OpenAI’s ChatGPT Work aims to make AI agents mainstream for office tasks
Reported by TechCrunch AI: OpenAI Uncovers a Covert Russian Influence Machine Using ChatGPT to Build Fake Expertise and Spread Pro-Moscow Narratives Across Social Media - Yahoo. Analysis and context written by TickrWire.
Evolving story · 3 updatesChatGPT Work agent launchTimeline →OpenAI’s head of product Thibault Sottiaux discusses the launch of ChatGPT Work, a platform for white-collar workers to use AI agents, its $20/month pricing, and the challenges of scaling AI adoption.

- ChatGPT Work is OpenAI’s platform for white-collar workers to use AI agents autonomously, led by head of product Thibault Sottiaux.
- The product launched as part of the $20-per-month ChatGPT Plus plan and has already reached 20 million users.
- OpenAI is reducing costs through efficiency improvements, including an 80% price cut with the Luna update.
- The design philosophy emphasizes minimal interfaces and natural interactions, such as voice and conversation.
- Security and safety remain top priorities, with OpenAI investing heavily in alignment and publishing benchmarks.
OpenAI’s latest push to bring AI agents into everyday workplaces centers on a product called ChatGPT Work, led by Thibault Sottiaux, the company’s head of product. Sottiaux oversees core offerings including the API, agent infrastructure, enterprise solutions, and both the classic ChatGPT interface and the newer Work platform. His team also manages Codex, the AI coding assistant that has become a staple for software engineers. The transition from Codex to ChatGPT Work reflects a deliberate strategy to expand AI’s utility beyond technical users to a much broader audience of white-collar professionals. This shift is not just about accessibility but also about proving that AI can handle complex, autonomous tasks safely and effectively in real-world settings.
The importance of ChatGPT Work lies in its positioning as a bridge between early adopters and the mainstream. Sottiaux emphasized that the product was designed to package AI agent capabilities in a way that feels intuitive, safe, and delightful to use, even on mobile devices. By launching it as part of the $20-per-month ChatGPT Plus plan, OpenAI is betting that the perceived value will justify the cost for users. The company argues that as AI delivers more utility, users will naturally see the subscription as a worthwhile investment. This economic motivation is central to OpenAI’s broader strategy of owning the direct relationship with end users, rather than relying solely on third-party integrations.
Public opinion about AI remains a critical factor in its adoption, and Sottiaux framed ChatGPT Work as a step toward normalizing AI in daily life. He described the product’s evolution as a diffusion process, starting with Codex, which was built for a forgiving technical audience, and now expanding to a general user base. The goal is to demonstrate that AI can do more than assist with writing or advice, it can autonomously complete entire tasks, from processing large documents to generating professional reports and conducting deep research. This shift aligns with OpenAI’s mission to bring everyone along in the AI revolution, not just early adopters.
The design philosophy behind ChatGPT Work emphasizes minimalism and natural interaction. Sottiaux explained that the team’s approach is to build highly capable models and then create interfaces that get out of the user’s way, allowing the AI to express its full potential. This philosophy is reflected in the product’s evolution, including the recent growth of ChatGPT Voice, which enables users to engage with the AI through natural conversation. The idea is to adapt the technology to human behavior, rather than forcing users to adapt to the technology. This user-centric approach is intended to reduce friction and make AI feel like a seamless part of daily workflows.
Adoption metrics suggest that OpenAI’s strategy is resonating. The company announced that ChatGPT Work has already reached 20 million users, a figure that underscores the growing appetite for AI-powered tools in professional settings. However, the product’s design also raises questions about how much autonomy users are comfortable ceding to AI. Analysts like Ethan Mollick, a Wharton professor who studies AI tools, have noted that ChatGPT Work aims for a “magic” level of automation, while competitors like Anthropic’s Claude Cowork take a more measured approach by presenting users with A/B tests and requiring more active decision-making. Sottiaux countered that the world appears ready for this level of automation, pointing to the rapid adoption as evidence.
The product’s development is iterative, with OpenAI continuously discovering new capabilities and refining the platform based on real-world use. For example, the release of GPT 5.6 marked a significant step forward, enabling the AI to process large volumes of documents, generate high-quality slides and reports, and conduct deep research, tasks that professionals routinely perform. This iterative process involves capturing feedback from users and the broader community to improve the product over time. The approach reflects a broader trend in AI development, where products evolve alongside the technology itself, with each iteration uncovering new possibilities and challenges.
Cost efficiency is another critical aspect of ChatGPT Work’s strategy. Sottiaux acknowledged that users often face a gap between the value they perceive and the tokens they consume, but OpenAI is actively working to address this through efficiency improvements. The company recently announced major price cuts, including an 80% reduction with the Luna update, which it describes as a permanent correction to make frontier capabilities more affordable. The goal is to ensure that users can access more utility over time without increasing costs, or even while spending less. This strategy is designed to make AI more accessible to a wider audience while maintaining profitability.
Security and trust are paramount concerns for users, particularly when granting AI access to sensitive data like emails or messages. Sottiaux emphasized that OpenAI’s investment in safety and alignment is a core part of its approach. The company publishes honest benchmarks on safety and alignment, and its models are designed to meet high standards in these areas. This focus on safety is intended to reassure users that they can trust the AI with their data, even as the product’s capabilities expand. The challenge, however, remains in balancing openness with security, especially as AI agents become more integrated into daily workflows.
Looking ahead, the trajectory of ChatGPT Work will depend on several factors. The product’s ability to deliver on its promises of autonomy, safety, and cost efficiency will be critical in winning over skeptics. Competitors are also advancing their own AI agent platforms, and the race to define the standard for white-collar AI tools is far from over. For developers, the product’s API and agent infrastructure offer opportunities to build on top of OpenAI’s work, while businesses will need to evaluate whether the $20 monthly plan delivers sufficient ROI. Investors, meanwhile, will watch closely to see if OpenAI can sustain its growth and monetization strategies as the AI market matures.
OpenAI’s agent infrastructure and API for ChatGPT Work provide tools to build on top of the platform.
The $20 monthly plan and autonomous task completion could drive productivity gains and cost savings.
Rapid user adoption and pricing strategies signal potential for sustained growth and monetization.
AI agents are becoming more accessible to non-technical users, reshaping how people work.
- Codex
- OpenAI’s AI coding assistant designed for software engineers, now part of the broader ChatGPT ecosystem.
- ChatGPT Work
- OpenAI’s platform for white-collar workers to use AI agents for autonomous task completion.
- Agent infrastructure
- The backend systems that enable AI agents to perform tasks autonomously within OpenAI’s products.
AI bias estimate: The source emphasizes OpenAI’s positive framing of AI adoption and downplays skepticism, such as concerns about token costs or user readiness. (Automated estimate, not a definitive judgement.)
- OpenAI Uncovers a Covert Russian Influence Machine Using ChatGPT to Build Fake Expertise and Spread Pro-Moscow Narratives Across Social Media - Yahoo ↗
- OpenAI restores 5-hour Codex and Work limits for ChatGPT Plus users ↗
- ‘The world seems to be ready’: An interview with OpenAI head of product Thibault Sottiaux ↗
- OpenAI bans Russian ChatGPT accounts in influence campaign - qz.com ↗
- OpenAI bans Russian ChatGPT accounts used in covert misinformation campaign - CNBC ↗
- Stampli cuts launch hours by 68% using ChatGPT Work ↗
AI ToolsAccel-backed Keenable is indexing the web for AI agents
AI ToolsNvidia says its Groq 3 LPX is four times faster than Cerebras, but the math is more complicated
AI ToolsAn AI boss fired its first employee but only after humans reminded it of its own rules
AI ToolsVercel Introduces ‘Is Agentic’, a Free Agent-Readiness Scoring Tool That Audits Public Websites Using Ora’s 100+ Checks
AI ToolsDecoding AI’s Open-Source Course Maps Three Ways to Run an Agent Loop and the Provider Economics Behind Each
HardwareOpenAI’s Jalapeño chip is built for fast inference at scale, benchmarks show
OpenAI shared initial benchmark data for its Jalapeño inference chip at the Hot Chips conference, demonstrating higher token generation and energy efficiency than current market standards.
SecurityUkraine opens its massive labeled battlefield dataset to British firms in a landmark AI weapons partnership
Ukraine has granted the United Kingdom access to its Avengers Labs platform, providing foreign tech companies with millions of annotated combat images to train military artificial intelligence.
RoboticsI spent a day at a robot “carnival” in Shanghai. Here’s what I saw.
A recent robotics festival in Shanghai highlighted China's rapid commercial progress in humanoid systems, where local firms now dominate global delivery numbers.
SecurityTaiwanese cybersecurity firm warns that AI tools have more than doubled Chinese state-backed cyberattacks
Taiwanese cybersecurity firm TeamT5 reports that Chinese state-backed hacking groups have more than doubled their attack frequency after adopting AI models such as DeepSeek.

Jalapeño’s first results show industry-leading speed and efficiency in AI inference
OpenAI’s Jalapeño custom inference chip delivers 1.5‑3.6× better power efficiency and lower latency across major language models, outperforming commercial competitors.

Disrupting a new covert influence campaign from Russia
OpenAI blocked Russian ChatGPT accounts that were used to push a fabricated Israeli think‑tank and a pro‑Russia sovereignty index, exposing a covert influence operation.