Maisa, a leader in developing hallucination-resistant AI agents, announces an initial $25 million (€21.4 million) investment led by Creandum, with participation from Forgepoint Capital through its European joint venture with Banco Santander.
The announcement comes just months after the company raised $5 million in a pre-seed round in December 2024, with participation from NFX and Village Global, backed by Mark Zuckerberg, Jeff Bezos, and Eric Schmidt. Both firms continued their support and also participated in the current round.
The new funding will enable Maisa to expand its teams in AI R&D, engineering, sales, and customer success, as well as extend its growing presence in Europe and North America.
Alongside the fundraising, Maisa is launching Maisa Studio, a process automation platform through agents that allows “citizen developers”—users who are experts in their business domain but lack IT training—to deploy powerful, fully auditable AI “digital workers” trained via natural language.
The platform is already being tested at scale within global banks, automotive manufacturers, and energy companies, among others, to execute complex, multi-step, compliance-sensitive workflows with full traceability and reliability.
Maisa Studio enables non-technical staff to integrate digital workers as easily as adding new colleagues. Users can quickly deploy a digital worker capable of executing complex processes requiring high expertise, from risk assessment and transaction reconciliation to monitoring supply chain disruptions, eliminating long hours of repetitive human labor and unreliable outcomes.
These digital workers do not require datasets or developers. Users simply input commands in natural language into the platform, and the agents learn on the fly through a method called HALP (Human-Augmented LLM Processing), a fast and enterprise-ready way to train digital workers.
Maisa was founded in 2024 by David Villalón, CEO, and Manuel Romero, CSO, two leaders in applied and foundational AI. Villalón previously served as Head of AI at Clibrain and Product Director at Voicemod, while Romero is one of HuggingFace’s top global contributors, with over 700 open-source models and 15 million monthly downloads.
Last month, Maisa was recognized by Gartner, alongside Google and Amazon, as a leading AI agent provider in its Hype Cycle for AI and Future of Work reports—the first time a Spanish startup appeared in one of these influential publications.
Despite the surge in AI adoption, most enterprise initiatives fail or remain unimplemented. A recent report by the MIT NANDA initiative revealed that 95% of generative AI pilot projects have minimal or no impact. IDC data shows that 88% of AI pilots do not reach production, and BCG reports that only 22% of companies have advanced AI beyond proof-of-concept to generate tangible value.
Moreover, an i4cp HR study indicates that over two-thirds (70%) of organizations struggle to equip their workforce with the necessary AI skills, and Bain found that 44% of executives cite a lack of internal AI expertise as a barrier to implementation.
Maisa removes these barriers, providing tools that enable fast, safe, and effective AI adoption.
David Villalón, co-founder and CEO of Maisa, stated:
"This investment confirms that enterprise AI must be trust-based. Our platform allows teams to use AI workers that not only operate with incredible accuracy and intelligence but also explain and demonstrate their logic while documenting every step of the workflow. This means users can scale AI at their own pace, safely, without needing a full development team."
Peter Specht, General Partner at Creandum, said:
"Maisa is solving one of AI’s most complex challenges: making its use reliable and safe in critical business operations. With several leading global companies already using the technology and strong demand, we have full confidence in significant large-scale growth."
Alberto Yepez, from Forgepoint Capital International, added:
"Maisa’s platform is specifically designed for highly regulated sectors, such as finance, where decisions must be traceable and results consistent. The company’s initial success demonstrates real market interest in well-implemented AI, and the launch of Studio today will allow anyone within an organization to use digital workers quickly and effectively."
Maisa’s system is model-agnostic, turning general-purpose LLMs into safe, specialized AI agents that operate reliably across regulated industries. The platform can run on Maisa’s secure cloud or be deployed privately to meet enterprise compliance requirements.
It is built around a computable system called the Knowledge Processing Unit (KPU), a proprietary reasoning engine that converts large language models (LLMs) into reliable task executors delivering accurate and dependable results.
Rather than relying on probabilistic guesses, Maisa’s digital workers use AI guided through step-by-step logical calculations to perform tasks. Each action is recorded in Maisa’s Work Chain, a transparent and auditable record of logic and execution, allowing companies to see exactly how each outcome was achieved, reducing hallucinations.
Early traction demonstrates the potential of this new automation approach:
These are not experiments or prototypes. They are operational, production-grade deployments within organizations with high regulatory and accuracy requirements.
Studio integrates immediately with over 450 third-party systems. It connects automatically to any documented API and, when updated to specification, to any undocumented custom API. Automations can be triggered via web, email, API, or webhook and are continuously updated through the same conversational interface used to create them. It also interacts with websites and legacy systems.
cloud technology axon
RSM has advised the shareholders of JG Ingenieros on a transaction tha...
