal due diligence, and M&A analysis.
By Hyung-seok Kang (redbk@itdonga.com)
* This article was written with support from Seoul National University of Science and Technology.
context, especially in specialized fields such as finance and investment.
- Powertasks, led by CEO Chang-hoon Moon, is developing Agentic-AI technology
called Outcode to address these issues by providing structured,
source-grounded research and analysis for professionals.
- The company aims to become a leader in AI-driven, evidence-based research
platforms for specialized industries worldwide.
Diverse general-purpose artificial intelligence (AI) services like ChatGPT, Gemini, and Claude have been released, shaking up daily life. They are useful for answering questions and generating ideas. However, problems like plausible falsehoods (hallucinations) and a lack of context have emerged. While this may be acceptable in daily life, it's a different story in specialized industries like investment and finance. A single, minor error can lead to enormous losses.
As a result, the industry is turning its attention to Vertical-AI, which has a deep understanding of specific fields, and 'Agentic-AI,' which autonomously performs complex tasks. Vertical-AI is noted for its advantages, including high accuracy, lower costs, and enhanced regulatory compliance, complementing the limitations of general-purpose large language models.
Market analysis firm Technavio identified language models optimized for specific industries and multi-agent systems as key AI market trends post-2025. This indicates growing demand for AI solutions equipped with deep industry knowledge, suggesting the market's required value from AI has shifted from novelty to practical utility.
AI technology startup Powertasks is also in the midst of building an Agentic-AI service aligned with this changing industry trend. It is preparing to transition from a no-code-based work automation platform to a research agent platform for professionals. What are Powertasks' AI service strategy and goals? We met with Powertasks CEO Chang-hoon Moon to hear his story.
Chang-hoon Moon, CEO of Powertasks / source=IT dongA
"We Find It to the End": Building an AI Like Sherlock Holmes
CEO Moon founded Powertasks after discovering the potential of AI tools to support professional research and investment analysis. The company was started with the conviction that AI could expand experts' capabilities by assisting with vast information exploration and sophisticated data structuring. His professional industry experience, spanning from engineering to management consulting and investment firms, became the foundation for the startup. He subsequently developed ‘Outcode’, an Agentic-AI service for professionals.
Outcode focuses not on speeding up simple tasks, but on securing analytical capabilities for experts. The focus is on value creation rather than cost reduction. Moon’s philosophy is that if AI can resolve the cognitive load and anxiety over incompleteness that come from analyzing massive documents, professionals can immerse themselves in insight and judgment.
Powertasks applied its 'Traverse Chain of Reasoning' technology to Outcode. CEO Moon likens this technology to the fictional detective Sherlock Holmes, due to how it collects, analyzes, and organizes all evidence related to solving a case "to the end." Furthermore, mindful of issues plaguing AI like context loss, bias, and hallucinations, the company has worked to ensure output reliability through multi-stage evaluation and evidence-tracking.
When a user uploads documents and inputs a query, Outcode understands the user's intent like a seasoned analyst and meticulously collects all related information. It then connects scattered clues from multiple documents to present a complete inference. It uncovers causal relationships and hidden information that are difficult to find with simple keyword searches or fragmented summaries.
What's noteworthy is Outcode's output. Most generative AI produces long, conversational sentences. In contrast, Outcode provides structured answers. This reflects the fact that professionals don't want long explanations, but clear, systematic information ready for immediate use. "Writing clearly and concisely is harder than writing at length," said Moon. "To conclude 'This is A' instead of giving a vague answer means the AI has perfectly reviewed vast amounts of data."
Outcode homepage / source=Outcode
Outcode can analyze hundreds of pages of business reports or complex financial documents in minutes, regardless of language. It considers the professional user's workflow, offering concise, structured reports, clear evidence, and information source citations.
As it targets professionals, its standards for security and citation are strict. "All data is encrypted at a bank-level during transmission and storage, and we operate on a 'Zero Data Retention' policy, meaning client data is not stored on AI service company servers or used for training," Moon said. "We build agent trust by citing every analysis step and the exact source of the information in the output."
"Professional Research Will Be Handled by Agents"
Powertasks plans to lead work innovation with its Agentic-AI technology specialized for document processing. The core is to overcome the limitations of Large Language Models (LLMs) through technology. Moon explained that the problem isn't the model itself, but how it's utilized.
Various AI tools, such as document summarizers and dictation services, are on the market. But why do they fail to perform adequately in professional work? Moon diagnoses that existing AI tools fail to reflect the work experience and level of completion required in specialized fields. His solution is as follows: thoroughly collect and connect all information, then repeatedly infer from an expert's perspective to produce a structured result. This implies clear limitations to a standard AI (GPT) processing structure without agentic functions.
Powertasks leads work innovation with Agentic-AI specialized in document processing / source=Powertasks
While people often perceive agents as replacements for simple repetitive tasks, Powertasks refutes this perception. Their stance is that a properly designed agent can perform diverse tasks like research, report generation, and memo writing, much like a human. As AI innovation accelerates and success cases emerge, CEO Moon revealed his ambition to lead the market with AI technology that will change the paradigm of professional research and analysis.
Goal: To Be No. 1 in 'Source-Grounded Research'
"Development of Agentic-AI specialized for specific industries is underway. We are building services in a way that Agentic-AI will replace or supplement complex tasks. Since this service is applicable to global industrial environments, we plan to actively pursue global market entry."
Powertasks successfully secured Pre-Series A funding, providing the momentum to advance its Agentic-AI technology and enter overseas markets. Although still in pre-launch testing, Moon explains that it is a source-based research agent platform technically on par with global companies.
Powertasks was also selected for the AWS Jungle Program, co-hosted by Seoul National University of Science and Technology (SeoulTech) and AWS, receiving technical support necessary for operating its AI service. Support also included various programs like professional investment mentoring and workshops. "Thanks to AWS's support, the completeness of Powertasks' service has improved," Moon said. "We expect it to be a great help in our future global market entry."
Powertasks' goal is to become the No. 1 company in the field of 'source-grounded research,' where information is analyzed based on clear provenance. This is an expression of confidence that it will set a new standard in the AI service market for professionals based on Outcode's technology. Furthermore, the company plans to expand the Outcode service into a platform that covers the entire workflow of specialized fields such as investment screening, legal due diligence, and M&A analysis.
By Hyung-seok Kang (redbk@itdonga.com)
* This article was written with support from Seoul National University of Science and Technology.
About Dong-A Ilbo