Digital Transformation · 30 September 2026
OpenAI launches Dots, a persistent agentic AI avatar
OpenAI has introduced Dots, an always-on agentic avatar that pursues user-set goals across devices and apps with minimal oversight, marking a shift away from single-session chatbots.
What happened
OpenAI has introduced Dots, an agentic avatar it describes as operating independently of any single device, app or screen. According to TechCrunch, Dots is designed to run continuously in the background, working towards goals that users set for it rather than waiting for a fresh prompt each time it is needed.
The defining feature is persistence with minimal oversight: rather than a chatbot confined to a single conversation window, Dots is pitched as a standing presence that keeps pursuing a user's objectives across contexts, checking in only as required.
Why it matters
Dots signals a further move by OpenAI away from single-session, prompt-and-response assistants towards agents that act with a degree of autonomy over time. For organisations building on top of OpenAI's tools, this changes the design problem: instead of optimising a single interaction, teams now need to think about how an agent behaves continuously, across multiple touchpoints, with a human checking in only intermittently.
That shift has real implications for how AI gets embedded into products and services. Interfaces, permissions and escalation paths all need rethinking when the "user" of a system might be an always-on agent rather than a person actively driving each step.
The Renascence take
An avatar that pursues goals with "minimal oversight" is as much a trust-design challenge as it is a technical one. The more autonomy an agent is given, the more the burden shifts to designers to make its intentions, progress and boundaries legible to the person who set it loose.
The interesting question isn't what Dots can do unsupervised — it's what happens the first time it gets something wrong while nobody's watching. Autonomy only earns trust when it comes with visible checkpoints: clear signals of what the agent is doing, why, and how a user can interrupt or correct it before drift becomes damage. Any organisation building on agentic tools like this should treat "minimal oversight" as a design brief, not a selling point — the real work is engineering the moments where a human glance is still worth more than full automation.
Sources
This briefing was written by our Newsdesk, synthesising reporting from the outlets below. Follow the links for the original coverage.
FAQ
Questions we get on this topic
More in Digital Transformation
Stay ahead of CX
Get the signal, not the noise.
The stories shaping customer experience — plus the Journal and Experience Loom — in your inbox.
