Here, Maisa’s co-founder, David Villalón, discusses the company’s origins, its approach to automation, and the journey of building an AI platform designed for enterprise-grade performance.

What does Maisa focus on and how was the company founded?

Maisa is an Agentic Process Automation Platform (APA or also RPA 2.0) that focuses on automating complex, knowledge-intensive business processes with full traceability, accountability, and reliability. Unlike traditional AI agent solutions, Maisa ensures deterministic execution through our unique ‘Chain-of-Work’ approach, which eliminates the risks of hallucinations and inconsistent outputs often associated with standard LLM or RAG-based systems.

The company was founded based on our firsthand experience with AI models and their limitations in business automation. We recognised that existing AI approaches, particularly those using LLMs, finetuning, RAG-based retrieval and multi-agent frameworks, struggled with reliability and scalability. We observed that as token outputs increased significantly, so did risks, which compounds errors over multiple steps, making conventional AI systems unreliable for enterprise use.

To solve these challenges, our team took inspiration from operating system architecture models, leading to the development of Maisa and our Knowledge Processing Unit (KPU). This unique approach ensures every decision and action taken by our AI Agents can be fully audited and explained. This foundation has positioned Maisa as the first accountable AI platform capable of automating high-stakes business processes with complete transparency and reliability.

Who is your target market and why

We primarily target enterprises that require reliable, transparent, and scalable AI-driven automation. Industries such as financial services, energy, cybersecurity, manufacturing, and all cross-sector supply chain management, benefit significantly from Maisa’s AI Agents, which function as digital workers. These industries demand high levels of compliance, risk management, and decision-making accuracy, making traditional AI solutions unsuitable due to their probabilistic nature.

Many businesses struggle with integrating AI seamlessly into their workflows without costly engineering overheads. Maisa’s platform is built to eliminate these inefficiencies by providing an all-in-one automation solution so that businesses can focus on outcomes rather than managing complex infrastructure.

Many businesses struggle with integrating AI seamlessly into their workflows without costly engineering overheads. Maisa’s platform is built to eliminate these inefficiencies by providing an all-in-one automation solution so that businesses can focus on outcomes rather than managing complex infrastructure.

With Maisa, onboarding, creating, and deploying AI workers happens in hours and days instead of weeks and months. Business process operators are placed at the heart of every implementation, making the experience of deploying a new AI worker as intuitive and seamless as onboarding a new colleague.

What challenge are you trying to overcome?

The AI industry is grappling with several significant challenges that hinder its widespread adoption and effectiveness. One of the primary concerns is reliability, as most AI solutions operate on probabilistic models, which can lead to inconsistent and unpredictable outputs. This lack of determinism makes it difficult for businesses to depend on AI for critical decision-making.

Transparency is another pressing issue. Many AI models function as opaque "black boxes," offering little to no insight into how they reach their conclusions. This absence of explainability not only raises concerns about trust but also makes it difficult to audit AI-driven resolutions.

Scalability further complicates AI adoption, as many systems struggle to integrate seamlessly into enterprise environments. Companies often find that AI implementations require extensive customisation, leading to slow and cumbersome deployment, often involving complicated tech stacks with guardrails and evaluations to mitigate inherent error risks in system design.

Setting up and maintaining AI solutions at scale is often a resource-intensive endeavour. Companies must invest in expert knowledge, substantial IT infrastructure and tech stack to deploy AI effectively, making the process both costly and complex.

As regulatory frameworks around AI continue to evolve, businesses face increasing pressure to ensure compliance. Many AI platforms fail to meet the growing demand for explainability, putting organisations at risk of non-compliance with emerging regulations.

Maisa addresses these issues by providing a robust, reliable, and enterprise-ready all-in-one AI solution enabled by a fundamentally more reliable architecture based on the KPU operating systems.

Can you tell me a bit about your entrepreneurial journey?

As mentioned, the journey of building Maisa began with a deep frustration over the limitations of existing AI solutions. 

My co-founder Manuel Romero and I used this as fuel to set-up Maisa, with the help of a small team, developing a proof of concept to showcase Maisa’s potential. That early work won the trust of top investors.

Having raised earlier last year, in December we announced our $5 million pre-seed round funded by NFX and Village Global (the VC backed by Mark Zuckerberg, Eric Schmidt and Jeff Bezos). They were joined in the round by local business angels, Sequoia scout and DeepMind PM Lukas Haas.

Through rigorous testing, refining, and achieving product-market fit, we have grown into an enterprise-grade automation platform that provides AI-driven solutions with complete accountability. Today, Maisa is scaling its impact across industries, proving that AI can be both powerful and transparent.

What's the one thing you would say to any startup founder that's just getting started?

The key to success is to challenge conventional wisdom. Many successful founders don’t just react to trends. Instead, they anticipate what’s coming next and think independently about how to solve fundamental problems. The AI industry, for example, has been heavily focused on LLMs and RAG-based approaches despite their inherent flaws. Instead of following the crowd, we reimagined AI systems and automation from the ground up, creating a solution that ensures full accountability.

Additionally, curiosity and adaptability are invaluable traits, and those dedicated to deep learning can achieve a substantial advantage. With 1,000 hours of focused effort, mastery in any field becomes attainable.

In the AI space, having a deeply technical founder - or at least a co-founder - is essential. Without a genuine understanding of the technology and its capabilities, you’ll struggle to stay ahead or fully harness its potential. An AI-native approach to building and scaling is foundational and fundamentally distinct from simply 'using AI.’ The most successful AI companies are led by those who don’t just adapt to advancements but actively shape them.

Finally, building a strong team is crucial. Even the best ideas won’t scale without the right people to execute them. Recruiting exceptional talent, fostering a culture of collaboration, and aligning around a shared vision are essential for turning an idea into a thriving company.

For any startup founder, the most important thing is to remain bold, challenge assumptions, and be willing to rethink how things are done. That’s where true breakthroughs happen.