ChatGPT's New Boss Bets a Billion Users Want Action

OpenAI is handing ChatGPT’s next act to the engineer who helped turn AI coding into one of its fastest-growing businesses, a signal that the chatbot is being rebuilt less as a Q&A box and more as a task-execution layer.

Thibault Sottiaux was appointed last month as OpenAI’s head of core products, overseeing ChatGPT, Codex, and the effort to combine them into what the company has called a “super app,” according to Wired. That phrase can sound inflated. In OpenAI’s case, it points to a clearer strategy: make ChatGPT the interface people use to act across software, not just ask questions.

Sottiaux’s promotion turns ChatGPT into a Codex-style execution problem

The power shift inside OpenAI matters because Sottiaux comes from the part of the company where AI already proved it can do work, not just talk about work. He helped build Codex, which Wired reports has become one of OpenAI’s fastest-growing revenue streams. Now he has to apply that same operating logic to a consumer product with nearly a billion weekly active users.

That’s a much harder brief. Developers tolerate rough edges if the output saves them time. Consumers don’t. If ChatGPT starts acting on someone’s email, calendar, files, or payments, the margin for confusion shrinks fast.

Sottiaux seems aware of the size of the jump.

“It’s incredibly exciting and mildly terrifying at the same time,” Sottiaux told Wired.

The thesis here is simple: OpenAI’s next ChatGPT won’t be judged mainly by whether it gives smarter answers. It will be judged by whether it can take useful action without making users feel exposed, confused, or out of control.

Codex gave OpenAI a playbook for agents that actually finish tasks

AI coding taught OpenAI that users value systems that can inspect context, plan steps, revise output, and complete work across tools. That is the real bridge between Codex and the ChatGPT overhaul.

Sottiaux joined OpenAI in 2024 after earlier work at Google, Google Maps, and Google DeepMind, where he helped build infrastructure and tools used by researchers working on projects including AlphaGo. At OpenAI, he first worked on tools for researchers, then moved into what became Codex.

The coding product’s appeal is not just that it writes code. It works inside a workflow. It can reason through files, change direction, debug, and respond to concrete user feedback. That is why coding is the obvious test bed for broader agent behavior.

Sottiaux told Wired the super app will largely be powered by Codex, converted into a more general-purpose agent and merged into ChatGPT. That is the technical heart of the story.

The super app bet is less WeChat clone, more universal work interface

OpenAI is borrowing the “super app” label, but its version appears structurally different from the Asian platforms usually associated with the term. Wired notes that super apps such as WeChat bundle messaging, payments, shopping, and other functions into one interface. OpenAI is aiming at something more agentic: one assistant that understands user intent and acts across existing systems.

That distinction matters. OpenAI does not own the consumer infrastructure that WeChat and Alipay built around in China. In the US and other markets, users already have email accounts, calendars, Slack, payment cards, and workplace software. So ChatGPT has to plug into the tools people already use.

OpenAI has already moved in that direction. Wired reports that it expanded a partnership with Visa for agentic payments, and previously built services connecting ChatGPT and Codex to email inboxes, Slack, and calendars. For adjacent fintech context, XOOMAR has also covered how Visa is positioning itself before AI agents hit checkout.

The strongest counterpoint is obvious: plugging into other people’s platforms leaves OpenAI dependent on systems it doesn’t control. That could limit reliability, permissions, and user experience. Still, the direction fits the product logic. If ChatGPT becomes the place where users express intent, the underlying app becomes less visible.

Earlier agents failed to become habits, and OpenAI knows it

The risk is not that OpenAI has never tried agents. The risk is that most users still don’t know when to trust one. Wired points to Operator, launched last year inside ChatGPT to navigate the web on a user’s behalf. It later became ChatGPT Agent, but Wired reports neither product saw significant adoption.

Sottiaux’s diagnosis is that those efforts were “too early.” The models were not reliable enough, so OpenAI had to restrict what they could do. He now says the technology is there.

That claim will need proof in ordinary use. A coding agent can fail in ways a developer understands and repairs. A personal agent that books the wrong reservation, files the wrong expense report, or acts on stale context creates a different kind of damage. The problem shifts from capability to confidence.

Sottiaux described the adoption curve as gradual:

“We have to bring the user along. Initially, maybe it’s a small thing that we can do for you, and then increasingly, build confidence that ChatGPT can do bigger and bigger things.”

That is the right product instinct. Quiet competence will matter more than spectacle.

ChatGPT’s scale creates a design problem, not just a growth opportunity

A product with nearly a billion weekly active users cannot be redesigned like a developer tool. ChatGPT’s reach gives OpenAI distribution, but it also makes every product choice more delicate. A confusing automation feature can hit casual users, enterprise users, students, and developers at once.

The company has already shuttered several stand-alone products, including its Sora video app and an AI platform for scientists, as it redirects resources toward the super app strategy, according to Wired. Several executives who led those teams have left, while Sottiaux’s influence has grown. He now reports directly to Greg Brockman, who is responsible for all OpenAI product teams while Fidji Simo, the company’s CEO of AGI deployment, is on medical leave.

That concentration says OpenAI is choosing integration over a portfolio of separate apps. The benefit is focus. The risk is that ChatGPT becomes overloaded.

OpenAI’s own GPT-5 positioning also points in this direction. In its GPT-5 announcement, the company described a unified system that routes between faster answers and deeper reasoning based on task complexity, tool needs, and user intent. That architecture fits a future where ChatGPT has to decide when to answer, when to think longer, and when to act.

Developers, enterprises, and consumers will grade the redesign differently

Sottiaux has to satisfy three audiences that want different versions of trust. Consumers want ChatGPT to feel personal without feeling invasive. Developers want capable tools, stable behavior, and clear boundaries. Enterprises need controls before they let an AI agent deeper into everyday operations.

Wired’s reporting supports the first two tensions directly. Codex grew with developers because it helped them execute. ChatGPT’s super app pitch now asks regular users to adopt similar habits, but without expecting them to understand how agents work under the hood.

For software teams, this is also why coding remains strategically important. Engineers are early users of agent workflows. The habits that start in code review, repository navigation, and debugging can spread into documents, spreadsheets, research, and operations. Readers tracking that developer-tool fight can see the same pressure in XOOMAR’s coverage of the Cursor vs. Windsurf AI coding split.

Competition is part of the backdrop, but the source only names Google and Anthropic as rivals OpenAI is trying to ward off. The deeper contest is product execution. Benchmarks may get attention, but the winning assistant has to become useful inside messy daily workflows.

The next ChatGPT will win or stall on controlled autonomy

Expect OpenAI to make agent behavior less visible, not more dramatic. Sottiaux said “a lot” of what will be made available in ChatGPT is already available in the Codex app, and OpenAI has said it plans to merge Codex into ChatGPT in the coming weeks. He would only say the super app is coming “soon.”

The most plausible near-term shape is incremental: more app connections, more persistent context, more task follow-through, and more coding-style workflow discipline inside general productivity. OpenAI also prefers small releases, Sottiaux told Wired, because in AI “you can’t really afford to do a big splash and be wrong.”

What would confirm the thesis? ChatGPT starts handling multi-step tasks across email, Slack, calendars, files, and payments while asking for confirmation at the right moments. What would weaken it? Confident mistakes, messy permissions, unclear memory, or users ignoring agent features the way they ignored earlier attempts.

Sottiaux summed up OpenAI’s ambition directly:

“OpenAI is known to take big, bold bets ahead of others, and this is us doing it again.”

The watch item now is whether OpenAI can make that bold bet feel boring in daily use. For an agent, boring is good. It means the work got done.

The Bottom Line

• OpenAI is positioning ChatGPT to become a tool that performs tasks, not just answers prompts.
• Sottiaux’s Codex background signals OpenAI wants to apply lessons from AI coding to mainstream consumer workflows.
• With nearly a billion weekly active users, even small changes to ChatGPT’s role could reshape how people interact with software.