Three enterprise AI companies connecting in-house knowledge [사진=구글 제미나이로 생성]
[디지털데일리 이안나기자] One employee spends an average of 1.8 hours a day searching for information within the company. Search through Slack, open and close Drive folders, and enter keywords repeatedly into the search box. Based on a McKinsey survey, in a year, 450 hours, or about 8 weeks’ worth of work time, are lost through ‘file search’. Even after introducing dozens of Software as a Service (SaaS) tools, information is still trapped within each system. It’s a data silo problem.
According to the industry on the 22nd, companies that aim to solve this problem with AI are rapidly emerging. While the generative AI craze, represented by ChatGPT, is centered around Big Tech, Glean, Sana, and Guru, a trio of emerging powerhouses that have built their own ecosystem in the AI search and automation platform market specialized in corporate internal data, are attracting attention. There is one thing in common. AI connects and retrieves information scattered within a company and even executes it.
Among these, the one showing the fastest growth is Glynn, based in Silicon Valley, USA. In December of last year, annual recurring revenue (ARR) officially surpassed $200 million. It has only been 9 months since it surpassed $100 million, which is 2 to 3 times faster than typical business-to-business (B2B) SaaS companies. The current corporate value is estimated at $7.2 billion (approximately 9.7 trillion won).
Glynn CEO Arvind Jain recently said on his official blog, “Simply introducing AI tools into Google or Microsoft (MS) product lines cannot guarantee value creation or widespread adoption,” and added, “What is needed is a tool that consistently provides the right answers so employees can proceed with their work without having to search for information.”
In fact, according to Gartner’s 2025 Global Labor Market Survey, 72% of IT leaders driving Microsoft Co-Pilot integration said that users are having difficulty integrating Co-Pilot into their daily work. More than half complained that the value did not meet expectations and that participation was low.
Glynn’s technical core is the ‘Enterprise Graph‘. It connects hundreds of business tools such as Slack, Jira, Salesforce, and Google Workspace to structure a relationship network that shows who handled what information, when, and in what context. The industry evaluates it as an enterprise implementation of the augmented graph search and generation (GraphRAG) technology, which has recently attracted attention.
Unlike the existing RAG, which is limited to simple keyword matching or vector search, this method improves context accuracy by learning relationships between documents and information flow within an organization in a graph structure. Recently, the ‘autonomous agent’ function has been brought to the fore. It works by proposing relevant documents before a meeting or analyzing Jira ticket status to warn of possible deadline delays. It directly performs actions such as updating Salesforce data, creating Jira tickets, and posting Slack reports.
Sana’s approach is different. The Sweden-based startup was acquired by Workday for $1.1 billion last year. Workday is a platform with 75 million corporate users in the HR, finance, and operations areas around the world. Sana’s AI features operate within the existing Workday environment without a separate subscription or login.
Joel Hellermark, founder and CEO of Sana, said at the time of the acquisition, “By joining Workday, we will be able to significantly advance our vision,” and added, “Together, we will change the way companies access knowledge, automate repetitive tasks, and learn with agentic AI.” While Glynn relies on top-down sales targeting CIOs and CISOs, Sana uses its existing user base of 75 million to directly penetrate end users.
Guru, founded in the United States in 2013, promotes ‘Verified Knowledge.’ It has a structure that first refers to the knowledge base verified by in-house experts before the AI generates an answer, and its customers include Nike and Spotify. It is securing markets centered on the financial, medical, and legal industries, where hallucination threats are fatal, and is positioning to target corporate demand that wants controllable AI rather than complete autonomy.
Even in Korea, interest in the technology stacks of these companies is increasing among some engineers and IT managers. Domestic small and medium-sized companies, such as Allgunize, which entered the global market with Agentic RAG, and Upstage, which leads the domestic AI consortium, are also releasing enterprise solutions in a similar direction, drawing attention to whether a market will be formed. However, internal data security issues and interconnection issues with existing legacy systems are considered prerequisites before actual introduction in Korea.
Copyright ⓒ Digital Daily. Reproduction and redistribution prohibited.
