Rubrik co-founder Soham Mazumdar left in 2023 and owns a new data startup called Wisdomai. The AI ​​data analytics provided by the company can provide business insights through structured, unstructured, and even "dirty" data, meaning typos or wrong data is not cleared.

Where and how data is used is actually the holy grail of enterprise business intelligence software, and why Coatue leads a huge seed round of $23 million. Madrona, GTM Capital, Anthology Fund and others also attended the meeting.

Instead of requiring the data analytics team to run the report, the business manager can ask Wisdomai issues and step in the details.

Mazumdar gives an example of the revenue leader wondering: “How will I close my quarter?” Wisdomai’s answer will provide a list of pending that the team should focus on, as well as information about delaying everyone, such as a list of questions each customer is waiting for.

"You can see CRO literally through our platform, five key strokes, rather than involving five people, including some analysts and a lot of time," Mazumdar told TechCrunch.

This is just an example of the type of question Wisdomai wants to answer.

Another early client was an oil and gas company that used Wisdomai on site to ask questions about production, from storing documents to telemetry.

Obviously, every business analytics tool already available (and many startups) also provides AI-driven natural language tips.

Wisdomai stands out for the founder's lineage - all of which have previously worked with Mazumdar at Rubrik. But the superpower of the platform is its accuracy, even for messy data, Mazumdar said. It can find answers in structured data such as databases and unstructured data stored in files.

It is also important that Widsomai does not provide hallucinations.

Most businesses pursue the accuracy of AI applications by focusing on training their AI models’ data as well as model size, timely engineering and perhaps real-time search techniques such as Retrieval Authorization Generation (RAG). However, they still have the risk of being fabricated answers.

Wisdomai uses Genai in query formation, not in creating answers. "Ultimately, Genai can hallucinate. The job we use Genai is to write small programs ... can query these different systems," Mazumdar said.

So if Widsomai's model hallucinations, all it has to do is write a fake query that fails to retrieve the data. The data itself - the answer to the question - will not be fabricated.

Wisdomai claims Conocophillips, Cisco and Descope are early customers and have customers who are using major cloud data storage services such as Snowflake, Google's BigQuery, Amazon's Redshift, Databricks and Postgres. Mazumdar said any data storage system can be trained through query logs and other source research query languages.