Genloop is creating AI agents that self-learn as the business evolves
This Santa Clara-based startup has developed technology that allows businesses to get directly in touch with their structured data through natural languages.
There are operational hazards in every industry. But dealing with data means spending time trying to get hold of all the reports, and the back and forth ends up costing an enterprise a lot.
In 2024, technologist Ayush Gupta noticed this problem and founded Genloop—a deeptech startup that has developed its AI agents that allow businesses to get directly in touch with their structured data using natural languages like English, for instance.
Gupta had been building language models eight years ago. But when the Gen AI and ChatGPT boom was happening, he noticed the transformational potential technology had. But there were some immediately noticeable problems.
These systems were probabilistic and hence, not reliable for enterprise use. His foundational goal for Genloop was to make these powerful systems more deterministic and reliable in enterprise settings.
“Our initial action was building a solution that continuously trains and optimises open-weight Large Language Models (LLMs). This product customises the LLMs by injecting business knowledge projection to ensure accurate performance for specific enterprise needs, ultimately forming GenLoop's core stack,” CEO Gupta tells YourStory.
The startup is based in Santa Clara, California, and raised about $300,000 this year in a pre-seed round led by TRTL and Pegasus Angel Accelerator. Launched in September this year, the 10-member team has amassed seven B2B clients.
Building the technology
The deeptech startup’s technology acts like an AI assistant built for big companies to access their structured data, like SQL data. Earlier, if a business wanted to ask a complex question, such as "What were the sales of this particular product in this particular region?", the popular AI models would often guess wrong because of their probabilistic nature.
Once Genloop optimised LLMs for enterprise-specific needs, it started building its flagship product, Data to Insight.
“This specific product addresses the significant challenge companies face when trying to converse with their vast amount of structured or SQL data (like large tables and columns). Our solution helps business users directly chat with their SQL data, bypassing the dependency on data science teams and eliminating the bottleneck created by static dashboards (like those built in Power BI or Tableau),” explains Gupta.
The founder says that by achieving reliable, accurate, and complex insights, GenLoop aims for “100% reliability, significantly reducing the ad hoc requests and saving time in decision-making”.
Genloop is offering the first enterprise-grade GenBI platform powered by self learning agents. It is building AI data analysts that help sales and operations teams decide and move faster. The core of this process is building a unified business memory that captures an enterprise’s tribal knowledge and continuously improves it with learning. Agents grounded in enterprise understanding work together to answer deep analysis requirements in simple conversations.
While it competes with companies like Thoughtspot, Wisdom, and Omni, the CEO says Genloop separates itself due to its self-learning agents that continuously learn with the evolving business.
“We are the only offering that can run completely on self-hosted customised LLMs/SLMs (Small Language Models) and deploy as an air-gapped solution,” he signs off.
Genloop is part of YourStory’s Tech30 cohort—a selection of India’s most promising startups of 2025—unveiled at TechSparks Bengaluru.
