Tokyo-based enterprise automation startup Yoom has closed its first external funding round, raising ¥700 million ($7 million) to accelerate a platform that pairs AI agents with rule-based automation for Japanese enterprises. The round announced on September 15, 2026 was led by DUAL BRIDGE CAPITAL, with Mizuho Capital, Shizuoka Capital, Harizzly and Canon Marketing Japan's CVC fund joining as backers.
Building AI staff for Japanese teams
Yoom, founded in June 2022 and headquartered in Chiyoda ward's WeWork Kanda Square, positions its no-code platform as a way to build "AI staff" that colleagues can co-work with. Users compose workflows that mix large language model reasoning with deterministic rules for data entry, document creation, email, spreadsheet operations and SaaS API calls. As of the raise, the platform counts more than 70,000 registered users executing over 5 million monthly tasks.
Why the mix of AI and rules matters
Chief executive Shun Hatozaki argues that pure agent frameworks struggle with the certainty enterprises expect from back-office workflows, while pure RPA cannot handle unstructured inputs. Yoom's approach lets business teams choose which parts of a process are governed by rules and which are decided by an AI agent, keeping compliance-heavy steps deterministic while letting reasoning models take the messier tasks such as email triage or document normalisation.
Where the money goes
Yoom said it will spend the raise on product development, deeper system integrations and hiring across engineering and go-to-market functions. Canon Marketing Japan's Future Investment CVC Fund — operated by Global Brain — is expected to help the startup channel through Canon's enterprise customer base, joining a broader wave of Japanese corporate venture pouring into workflow AI.
The raise arrives amid record global AI-automation demand documented in Perplexity's on-device push, Salesforce's Koa CRM reasoning model with NVIDIA and Sakana AI's Fugu Ultra v2.0.
Reporting based on coverage from Tech Startups and PR TIMES.