OpenAI is not building a chatbot upgrade. It’s building a digital employee — one that files your expense reports, drafts your emails, and, yes, places your dinner order without you lifting a finger. That’s the pitch behind Dot, the company’s new agentic product that The Verge got an early hands-on with, and the ambition here is considerably larger than anything ChatGPT has done before. For anyone watching OpenAI expand beyond a simple prompt-and-response interface, ChatGPT features keep stacking up fast — and Dot may be the most consequential one yet.
According to The Verge’s hands-on, Dot is designed to operate as a persistent, memory-enabled agent that lives inside ChatGPT and can act on your behalf across a wide range of tasks — from navigating enterprise software and scheduling meetings to placing food delivery orders through third-party services. It’s the kind of thing that sounds like a party trick until you realize it’s wired into your entire workflow.
Enterprise Backbone, Personal Touch
The core use case for Dot is enterprise productivity. The agent can connect to workplace tools, parse documents, manage tasks across applications, and take autonomous action without waiting for a user to prompt it step by step. Think of it less as a question-answering assistant and more as a background process that monitors your context and executes tasks when conditions are met. OpenAI is explicitly positioning this as a tool for business users who need AI that operates across systems, not just within a chat window.
That enterprise angle is reinforced by the pricing. As Fortune reported, OpenAI unveiled Dot alongside a $500-per-month plan targeted at power users and business customers — a significant jump from existing tiers and a clear signal about who the primary audience is. At that price point, OpenAI is not selling a novelty. It’s selling infrastructure, competing directly with workflow automation tools, enterprise copilots, and the broader category of AI-powered productivity software that Microsoft, Google, and Salesforce have all been racing to build out.
Dinner Orders and the Blur Between Work and Life
Here’s where Dot gets genuinely interesting — and a little unsettling, depending on your comfort level with AI autonomy. The agent doesn’t clock out when your workday ends. It can place food delivery orders, handle personal scheduling, and manage tasks that have nothing to do with your job. That dual-mode capability — rigorous enterprise tool by day, personal assistant by night — is what separates Dot from most business-focused AI products, which tend to draw a firm line between professional and personal contexts.
The design philosophy here mirrors what OpenAI has been building toward for months: AI that doesn’t wait to be asked. Dot’s memory layer means it accumulates context over time, learning your preferences, your calendar patterns, your recurring needs — and acting on them proactively. That is a fundamentally different relationship with software than anything most enterprise customers are used to. It also raises real questions about data handling, permissions boundaries, and what happens when an autonomous agent makes a mistake with your credit card or your calendar. OpenAI hasn’t fully answered those questions publicly yet, but the product exists and the capability is real. As OpenAI investment continues to pour in at a scale few companies have ever seen, the pressure to ship transformative products — not incremental ones — is enormous. Dot looks like the company’s answer to that pressure.
