Knowledge Work, a Japanese startup developing a sales-specialized AI platform called the "Sales AI A...
The AMW Read
New strategic capital and lineup of large Japanese corporates significantly update the agent player map in Japan, despite the modest round size being below the cross.§D threshold.
Knowledge Work, a Japanese startup developing a sales-specialized AI platform called the "Sales AI Agent OS," announced on August 4 the first close of its Series C round, raising ¥3.5 billion (approximately $24 million). New investors include Ricoh Japan, Canon Marketing Japan MIRAI Fund, NTT Docomo Ventures, Sumitomo Mitsui Banking Corporation, Japan Post Bank Asset Management, Mitsubishi UFJ Innovation Partners, Sumitomo Mitsui Trust Bank, Fukoku CVC Fund, Dentsu Ventures SGP Fund, and Hakuhodo DY Ventures. Existing investors Glovis Capital Partners, DNX Ventures, WiL, Salesforce Ventures, and For Startups Capital also participated.
This round brings the company's total funding to a level that underscores the growing market for verticalized AI agents in Japan. The Sales AI Agent OS is built on four layers: an ontology layer that connects sales data in business context, a strategy layer that feeds company-specific sales strategies into the AI, an agent layer for implementation, management, and analysis, and an interface layer for user UX. This architecture prioritizes task-specialized agents over general-purpose AI, a design choice that has attracted investment from major corporate venture arms in logistics, media, and finance.
The capital will be used to accelerate strategic partnerships with leading companies across industries, merging the investors' client bases and solutions with Knowledge Work's AI technology and sales enablement expertise. For enterprise buyers, the complexity and deployment of sales use cases makes bespoke agent foundations more convincing than generic copilots. The investor lineup, mixed with banks, insurers, trading houses, and media majors, suggests that Japanese incumbents are preparing for adoption of AI agents, funding infrastructure rather than just buying point solutions. As public markets test software valuations, the architecture of task-specific agent OS could make or break the profitability of the application layer in 2026.