Enterprise AI Revolution: How Eragon’s Bold Vision Aims to Replace Software with Simple Prompts

In a significant move that challenges decades of software design, startup Eragon has secured $12 million to pursue a radical vision: making enterprise software interfaces disappear entirely, replaced by natural language prompts. Founded in August by former Oracle and Salesforce executive Josh Sirota, the company operates from a live-work loft in San Francisco, aiming to build what it calls an “agentic AI operating system” for business. This funding round, valuing the company at $100 million post-money, signals growing investor confidence in a future where complex business tools are accessed not through buttons and menus, but through conversation.

Eragon’s Vision for an Agentic AI Operating System

The core thesis driving Eragon is stark and simple. “Software is dead,” declares founder Josh Sirota. He argues that traditional graphical user interfaces—with their buttons, dialog boxes, and pull-down menus—represent a bygone era. Instead, Eragon is developing a platform that offers the entire suite of common business software, from Salesforce and Snowflake to Tableau and Jira, through a single large language model (LLM) interface. This approach fundamentally reimagines how employees interact with the digital tools that power modern corporations.

Furthermore, Sirota’s background provides crucial context for this ambitious project. His experience on go-to-market teams at major enterprise software firms gives him intimate knowledge of the pain points he now seeks to eliminate. Investors, including Arielle Zuckerberg at Long Journey Ventures and Axiom Partners, cited this “founder-market fit” as a key reason for their backing. Sandhya Venkatachalam of Axiom Partners stated, “We see enormous potential for Eragon to become the connective tissue for how modern teams operate and make decisions.”

The Technical Architecture and Security Promise

Eragon’s technical execution is led by a team including Berkeley and MIT PhDs, Rishabh Tiwari and Vin Agarwal. Their strategy involves post-training open-source models like Qwen and Kimi on specific customer datasets. Crucially, the company emphasizes a security-first model where a client’s data never leaves its own servers. The AI models are trained and deployed within the customer’s own secure cloud environment, and the company retains ownership of its own model weights—the fundamental parameters that dictate an AI’s behavior.

This architecture directly addresses one of the biggest hurdles to enterprise AI adoption: data security and control. Nico Laqua, CEO of insurance startup Corgi, an early Eragon user, highlighted this advantage. “Most of the data we have needs to remain secure and behind our own cloud,” Laqua said. “Eragon trains state-of-the-art models for us on our data and deploys it in our own environment.” Sirota believes models trained on decades of proprietary corporate data will themselves become invaluable strategic assets.

A Real-World Demonstration and Competitive Landscape

During a demonstration, Sirota showed how Eragon functions in practice. To onboard a new customer, Dedalus Labs, he simply issued a natural language prompt. The system then automatically assigned user credentials, spun up a new cloud instance, and initiated a tailored onboarding workflow. The platform can generate dashboards on demand, analyze potential deal slippage, or suggest supply chain optimizations, all triggered by conversational commands.

However, this vision is not without its challenges. The industry faces well-documented issues with AI hallucinations, edge-case failures, and the auditability of AI-driven decisions. The competitive landscape is also intensifying. Notably, Nvidia CEO Jensen Huang recently presented a similar vision at Nvidia’s GTC conference, announcing the NemoClaw initiative to help AI agents operate within secure enterprise systems. Huang argued that “every single SaaS company will become Agentic as a Service,” validating the market trend while also signaling the arrival of powerful competitors.

The Broader Shift in Enterprise Software Paradigms

Sirota frames this evolution as analogous to the historical shift from mainframes to personal computers. While frontier AI labs offer powerful, centralized models via API, he contends that true mass corporate adoption requires local, bespoke tools that companies can control. The promise of Eragon is to provide the foundational operating system upon which these specialized corporate agents can be built and managed.

Despite an MIT-cited statistic that 95% of corporate AI trials fail, Sirota remains confident. He humorously suggests the high failure rate stems from executives not fully understanding their employees’ daily tasks. Eragon aims to bridge that gap by providing an intuitive interface that translates executive queries into actionable insights and automated workflows. The startup is currently deployed in several large enterprises and dozens of smaller startups, serving as its own first customer to refine the product.

Conclusion

Eragon’s $12 million funding round represents a bold bet on a prompt-driven future for enterprise software. By advocating for the death of traditional interfaces and championing an agentic AI operating system, Josh Sirota and his team are positioning themselves at the forefront of a potential workplace revolution. The coming years will test whether natural language can truly unseat decades of GUI design, and whether enterprises will trust AI agents with their most critical operations. The journey from a San Francisco loft to challenging software giants is underway, and the stakes for the future of white-collar work have never been higher.

FAQs

Q1: What is an “agentic AI operating system”?
An agentic AI operating system is a platform that uses artificial intelligence to autonomously execute tasks and workflows based on high-level user commands or prompts. Instead of a human manually clicking through software, AI agents interpret the goal and perform the necessary actions across various business applications.

Q2: How does Eragon handle data security and privacy?
Eragon’s architecture is designed for data sovereignty. A company’s data remains within its own servers and security perimeter. Eragon trains the AI models directly on this internal data, and the resulting model weights—the core of the AI’s intelligence—are owned and controlled by the client, not by Eragon.

Q3: What is the main difference between Eragon and using an API from a company like OpenAI?
The key difference is control and customization. Using a generic API means your queries are processed on the provider’s servers with their model. Eragon enables companies to build, own, and deploy custom AI models trained specifically on their proprietary data, operating entirely within their own secure environment.

Q4: What kinds of business tasks is Eragon designed to automate?
The platform aims to handle a wide range of tasks, from generating business intelligence dashboards and analyzing sales pipelines to managing onboarding workflows, approving invoices, and optimizing supply chain logistics—all initiated through simple natural language prompts.

Q5: Who are Eragon’s main competitors?
Eragon faces competition from multiple angles: large “frontier” AI labs offering enterprise APIs, other startups building “AI wrapper” products, and now major tech infrastructure companies like Nvidia, which are launching tools to help businesses build their own agentic systems, as seen with the NemoClaw initiative.