CUBIG wins Gartner recognition for agentic AI tools

CUBIG has been recognised in two Gartner Emerging Tech reports on agentic AI, where the South Korean company was identified as both a Sample Provider and a Tech Innovator.

The recognition comes as Gartner's research highlights enterprise data and governance as central to moving AI projects from pilot schemes into production.

CUBIG, which recently launched in the UK, was cited in Gartner's report on prominent industry use cases for agentic AI and in a separate report on technology innovators in solution accelerators for agentic AI.

CUBIG develops Syntitan, which it describes as a data operating layer designed to help businesses prepare, validate and connect data for AI training, evaluation, execution and governance. Its products also include DTS and LLM Capsule, tools intended to handle restricted, scarce or low-quality data and to let AI systems work with sensitive operational data without moving it in raw form.

Gartner's findings reflect a broader shift in enterprise AI adoption, with attention moving beyond model development and computing infrastructure to the condition of corporate data and the controls around it. According to the analyst firm, technologies that provide enterprise context, semantic data layers, governance and operational orchestration are becoming more important for production-scale agentic AI.

One Gartner report said: "Vendors that integrate enterprise context deliver greater value than those that only focus on model optimization".

It also forecast: "By 2027, 40% of enterprises will demote or decommission autonomous AI agents due to governance gaps identified only after production incidents occur".

Data constraints

For many organisations, the issue is not access to information but whether data can be used safely, consistently and in line with regulatory or internal policy requirements. It may remain fragmented across systems, vary in quality, or sit outside the workflows where AI tools are expected to operate.

The challenge has become more pressing as businesses try to deploy autonomous or semi-autonomous AI systems in heavily regulated sectors. CUBIG works with clients in financial services, healthcare, telecommunications, manufacturing and the public sector.

Gartner's research also referred to a deployment of CUBIG's technology at a leading South Korean life insurer. According to Gartner, "The solution increased classification accuracy by 85.9% to 90%, a 200% improvement over rule-based systems' 50% to 60% accuracy", enabling the retention of high-utility behavioural insights beyond the six-month legal limit.

Founders' view

CUBIG's founders argue that the next phase of enterprise AI competition will depend less on selecting the newest model and more on making data usable in live operating environments. The company was founded in South Korea in 2021 and has been expanding internationally.

"Enterprise AI does not fail only because models are incapable. It often fails because the data state behind an AI run was never designed to be reused, traced or reproduced," said Bae Ho, Founder and Chief Executive Officer of CUBIG.

"Organisations have invested heavily in data platforms, governance frameworks and AI infrastructure. Yet there remains a critical gap between managed data and operational AI. We believe this recognition from Gartner reinforces the need for an AI-ready data operating layer that makes enterprise data usable, traceable and ready for AI workflows."

CUBIG says its software sits between raw enterprise data and AI execution. In practice, that means creating a layer that allows the data states used by AI systems to be checked, reproduced and governed, particularly where information is sensitive or cannot be freely transferred.

That focus on traceability has become increasingly important for companies using agentic AI in core business processes, where an error can have operational or regulatory consequences. Governance concerns have also grown as enterprises test systems that can make or influence decisions with less human intervention.

"Agentic AI may change how organisations operate, but autonomous systems are only as reliable as the data states they run on," said Ho.

"As enterprises move from experimentation to production, data readiness, traceability and operational trust become foundational requirements. AI-ready data is not just a preparation step. It is part of the operating layer for scalable enterprise AI."