WisdomAI, a new AI data analytics startup founded by Rubrik co-founder Soham Mazumdar, has secured a new $50 million Series A funding round. The investment was led by Kleiner Perkins and included participation from new investor NVentures, which is Nvidia’s venture capital arm. This funding arrives approximately six months after the startup announced a seed round of $23 million led by Coatue.
The company provides AI-driven data analytics that can answer business questions using structured, unstructured, and even “dirty” data. Dirty data refers to information that has not been cleaned of typos or errors. A business user can ask questions in natural language, for example, “How many customers do I have in my pipeline and what is preventing them from closing this quarter?”
WisdomAI employs a clever method to eliminate the LLM hallucination problem. The company does not use large language models to write answers to questions. Instead, LLMs are only used to write the query that will be sent to a data warehouse to retrieve information. This means if the LLM makes a mistake, it results in an ineffective query rather than inventing false answers.
The startup has developed its own logic called the “enterprise context layer,” which studies customer data to understand it. All of the company’s co-founders previously worked with Mazumdar at the data security company Rubrik, giving them deep experience with enterprise storage warehouses. Mazumdar left Rubrik in 2023.
Since its formal launch in late 2024, WisdomAI has grown from two enterprise customers to around forty. The company counts businesses such as Descope, ConocoPhillips, Cisco, and Patreon among its customers.
CEO Mazumdar also reported that the young company is experiencing growing usage within its enterprise customers. Some customers have doubled their usage within two months. Another customer began with ten seats and expanded to four hundred and fifty, which includes almost everyone in that company.
In the last six months, WisdomAI added an agentic feature that alerts users in real time to important changes in the situations they are monitoring. Mazumdar described creating an agent that watches product usage metrics and ticket information, a process that took him about five minutes. Instead of sending a daily or hourly report, the system alerts him when something interesting happens. He stated that the magic with analytics is that it has always been a static report, but they are making it dynamic and proactive.