Every new technology creates a new environment for us to work in, but it’s not clear how AI will do that. One possibility is that the interface disappears completely.
That’s Josh Sirota’s vision. He founded the startup Eragon in August, which just raised $12 million at a $100 million post-money valuation to build an agent AI operating system for enterprise customers.
There is a simple theory. “Software is dead,” Sirota says. Buttons, dialog boxes, and pull-down menus are a thing of the past; the future of business is driven by prompts. Eragon seeks to offer its entire suite of business software, including Salesforce, Snowflakes, Tableaus, and Jiras, through its LLM interface.
Mr. Sirota, who previously worked on go-to-market teams at Oracle and Salesforce, admits he had a bit of a quarter-life crisis before moving to San Francisco and launching Eragon with a small team in a live work loft across from Giants Stadium. On a recent sunny Wednesday, the dining room table contained a bottle of Moet, a few Mac minis, and a copy of Eragon, the fantasy novel by Christopher Paolini that inspired the company’s name. This is also based on the tradition of Palantir and Anduril, borrowed from the fictional world.
Mr. Sirota’s experience implementing the world’s first enterprise software convinced investors of his “founder-market fit.” His backers include Long Journey Ventures, Soma Capital, Arielle Zuckerberg of Axiom Partners, and strategic angels Mike Knoop and Elias Torres.
“We think there’s a huge potential for Eragon to become the connective tissue of how modern teams operate and make decisions,” Axiom’s Sandhya Venkachalam said. Eragon’s technical talent includes Rishabh Tiwari, a computer science doctoral student at Berkeley, and Bin Agarwal, an MIT doctoral student. They work together to build the company’s technology stack.
In Eragon’s Customer Center of Excellence (a shabby white sofa), Sirota shows how the company eats its dog food. Eragon post-trains open source models like Qwen and Kim on customer data sets and links them to corporate email accounts and other resources. When Sirota wants to acquire a new customer (he’s demonstrating the tool this week at Dedalus Labs, which has adopted it), he asks with a natural language prompt, and the software automatically assigns each new user’s credentials, launches a new Eragon instance in the cloud, and begins the onboarding workflow.
tech crunch event
San Francisco, California
|
October 13-15, 2026
Mr. Sirota expects that software executives will ask Mr. Eragon to push agents to take action, such as asking them to analyze what deals are likely to fail or taking steps to improve supply chain lead times. Need a dashboard? Ask Eragon to spin it up.
While the demo is convincing, it’s easy to imagine edge queries that confuse the software and failures that are difficult to audit. Mr. Sirota used Eragon to demonstrate automatic invoice approval. The system processes your invoice when it arrives in your inbox. So this reporter considered filing a claim to see what would happen. (Dear reader, I didn’t know.)
The security concerns raised by AI agents are significant, but for now the company is trying to solve real-world workplace glitches. Eragon is currently used by a few large companies and dozens of startups. Nico Laqua, CEO of insurance startup Corgi, which spun off from Y Combinator last year and raised $180 million, calls Eragon “the most enterprise-ready AI on the market.”
“Most of the data we have needs to be kept secure behind our own cloud,” Laqua said. “Eragon trains state-of-the-art models based on our data and deploys them in our environments.”
This is the core of Mr. Eragon’s pitch. A company’s data remains within its own servers and security environment, and it owns its own model weights (the underlying parameters that define the AI’s behavior). Sirota expects that models trained on years or decades of company data will become valuable assets in their own right. And while Frontier Labs may have the most capable model, as long as companies need to access models via API without owning the configuration, Sirota believes Eragon has an edge in the market.
He likens the evolution of AI software to the transition from mainframes to personal computers. While Frontier Labs offers powerful, centralized services, enterprise mass deployments rely on bespoke, purpose-built local tools. Enterprises will need and want to control agents and models for specific purposes.
Days later, Nvidia CEO Jensen Huang echoed similar sentiments at GTC, Nvidia’s annual developer conference, arguing that enterprise agent AI tools will replace existing approaches to white-collar work. “It’s no different than how Windows made it possible to create personal computers…Every SaaS company will become Agentic as a Service.”
Huang’s comments relate to NemoClaw, Nvidia’s new initiative aimed at making it easier for OpenClaw agents to work within secure enterprise systems. This is a sign both that Sirota is on to something and that competition will be fierce with everyone from Frontier Labs to model rappers.
Undaunted, Sirota said he expects Eragon to be a billion-dollar company by the end of the year. He’s aware of the oft-quoted MIT figure that 95% of AI companies’ trials fail at adoption, but jokes that it’s because senior executives don’t know what their employees do all day. Eragon aims to give them something they can actually work with.
