Datumo Raises $15.5M to Compete with Scale AI - Salesforce Backed

Generative AI Safety Concerns and Datumo's $15.5 Million Funding

A recent McKinsey report indicates that a majority of organizations feel unprepared to implement generative AI in a secure and responsible manner. A key challenge identified is explainability – the ability to understand the reasoning behind AI’s decisions.

While 40% of those surveyed consider this a substantial risk, only 17% are currently taking active steps to address it.

Datumo: Building Safer AI Solutions

Datumo, a company originating in Seoul as an AI data labeling firm, is now focused on assisting businesses in developing more secure AI systems.

They provide tools and data designed for testing, monitoring, and refining AI models, all without requiring specialized technical skills.

On Monday, Datumo secured $15.5 million in funding, bringing their total raised to approximately $28 million. Investors include Salesforce Ventures, KB Investment, ACVC Partners, and SBI Investment.

From Data Labeling to Comprehensive AI Evaluation

David Kim, CEO of Datumo and a former AI researcher at Korea’s Agency for Defense Development, initially sought a solution to the time-intensive process of data labeling.

His solution was a reward-based application enabling individuals to label data during their free time and earn compensation.

The concept was validated at a startup competition hosted by KAIST (Korea Advanced Institute of Science and Technology).

Datumo, previously known as SelectStar, was co-founded by Kim and five KAIST alumni in 2018.

Prior to the app’s completion, the startup secured tens of thousands of dollars in pre-contract sales through customer discovery, primarily from businesses and startups connected to KAIST alumni.

Within its first year, Datumo exceeded $1 million in revenue and secured several significant contracts.

Expanding Services and Client Base

Today, Datumo serves prominent Korean companies including Samsung, Samsung SDS, LG Electronics, LG CNS, Hyundai, Naver, and SK Telecom.

Initially, clients requested assistance beyond basic data labeling, asking Datumo to assess and compare AI model outputs.

Michael Hwang, Datumo’s co-founder, explained to TechCrunch that this led to the realization they were already performing AI model evaluation without explicitly recognizing it.

Datumo subsequently focused on this area and released Korea’s first benchmark dataset dedicated to AI trust and safety.

“We began with data annotation, then broadened our scope to include pretraining datasets and evaluation as the LLM ecosystem developed,” Kim stated to TechCrunch.

A Growing Market and Competitive Landscape

Meta’s substantial investment – approximately $14.3 billion – in data-labeling company Scale AI underscores the importance of this market.

Shortly after this acquisition, OpenAI, a competitor to Meta, discontinued its use of Scale AI’s services, indicating intensifying competition for AI training data.

Datumo shares similarities with Scale AI in providing pretraining datasets and with Galileo and Arize AI in AI evaluation and monitoring.

However, it distinguishes itself through its proprietary licensed datasets, particularly data extracted from published books, which the company asserts offers valuable structured human reasoning but presents significant cleaning challenges.

Datumo Eval: A No-Code Evaluation Platform

Datumo also offers a comprehensive evaluation platform, Datumo Eval, which automatically generates test data and evaluations to identify potentially unsafe, biased, or inaccurate responses without requiring manual scripting.

This signature product is a no-code evaluation tool specifically designed for non-developers working in policy, trust and safety, and compliance roles.

Securing Investment and Future Plans

Kim attributed attracting investors like Salesforce Ventures to a fireside chat hosted by the startup with Andrew Ng, founder of DeepLearning.AI, at a South Korean event.

Sharing the session on LinkedIn brought it to the attention of Salesforce Ventures, leading to meetings and a soft commitment for funding.

The entire funding process took approximately eight months, according to Hwang.

The new funding will be allocated to accelerating R&D, particularly in automated evaluation tools for enterprise AI, and expanding global operations across South Korea, Japan, and the U.S.

Datumo, currently employing 150 individuals in Seoul, also established a Silicon Valley presence in March.

We’re always looking to evolve, and by providing some insight into your perspective and feedback into TechCrunch and our coverage and events, you can help us! Fill out this survey to let us know how we’re doing and get the chance to win a prize in return!