Knowledge Work Closes Series C With ¥4.8 Billion in Bank Debt to Build Sales AI Agents
Tokyo-based Knowledge Work has completed its Series C round after raising ¥4.8 billion in debt financing, including credit lines, in a second close announced on October 23 through PR TIMES. Combined with ¥3.5 billion raised through a third-party share allotment in the first close, the round brings the company's cumulative equity funding to ¥9.74 billion and its cumulative debt funding, including credit lines, to ¥6.49 billion.
The lenders are Mizuho Bank, Sumitomo Mitsui Banking Corporation, MUFG Bank, Sumitomo Mitsui Trust Bank, Shoko Chukin Bank, the Shizuoka Bank and the Bank of Fukuoka. Knowledge Work, led by chief executive Koji Asano, says it will put the money into strategic investments that strengthen how it builds and runs sales AI agents for customers.
Founded in 2020, the company has shifted its business model from software-as-a-service to what it calls an AI company. It now sells three lines: a sales AX consulting series that builds and operates AI agents tailored to a client's sales work, a sales AI product series that quickly delivers agents for standard sales tasks, and a sales AI agent OS series that serves as the foundation for more accurate agents. Customers include NTT Docomo Business, Mizuho Bank and Nissin Foods, mostly large companies deploying it to thousands of staff.
What it means in practice
The financing mix is the signal. Using bank loans rather than another equity round lets a startup avoid further dilution for existing shareholders, and the presence of Japan's three megabanks alongside regional lenders suggests banks now see recurring enterprise AI revenue as something they can lend against. Sales is a natural target in Japan, where companies face labour shortages and a heavy load of manual reporting, proposal writing and follow-up. The deal fits a wider corporate push, from NTT Docomo Business spreading more than 200 AI assistants across its units to Sumitomo Life planning AI-tailored insurance proposals for its sales staff.
The open question is how much of the promise shows up in results. Knowledge Work has not published revenue or productivity data from its agent deployments, and sales agents depend heavily on well-organised customer data and on salespeople trusting what the system suggests.
What to watch: whether Knowledge Work discloses concrete gains at large customers, how widely its agent OS spreads beyond consulting projects, and whether more Japanese AI startups choose bank debt over new equity.