A new startup, Eragon, is betting that the future of enterprise technology will move beyond traditional software interfaces, as it develops an artificial intelligence driven operating system built entirely around natural language prompts. The company, founded by Josh Sirota, has raised $12 million in fresh funding at a $100 million valuation to advance what it describes as an “agentic AI” platform for businesses.
Sirota’s vision is rooted in a bold premise that conventional software tools, defined by dashboards, menus and buttons, will eventually give way to systems where users simply instruct machines through conversation. The platform aims to replace or integrate widely used enterprise tools such as Salesforce, Snowflake, Tableau and Jira, allowing companies to manage operations, analyse data and automate workflows through a single AI interface.
The company’s approach centres on deploying customised AI models within a client’s own infrastructure, ensuring that sensitive business data remains secure and under corporate control. Eragon trains models on internal datasets and enables automated actions such as onboarding new customers, generating reports, and even processing invoices, all initiated through simple prompts rather than manual workflows.
Industry observers say the concept reflects a broader shift toward agent based artificial intelligence systems capable of both analysis and execution. Speaking at a recent developer event, Jensen Huang, chief executive of Nvidia, noted that enterprise AI agents could fundamentally reshape white collar work, much like personal computing transformed office productivity decades ago.
Despite the optimism, questions remain around reliability, security and adoption, as many AI deployments in corporate settings have struggled to scale effectively. Still, Eragon’s early traction among startups and large firms suggests growing interest in tools that simplify complex operations, and the company believes its approach could redefine how businesses interact with technology in the years ahead.