Isotopes emerged from stealth on Thursday with a healthy $20 million seed round. The company offers an AI agent designed to solve a persistent problem in data analytics. The people who manage big data infrastructure are often not the ones who need to use the data to make business decisions.
Using large language models, Isotopes allows business managers to ask questions of their data in natural language. Its agent, Aidnn, can provide answers and even draft complex planning documents. It gathers data from wherever it is stored, including finance apps, ERP and CRM systems, and cloud storage.
While there are many agentic business analytics offerings, Isotopes’ co-founders have a unique pedigree. This background has resulted in a product so sophisticated the startup has already applied for ten patents, according to co-founder and CEO Arun Murthy.
Over twenty years ago, Murthy worked at Yahoo on the team that built the open-source project Hadoop. Hadoop spurred the initial big data frenzy of the 2010s. Yahoo later spun it out into a company called Hortonworks, with Murthy as a co-founder and chief product officer. Hortonworks went public just four years after its launch. The rise of new cloud storage technology eventually impacted Hadoop’s market, leading Hortonworks to merge with its biggest rival, Cloudera. The merged company was taken private in 2021.
Murthy spent a few years at Cloudera, managing about two hundred people. Even there, he witnessed the enduring problem of data access. He recalls embarrassing quarterly conference calls where Wall Street analysts grilled executives on operating details they could not answer because they lacked access to the data.
In 2021, Murthy left Cloudera with no definite plan. A venture capitalist introduced him to Scale AI’s Alexandr Wang. After several conversations and some consulting work, Murthy joined Scale as its chief technology officer. He describes this time as akin to getting a PhD, providing a deep understanding of what drives AI models and how to improve them.
When his old colleague from Hortonworks, Prasanth Jayachandra, contacted him, the two decided to launch their own AI startup. They persuaded a third co-founder, Gopal Vijayaraghavan, also from their Hortonworks days, to join them. They founded Isotopes in late 2024. Their seed round was led by Vab Goel at NTTVC.
The founders’ extensive backgrounds have enabled them to build a powerful agent. It can find data from any storage system, such as Salesforce or Snowflake, and then clean that data. The agent also maintains significant context memory to be useful for complex, multi-step tasks.
This is far beyond a simple chatbot. For example, if a team asks Aidnn to draft a report on monthly recurring revenue trends, the agent must execute a complex plan. This involves extracting metadata, reading the data, cleaning and normalizing it, joining data sets, prorating revenue, and aggregating the results. The agent shows its steps, reasoning, and assumptions, and it points out anomalies in the data. It will even make recommendations on how to proceed. Isotopes also promises enterprise customers can deploy the agent without sharing their data with external AI model makers.
Despite its sophistication, Isotopes faces significant competition. Incumbents like Salesforce’s Tableau already offer their own agents as part of a major AI push. Furthermore, many other startup founders with impressive pedigrees are also competing in the same market.