OpenAI released ChatGPT Work last month as a $20-per-month tool designed to give non-technical workers the same autonomous AI agent capabilities that software engineers already use for coding. The product represents one of the company's biggest bets on expanding AI beyond the engineering department, but early data reveals a stark adoption gap: while 98% of OpenAI's own employees use the underlying Codex tool, fewer than 1% of individual subscribers and just 17% of organizational subscribers have adopted the agentic version.

What Makes AI Agents Different From Regular Chatbots?

An AI agent isn't just a tool that answers questions. It's a system that can autonomously complete multi-step tasks over longer periods of time, accessing your email, Slack, spreadsheets, and other digital tools to accomplish real work. For software developers, this shift happened years ago through command-line interfaces that let AI write code. For everyone else, the transition has been much slower.

Andrew Ambrosino, the lead engineer for OpenAI's desktop app, explained the trust required to make agents useful. "If I'm asking it to write a document, is there a possibility that it's going to pull from a private DM on that subject and not know that it's not supposed to share some info? Yes," he said. "I'll do it for the job. I will take the personal hit here and there if I have to. And I haven't had to."

"If I'm asking it to write a document, is there a possibility that it's going to pull from a private DM on that subject and not know that it's not supposed to share some info? Yes. I'll do it for the job. I will take the personal hit here and there if I have to. And I haven't had to."

Andrew Ambrosino, Lead Engineer for OpenAI's Desktop App

Why Is the Adoption Gap So Dramatic?

The problem isn't that non-engineers don't want AI help. It's that the tools weren't built for them. When accountants and finance teams first started using Codex, the interface was designed for programmers, showing technical readouts about code changes that made no sense to non-technical workers. OpenAI spent months rebuilding the experience to be more general-purpose and user-friendly.

Thibault Sottiaux, who leads OpenAI's core product work including ChatGPT Work, emphasized the company's mission: "In this new factor, ChatGPT can actually do entire, very complicated tasks for you all autonomously in a way that is delightful and safe. It's the very mission of OpenAI, to bring everyone along."

"In this new factor, ChatGPT can actually do entire, very complicated tasks for you all autonomously in a way that is delightful and safe. It's the very mission of OpenAI, to bring everyone along."

Thibault Sottiaux, Product Lead at OpenAI

The disconnect between internal adoption and external adoption matters enormously for OpenAI's business model. Agents that work autonomously for longer stretches consume more tokens, the units of text that AI models process. More token consumption means higher revenue per user. But reaching new professions is crucial if the company is to justify its massive investments in training and computing infrastructure.

How Is OpenAI Trying to Close the Adoption Gap?

OpenAI is taking several approaches to make AI agents accessible to non-engineers:

• Simplified Interfaces: The team is building experiences that abstract away technical complexity, letting users simply describe what they want without needing to understand how the AI accomplishes it, similar to how "vibe coding" tools let users skip actual programming syntax.
• Visual Discoverability: ChatGPT Work includes buttons and visual elements to help users discover features, even though the long-term goal is for users to ask the model directly. Ambrosino compared this to skeuomorphism, the design practice of making digital tools look like physical objects they replaced, which actually helped people transition to new technology.
• Real-World Use Cases: OpenAI is focusing on routine, data-intensive coordination tasks where agents excel, such as assembling investment memos from communications and analysis, creating weekly metrics reports, and turning spreadsheets into planning tools.

Akshay Nathan, who leads the product engineering team at OpenAI, explained the value proposition: "There is a deluge of information for the average worker or employee of any of these companies, including myself. We're actually quite limited by our ability to parse everything that's available to us, and then take action on it. That information lives in all these system records tools like Salesforce. The value of ChatGPT is you already have access to this, but now you truly have access to it."

"There is a deluge of information for the average worker or employee of any of these companies, including myself. We're actually quite limited by our ability to parse everything that's available to us, and then take action on it. The value of ChatGPT is you already have access to this, but now you truly have access to it."

Akshay Nathan, Product Engineering Lead at OpenAI

What's at Stake for the Broader AI Industry?

OpenAI's challenge reflects a broader industry problem. Christian Catalini, writing on Andreessen Horowitz's "It's time to build" blog, warned that "if the labs cannot rapidly get ahold of the key complementary assets needed to scale AI in the market, value will accrue elsewhere." This means that if OpenAI and other large AI labs cannot expand beyond software engineering, specialized competitors like Harvey (for legal work) and Clay (for sales) may capture those markets instead, using whichever AI model works best at the time.

The stakes are high. Software engineering represents a tiny fraction of professional work. If AI companies are to justify their investments in training and computation, they need to enable AI agents across accounting, medicine, law, sales, operations, and dozens of other fields. Right now, ChatGPT Work is used by just 20 million people, compared to more than a billion users who use ChatGPT for basic prompting.

Sottiaux acknowledged the value equation: "The more value and the more utility that we generate for users, the more they will be willing to also pay for some part of that utility, and that's how we've always seen ChatGPT as well. You sit there and you're like, 'of course I want to pay $20 bucks a month for this,' because the value that you get is so much more."

Sottiaux

Whether OpenAI can bridge the adoption gap will determine not just the company's growth trajectory, but whether AI agents become a transformative tool for the broader workforce or remain a niche capability for software developers.