Writer launches a ‘super agent’ that actually gets sh*t done, outperforms Open AI on key benchmarks

Writer, a prominent enterprise AI firm currently valued at $1.9 billion, has launched its latest innovation—an autonomous 'Action Agent.' This groundbreaking technology is designed to independently handle intricate, multi-step business tasks across a wide array of software platforms, marking a significant leap in the ongoing corporate AI competition. Unlike traditional AI chatbots that merely respond to queries, the Action Agent is capable of carrying out entire projects autonomously. It can browse websites, analyze data, create presentations, write code, and manage tasks across an organization’s technological framework without the need for human intervention. May Habib, CEO and co-founder of Writer, highlighted the difference: "Other AI chatbots can tell you what to do. Action Agent does it." This launch positions Writer as a strong contender in the enterprise AI market, competing with giants like Microsoft’s Copilot and OpenAI’s ChatGPT. The Action Agent is tailored for the business sector, incorporating robust security features and audit trails crucial for regulated industries such as banking and healthcare. What sets Writer's Action Agent apart is its operation at what the company refers to as “level four orchestration,” a higher echelon of AI automation compared to the more common level one or two systems. Matan-Paul Shetrit, Writer’s head of product, explained, "What we’ve done here is full orchestration. This is an agent that calls agents, writes its own tools when needed, and executes with full visibility." During a demonstration, the capabilities of the Action Agent were showcased as it successfully performed a clinical trial site selection. This typically labor-intensive process, requiring weeks of human research, was accomplished with astonishing efficiency. The agent analyzed demographic data across multiple cities, ranked locations based on suitability, and produced detailed reports—all autonomously. The Action Agent’s performance has been validated by impressive benchmark results. It scored 61% on GAIA Level 3, which is the most rigorous benchmark for AI agents, outperforming competing platforms including OpenAI’s Deep Research. Furthermore, it achieved a top score on the CUB (Computer Use Benchmark) leaderboard, demonstrating its aptitude for handling complex reasoning tasks that have historically challenged AI systems. Writer's enterprise-centric approach distinguishes it from many consumer-focused AI tools. The Action Agent is built on the company’s existing platform, which already supports numerous major corporations like Accenture, Vanguard, Qualcomm, Uber, and Salesforce. This focus on business needs emphasizes security, governance, and transparency, allowing organizations to see exactly how the agent operates and makes decisions. The integration capabilities of the Action Agent are significant, as it connects with over 600 enterprise tools, utilizing the Model Context Protocol (MCP) enhanced with enterprise-grade security controls. Writer has collaborated with Amazon Web Services and other industry leaders to elevate MCP to meet enterprise standards. In a strategic move, Writer has decided to provide the Action Agent free to its existing customers, challenging conventional software pricing models. This decision reflects the company's confidence in its cost-effective development approach, having spent just $700,000 to train its Palmyra X4 model, compared to significantly higher costs for similar models by competitors. As the enterprise AI market is projected to grow from $58 billion to $114 billion by 2027, Writer is at the forefront, directly competing with Microsoft, Google, and OpenAI. However, Writer aims to carve out a niche by focusing specifically on enterprise needs, emphasizing security and reliability over raw capabilities. Looking ahead, Writer plans to significantly expand the Action Agent’s functionalities, adding connections to 80 more enterprise platforms and third-party data providers, thus enabling a comprehensive suite of tools for its users. The gradual rollout of these capabilities is designed to incorporate customer feedback and enhance the system's effectiveness while adhering to stringent security standards. The launch of the Action Agent signifies a transformative moment in the enterprise AI landscape, transitioning AI from a novelty to essential business tools. As companies adapt to this shift, the successful integration of AI alongside human workers could provide a substantial competitive edge for those willing to embrace this new paradigm.