TechCrunch Startup News Maisa AI gets $25M to fix enterprise AI’s 95% failure rate
Sep 1, 2025
Discover how Maisa AI is tackling the staggering 95% failure rate of enterprise AI by offering accountable, customizable agents. The discussion highlights the necessity of trust and transparency in automation to improve outcomes. Additionally, insights into scaling operations in regulated sectors and expanding into international markets reveal the strategies startups are adopting to gain a competitive edge. Exciting funding news ushers in a new era for innovation, alongside a glimpse of the Audi Q6 e-tron.
AI Snips
Chapters
Transcript
Episode notes
Enterprise AI Pilots Mostly Fail
- 95% of generative AI pilots are failing, highlighting a systemic problem in enterprise adoption.
- Maisa AI positions agentic, accountable systems as the alternative to opaque models.
Process-First Approach Reduces Hallucinations
- Maisa builds processes (a 'chain of work') to reach answers instead of only generating responses.
- This process-first approach aims to reduce hallucinations and increase accountability.
Founders Built Maisa From Hallucination Pain
- Co-founders Villalón and Manuel Romero built Maisa after seeing hallucination failures at their prior startup Silbrain.
- Their hands-on experience shaped Maisa's focus on trust and accountable digital workers.