Maisa AI, a startup focused on enterprise automation, has raised $25 million in seed funding to address what researchers say is a major issue for companies: the failure of generative AI pilots. A report from MIT’s NANDA initiative found that 95% of enterprise AI projects are currently unsuccessful.
The funding round was led by European venture capital firm Creandum, with participation from U.S. firm Forgepoint Capital International through its joint venture with Banco Santander. Earlier, Maisa had secured a $5 million pre-seed round led by NFX and Village Global.
Founded in 2024 by David Villalón and Manuel Romero, Maisa AI builds accountable AI systems designed to avoid the errors and “hallucinations” common in current models. Its new product, Maisa Studio, allows companies to create digital workers trained with natural language. Instead of generating direct answers, the system builds processes step by step — what Maisa calls “chain-of-work.”
Maisa also uses HALP (Human-Augmented LLM Processing), which lets users supervise how tasks are executed, and a Knowledge Processing Unit (KPU) designed to reduce hallucinations. Current clients include a major bank, as well as companies in car manufacturing and energy.
With its dual base in Valencia and San Francisco, the company positions itself as an alternative to traditional robotic process automation. Maisa says its tools give enterprises both trustworthiness and flexibility, offering deployment either in the cloud or on-premise.
The startup plans to expand its workforce from 35 to about 65 employees by early 2026 and expects significant growth in the last quarter of 2025 as it begins onboarding new customers from its waiting list.
Maisa faces competition from other AI workflow startups like CrewAI, but its focus on accountability and regulated industries could set it apart. “We are going to show the market that there is a company delivering what has been promised, and that it’s working,” CEO Villalón said.
