Why You Care
Ever wonder why so many promising AI projects never quite deliver? It turns out, a staggering 95% of generative AI pilots at companies fail to move beyond the pilot stage, according to the announcement. This isn’t just a small hiccup; it represents a massive lost opportunity for businesses. Imagine the productivity gains and new capabilities you could unlock if AI actually worked as promised. This new creation from Maisa AI could directly impact your business’s future AI success.
What Actually Happened
Maisa AI recently secured $25 million in funding to tackle a essential issue: the high failure rate of enterprise AI projects, as mentioned in the release. The company aims to make generative AI more reliable and effective for businesses. Maisa AI’s founders, David Villalón and Manuel Romero, previously worked together at Spanish AI startup Clibrain, the company reports. They teamed up in 2024 to build a approach after witnessing firsthand the unreliability of AI, according to the announcement. Their focus is on building processes around AI rather than just generating responses.
Maisa AI’s core creation is its ‘chain-of-work’ approach. This means they use AI to build the execution process, not just the final output, as detailed in the blog post. This differs from other platforms that primarily focus on ‘vibe coding’. The company also developed a system called HALP (Human-Augmented LLM Processing). This custom method involves asking users about their needs, then outlining the steps digital workers will follow, the technical report explains. This interactive process helps ensure AI aligns with user requirements.
Why This Matters to You
This new funding for Maisa AI is significant because it directly addresses a major pain point for businesses looking to adopt generative AI. If you’ve tried implementing AI and found it lacking, Maisa AI’s approach offers a potential approach. They are not just creating more AI models. Instead, they are focusing on the process of how AI delivers results. This could mean more predictable and reliable AI outcomes for your organization.
Maisa AI’s CEO, David Villalón, explained their unique approach: “Instead of using AI to build the responses, we use AI to build the process that needs to be executed to get to the response — what we call ‘chain-of-work’,” the company reports. This shifts the focus from raw AI output to a structured, auditable workflow. For example, imagine you need to generate complex financial reports. Instead of hoping an AI provides the correct numbers directly, Maisa AI’s system would outline the data sources, calculations, and verification steps the AI will take. This transparency builds trust.
Maisa AI also offers deployment flexibility, providing options for secure cloud or on-premise deployment, as mentioned in the release. This is crucial for companies with strict data privacy or security requirements. How much more confident would you feel deploying AI if you knew the process was transparent and auditable?
Here’s how Maisa AI’s approach compares:
| Feature | Traditional Generative AI | Maisa AI’s Approach |
| Primary Focus | Generating direct responses | Building the execution process |
| Reliability | Often unpredictable | Aims for higher predictability |
| Transparency | Limited | Increased process visibility |
| Error Handling | Prone to ‘hallucinations’ | Human-augmented review (HALP) |
| Deployment Options | Varies | Cloud or On-Premise |
The Surprising Finding
Perhaps the most surprising aspect of this creation is the sheer scale of generative AI project failures. The study finds that a 95% failure rate for generative AI pilots is a stark reality. This challenges the common assumption that simply deploying an AI model will automatically lead to success. Many might believe that the system itself is the main hurdle. However, the data indicates that the problem lies more in the implementation and reliability of these systems in real-world enterprise settings. This high failure rate suggests that while AI models are , their practical application often lacks the necessary structure and control. It highlights a essential gap between AI’s potential and its actual utility without proper process integration.
What Happens Next
Maisa AI’s new funding suggests a focus on expanding its enterprise solutions over the coming months, according to the announcement. We can expect to see more companies piloting their HALP system and ‘chain-of-work’ approach. For example, a large manufacturing firm might use Maisa AI to automate complex supply chain optimization tasks, ensuring each step is . This could lead to clearer, more reliable AI-driven decisions.
Businesses should consider exploring solutions like Maisa AI’s to improve their AI adoption rates. The company reports that they hope to position themselves as a more form of robotic process automation (RPA). This could unlock significant productivity gains without the rigid rules of traditional RPA. The team revealed that they are not skeptical about AI itself. However, they believe it’s not feasible for humans to review “three months of work done in five minutes.” This perspective underscores the need for automated verification and process building. This approach could reshape how enterprises integrate AI, moving towards more structured and auditable deployments in the coming quarters.