Welcome, Datumo!
AI trust and safety at scale.
August 11, 2025
- Founders: David Kim, Michael Hwang, Hawk Shin, Yuntaek Lee, Haseung Jeon, Muni Jeon
- Sector: Generative AI
- Location: Seoul, South Korea
The Opportunity
Generative AI is advancing at breakneck speed, with new services launching daily on top of foundational models. But for every company building with LLMs, one critical challenge remains: ensuring the reliability and safety of AI-generated outputs.
While LLMs can produce highly sophisticated results, they also carry well-documented risks:
- Hallucinations and factual inaccuracies
- Biased or ethically problematic responses
- Unintended legal, reputational, or societal consequences
Without a systematic and automated evaluation process, these risks can erode user trust, damage brand credibility, and slow adoption.
Enter, Datumo.
The Solution
Datumo’s Eval platform delivers automated, end-to-end trustworthiness evaluation for generative AI services. Eval enables early detection and mitigation of risks before deployment — bringing the rigor of structured quality and safety assurance to the fast-moving world of AI.
Key features include:
- Customizable Evaluation Criteria: Datumo Eval automatically generates optimized evaluation metrics and questions tailored to a user’s specific needs and preferences.
- Large-Scale Question Generation: Datumo Eval is capable of generating over 1 million questions to comprehensively cover diverse user utterance patterns.
- Automated AI Response Evaluation & Analysis: Datumo Eval measures LLM performance through a mix of objective benchmarks, human review, and model-judged comparisons—combining quantitative scoring on datasets like MMLU or GSM8K, qualitative assessments of fluency, coherence, and usefulness, and “LLM-as-a-judge” evaluations to determine the better response.
- Red Teaming Capabilities: Datumo Eval actively probes for unsafe or harmful outputs to detect and block vulnerabilities in advance.
Why We’re Backing Datumo
Datumo is angling to become the leading company in the AI trustworthiness evaluation domain — verifying and enhancing the quality and safety of generative AI–powered services before deployment.
From Datumo’s launch in 2018, the team focused its business on AI and data, long before generative AI became widespread. Founder David Kim and his team have garnered significant interest across the Korean market. Datumo has rapidly gained market share among large Korean corporations (e.g., Samsung, SK Group, Shinhan Bank, Hana group) and diversified their business model to meet evolving enterprise needs as AI technology has matured. Conversations with customers and stakeholders consistently show that Datumo is seen as a mission-critical tool in their ongoing AI efforts.Asia.
In an era when AI safety concerns are the largest hurdle to widespread adoption, Datumo provides a generative AI evaluation solution that enables organizations to deploy AI-powered services across any domain at scale confidently.
What’s Ahead
The AI data infrastructure and LLM evaluation space are drawing significant global attention — and Datumo is at the forefront. We see strong potential for Eval’s international expansion, particularly across Korea, Japan, and the U.S., as demand for trusted AI deployment continues to rise.
We’re proud to be investing in Datumo’s $15.5M Series B alongside KB Investment, Shinhan Venture Investment, Kiwoom Investment, SBI Investment, Samsung Securities, Moorim Capital, Infobank Partners, and ACVC Partner, and excited to support their journey to define the future of AI trust and safety at scale.
Welcome to Salesforce Ventures, Datumo!
