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

Writer, the enterprise artificial intelligence company valued at $1.9 billion, launched an autonomous “super agent” Tuesday that can independently execute complex, multi-step business tasks across hundreds of software platforms — marking a significant escalation in the corporate AI arms race.

The new Action Agent represents a fundamental shift from AI chatbots that simply answer questions to systems that can autonomously complete entire projects. The agent can browse websites, analyze data, create presentations, write code, and coordinate work across an organization’s entire technology stack without human intervention.

“Other AI chatbots can tell you what to do. Action Agent does it,” said May Habib, Writer’s CEO and co-founder. “It’s the difference between getting a research report and having your entire sales pipeline updated and acted upon.”

The launch positions San Francisco-based Writer as a formidable competitor to Microsoft’s Copilot and OpenAI’s ChatGPT in the lucrative enterprise market, where companies are racing to deploy AI systems that can automate knowledge work. Unlike consumer-focused AI tools, Writer’s agent includes enterprise-grade security controls and audit trails that regulated industries like banking and healthcare require.

Writer’s Action Agent fundamentally differs from existing AI assistants by operating at what the company calls “level four orchestration” — the highest tier of AI automation. Most current enterprise AI tools operate at levels one or two, handling basic tasks like answering questions or retrieving documents.

“The reality is most of the market is anywhere between one to two,” explained Matan-Paul Shetrit, Writer’s head of product, in an interview with VentureBeat. “What we’ve done here is full orchestration. This is an agent that calls agents, writes its own tools when needed, can execute on that with full visibility.”

The distinction goes far beyond simple automation capabilities. While traditional AI assistants like ChatGPT or Copilot are “very much built for like a Q and A experience,” Shetrit noted, Action Agent is designed for execution. “The difference is, one is not just about like, let me do this back and forth brainstorming, but more like, once and if I want to do the brainstorming, I can also act on it.”

The agent operates within its own isolated virtual computer for each session, allowing it to independently browse web pages, build software, solve technical problems, and execute complex multi-step plans. When asked to perform a product analysis, for example, Action Agent will automatically process thousands of customer reviews, perform sentiment analysis, identify themes, and generate a presentation — all without human guidance.

The system’s capabilities extend to generating its own tools when existing ones prove insufficient. “It can action whether or not it has MCP or any tool access, because it can just generate its own tools on the fly for the purpose of the task,” Shetrit explained.

During a demonstration, Shetrit showed the agent conducting clinical trial site selection — a process that typically requires weeks of human research. The agent systematically analyzed demographics across multiple cities, ranked locations by suitability criteria, and generated comprehensive reports with supporting evidence.

“This is weeks worth of work by these companies,” Shetrit noted. “It’s not something that’s trivial to do.”

Writer’s claims about the agent’s capabilities are backed by impressive benchmark results. Action Agent scored 61% on GAIA Level 3, the most challenging benchmark for AI agent performance, outperforming competing systems including OpenAI’s Deep Research. The agent also achieved a 10.4% score on the CUB (Computer Use Benchmark) leaderboard, making it the top performer for computer and browser use tasks.

These results demonstrate the agent’s ability to handle complex reasoning tasks that have traditionally stumped AI systems. GAIA Level 3 tests require agents to navigate multiple tools, synthesize information from various sources, and complete multi-step workflows — precisely the kind of work that enterprises need automated.

The performance stems from Writer’s Palmyra X5 model, which features a one-million-token context window — enough to process hundreds of pages of documents simultaneously while maintaining coherence across complex tasks. This massive context capability allows the agent to work with entire codebases, lengthy research reports, and comprehensive datasets without losing track of the overall objective.

Writer’s enterprise focus sets it apart in a market dominated by consumer-oriented AI companies attempting to adapt their products for business use. The company built Action Agent on its existing enterprise platform, which already serves hundreds of major corporations including Accenture, Vanguard, Qualcomm, Uber, and Salesforce.

The distinction proves crucial for enterprise adoption. While consumer AI tools often operate as “black boxes” with limited transparency, Writer’s system provides complete audit trails showing exactly how the agent reached its conclusions and what actions it took.

Shetrit emphasized this transparency as essential for regulated industries: “If you start talking about some of the largest companies in the world, whether it’s banks or pharmaceutical companies or healthcare companies, it’s unacceptable that you don’t know how these autonomous agents are behaving and what they’re doing, and you can audit and have a few full visibility on what, what the hell is happening in that in that box.”

The system provides “full traceability, auditability and visibility,” allowing IT administrators to set fine-grained permissions controlling which tools each agent can access and what actions they can perform.

Action Agent’s ability to connect with over 600 enterprise tools represents a significant technical achievement. The agent uses Model Context Protocol (MCP), an emerging standard for AI tool integration, but Writer has enhanced it with enterprise-grade controls that address security and governance concerns.

Writer has been working closely with Amazon Web Services and other industry players to bring MCP to enterprise standards. “There’s still place to bring it to enterprise grade,” Shetrit noted, referencing recent issues with MCP implementations at companies like Asana and GitHub.

The company’s approach allows granular control that extends beyond simple user permissions. “It’s not just by a user. It will also have it by the specific agent,” Shetrit explained. “So as an IT persona or a security persona, I have the controls I need to feel comfortable with this data access.”

For example, administrators can permit certain agents to publish messages to Slack while preventing them from deleting messages. “You need that fine grained control, and that’s something we’re baking in as part of the system,” Shetrit said.

The company pre-announces support for over 600 different tools, with each tool offering fine-grained control both at the integration level and for specific agents. This capability allows Action Agent to coordinate work across an organization’s entire technology ecosystem, from customer relationship management systems to financial databases.

Writer’s decision to offer Action Agent free to existing customers challenges traditional software pricing models and reflects broader shifts in the AI industry. The move comes despite the significant computational costs associated with the agent’s extensive token usage.

“Token pricing is extremely problematic when you start thinking about enterprises,” Shetrit explained. “They need a budget line item. They need to figure out the cost structure. This highly variable cost model does not work for these companies, and that is why we’ve been moving away from this for a while now.”

The strategy reflects Writer’s confidence in its cost-efficient model development. The company spent just $700,000 to train its Palmyra X4 model, compared to an estimated $4.6 million for a similarly sized OpenAI model. This efficiency stems from Writer’s use of synthetic data and innovative training techniques that reduce computational requirements.

Writer’s reasoning for the free offering goes beyond competitive positioning. “We think this shows the full value of the ecosystem and the platform, and really starts delivering on the promise of AI,” Shetrit said. Internal users have reported being more excited about this AI product than any previous AI tool they’ve used, including other copilot systems.

Writer’s Action Agent launch escalates competition in the rapidly expanding enterprise AI market, projected to grow from $58 billion to $114 billion by 2027. The company competes directly with Microsoft’s Copilot suite, Google’s enterprise AI offerings, and OpenAI’s business products, but targets a different market segment with its enterprise-first approach.

The competitive positioning reflects a broader industry split between companies building general-purpose AI systems and those focusing specifically on enterprise needs. Writer’s approach prioritizes security, governance, and reliability over raw capability, betting that enterprise customers will choose specialized tools over consumer products adapted for business use.

“Most of their focus is on the consumer realm versus us, which was like, this is not where we’re at,” Shetrit emphasized regarding competitors. “We are fully on the Enterprise B to B side.”

This focus has paid off financially. Writer raised $200 million in Series C funding in November 2024 at a $1.9 billion valuation, nearly quadrupling its previous valuation. The round was co-led by Premji Invest, Radical Ventures, and ICONIQ Growth, with participation from major enterprise players including Salesforce Ventures, Adobe Ventures, and IBM Ventures.

Writer’s vision extends beyond current automation to fundamentally reshape how enterprises operate. The company identifies two clusters of use cases emerging in large organizations: traditional “90% workflow, 10% AI” optimization and new “90% AI, 10% workflow” experiences that unlock entirely new capabilities.

“Each employee will have a thing like this next to them that helps them do their work, helps them automate a lot of it, so you can do much higher leverage work across the organization,” Shetrit predicted.

This transformation addresses a critical shift in enterprise software expectations. As employees become accustomed to sophisticated AI tools in their personal lives, enterprise software must match or exceed that quality. “You cannot afford for enterprise software to not be as good, and in a lot of cases, significantly better,” Shetrit noted.

The shift is already changing internal dynamics at Writer itself. “Historically, as a PM, I can say that execution was the bottleneck. So I can always say no, because I don’t have capacity. Capacity is no longer the bottleneck,” Shetrit explained. When his product managers claim they don’t have time for projects, he now uses Action Agent to generate “at least 80% and 70% and 90% of the work for them so they can start working on it.”

This represents a fundamental change from “scarcity to an abundance mentality” that will require “a lot of retraining element that has to happen within the org.”

Writer’s collaboration with Uber on Action Agent development illustrates how the company leverages customer relationships to improve its technology. Uber’s AI Solutions team provided operational expertise for scaling high-quality annotations across complex enterprise domains, while simultaneously validating the agent’s capabilities in real-world use cases.

“Our collaboration with WRITER allowed us to contribute our deep operational expertise in high-quality data annotation to help shape an agent capable of tackling the most complex enterprise challenges,” said Megha Yethadka, GM and Head of Uber AI Solutions.

This partnership model allows Writer to develop agents that solve actual enterprise problems rather than theoretical use cases. The approach has generated diverse applications across industries, from HR candidate sourcing and securities analysis to clinical trial site selection and competitive intelligence.

Shetrit noted that customer creativity continues to surprise the team: “I’m sure, because that’s the nature of platform and technology, is if we have this conversation again in a week after tomorrow, I’ll have completely different use cases, because our customers will be very, very creative in how they use them.”

Writer plans to expand Action Agent’s capabilities significantly over the coming weeks. The company will add connections to 80 enterprise platforms and third-party data providers like PitchBook and FactSet, enabling access to the full suite of 600+ agent tools.

The rollout strategy reflects lessons learned from enterprise AI deployments. Rather than launching with full capabilities, Writer is starting with core functionality and gradually adding integrations based on customer feedback and real-world testing.

Action Agent is available immediately in beta to Writer’s existing customer base, with a 14-day trial available for new users. The gradual rollout allows the company to refine the system based on enterprise feedback while maintaining the security and reliability standards that regulated industries require.

The launch signals a pivotal moment in the enterprise AI revolution, where autonomous agents are moving from experimental curiosities to mission-critical business tools. As traditional software vendors scramble to add AI features to existing products, Writer’s agent-first approach may determine which companies successfully navigate the transition from human-driven to AI-augmented work.

But perhaps the most telling sign of this shift came from Shetrit himself during the interview: “We will all become, you know, quote, unquote, managers of these fleet of agents, whether they’re humans or synthetic agents.” In this future, the companies that learn to orchestrate AI agents alongside human workers may find themselves with an insurmountable advantage over those still clinging to purely human-driven processes.