Funding
Maisa Raises $25M to Power Trustworthy, Transparent âDigital Workersâ

Enterprise AI is plagued by failureâstudies show up to 95% of generative AI pilots collapse. But Maisa, bridging Valencia and San Francisco, is engineering accountability into automation. Following a $25 million seed round led by Creandum with backing from Forgepoint Capital, NFX, and Village Global, Maisa is poised to transform AI workflows with agentic process automation thatâs as auditable as it is intelligent.
Inside the Engine: KPU & Chain-of-Work â Building Trust from the Ground Up
At the core of Maisaâs platform is the Knowledge Processing Unit (KPU)âa novel architecture that reconceives AI reasoning. The KPU includes:
- A Reasoning Engine, powered by a large language model, which plans out multi-step workflows.
- An Execution Engine that carries out these plans and feeds outcomes back for recalibration.
- A Virtual Context Window that streamlines information flow, focusing the model only on relevant data to mitigate hallucinations.
The genius of the KPU is in how it reshapes the role of the LLM. Instead of acting as a probabilistic text generator, the model is treated as one component inside a disciplined computational framework. The KPU orchestrates reasoning like an operating system, breaking problems into manageable, auditable steps and ensuring each action can be validated. This structure turns a once-opaque model into a predictable system of logic, where errors can be detected, corrected, and prevented from cascading.
Complementing the KPU is the Chain-of-Workâa meticulously logged audit trail tracking every decision, action, and tool involved in a digital workerâs process. Unlike typical AI outputs that leave users guessing, the Chain-of-Work functions like a black box recorder for automation. Every calculation, every data pull, every reasoning step is preserved in detail. This gives enterprises the ability to not only trust outcomes, but to retrace them, replay them, and refine them over time. For compliance-heavy industries, this isnât just a convenienceâit is the foundation of safe deployment at scale.
Maisa Studio in Action: No-Code Agents Rooted in Transparency
Leveraging KPU and Chain-of-Work, Maisa Studio enables âcitizen developersâânon-technical staffâto deploy digital workers using plain language instructions. Through Maisaâs HALP (Human-Augmented LLM Processing), the system interacts to clarify intent, constructs workflows, integrates across hundreds of APIs, and begins learning dynamicallyâall without developers or datasets.
In practice, this has already delivered meaningful impact: a financial services firm cut false positives by 99% and achieved a 10Ă productivity boost per employeeâwith full deployment accomplished in just three onboarding sessions. Global banks, automakers, and energy companies are piloting the platform to automate compliance-heavy processes at scale, where transparency and auditability are non-negotiable.
Why It MattersâA Vision for AI You Can Trust and Scale
Maisa doesnât just offer automationâit delivers accountable AI that embeds trust by architecture. In industries burdened by regulation, opacity, and high stakesâfinance, healthcare, energyâthe ability to trace every automated decision is paramount. The Chain-of-Work enables businesses to inspect, audit, and validate AI logic at any step, meaning regulators and internal teams need not guess at how conclusions were reached. Instead, they can verify with precision.
Meanwhile, the KPUâs design systematically suppresses hallucinations by isolating reasoning from data noise and structuring execution. This reliability sheds the unpredictability that often makes enterprises wary of AI. Rather than producing outputs shrouded in mystery, Maisaâs agents deliver logical, predictable, and consistent results.
Looking ahead, this platform represents a fundamental shift: AI becomes a trusted collaboratorâone whose reasoning is transparent, whose actions are traceable, and whose âthought processâ can be refined and audited. As Maisaâs system is model-agnostic, organizations retain flexibility to adopt stronger models in the futureâwithout losing the rigor and oversight of the KPU framework. That adaptability lays the groundwork for sustainable, scalable deployment across evolving enterprise needs.
In essence, Maisa offers a blueprint for AI that isnât just powerfulâbut responsible and resilient. In a world where most AI projects fail to deliver, this technology charts a rare path forwardâcombining innovation with integrity.