AI Data Startup WisdomAI Secures 23 Million Dollars with a Smart Approach to Avoid Hallucinations
Soham Mazumdar, co-founder of Rubrik who left the company in 2023, has launched a new data startup called WisdomAI. The company specializes in AI-powered data analytics that can generate business insights from a variety of data sources, including structured, unstructured, and even "dirty" data, which refers to data that has not been cleaned of errors or inconsistencies.
WisdomAI’s approach to working with data as it exists in the real world is seen as a major breakthrough in enterprise business intelligence software. This capability was a key reason why Coatue led the company’s impressive 23 million dollar seed funding round. Other investors like Madrona, GTM Capital, Anthology Fund, and others also participated in the round.
Instead of relying on data analysts to run reports, WisdomAI allows business managers to ask direct questions and access the data they need with ease. For example, Mazumdar describes how a chief revenue officer (CRO) could ask, "How am I going to close my quarter?" WisdomAI would then provide a detailed list of pending deals to focus on, as well as information on what’s delaying each deal, including the specific questions each customer is waiting to have answered.
Mazumdar explained, "You can get the CRO to literally see all the way down to this last level of detail through our platform with, like, five key strokes, as opposed to a process which involves five individuals, including some analysts, and a whole lot of time."
This is just one example of the types of questions that WisdomAI aims to answer. The company already has early adopters, including an oil and gas firm with thousands of field workers using WisdomAI to ask questions about production. These workers can tap into data from sources ranging from stored documents to telemetry.
While other business analytics tools also utilize AI-powered natural language prompts, WisdomAI stands out thanks to the pedigree of its founders, all of whom worked with Mazumdar at Rubrik. But the true strength of the platform lies in its ability to provide accurate answers, even when working with messy or incomplete data. Mazumdar emphasizes that WisdomAI can pull useful insights not only from structured data in databases but also from unstructured data stored in files.
A key feature that sets WisdomAI apart from many other AI systems is its ability to avoid hallucinations. In the world of AI, hallucinations refer to situations where a model generates responses that are not based on real data or facts. Many companies focus on improving AI accuracy by refining their models, training data, and techniques like retrieval-augmented generation (RAG), but the risk of generating fabricated answers remains.
WisdomAI takes a different approach. While it uses Generative AI (GenAI) to assist in forming queries, it does not rely on GenAI to generate answers. Mazumdar explains, "Ultimately, GenAI can hallucinate. What we use GenAI to do is to write small little programs… that can query these different systems." This means if WisdomAI’s model were to hallucinate, it would simply generate an invalid query that fails to retrieve data — the data itself will always be accurate.
The company claims high-profile early customers such as ConocoPhillips, Cisco, and Descope, and is working with organizations using major cloud data storage services like Snowflake, Google’s BigQuery, Amazon’s Redshift, Databricks, and Postgres. WisdomAI is designed to be adaptable, allowing it to be trained on any data storage system by studying query language and logs.
With its focus on data accuracy, ease of use, and the avoidance of hallucinations, WisdomAI is positioning itself as a significant player in the rapidly evolving world of AI-driven business analytics.