A Continuous AI Environment for Home and Work
A product concept from Alan Bendall | September 2026
I would like OpenAI to explore an assistant that stays with a person across the rooms of a home, a workplace, and the devices they already use. The central product would be a continuous ChatGPT experience with permission to act through compatible software and connected devices. A television with a capable computer inside could be the main screen at home. Small room units would make the same assistant available by voice throughout the house. At work, a computer would provide a full keyboard, screen, and business tools. The conversation and authorized context would follow the person, while home and business data would have separate permissions.
The experience I want
Imagine I run a mechanic shop. At my desk, I ask the assistant to update a parts spreadsheet and check tomorrow’s appointments. While I am in the garage, I ask a follow-up question without repeating which parts I mean. At home that evening, I can check the shop schedule on the TV, continue planning a trip on my phone, and ask a bedroom unit to turn off the lights. If I am away, I can check my thermostat or security system from the same assistant. Each device is another way to reach the same account and the right context for that task.
Today, even a simple request such as playing a specific artist can fail on a smart speaker. I want the easy commands to be dependable and the harder conversations to remember what I meant. The voice and personality should feel comfortable enough to live with every day, and users should be able to choose them.
A computer that helps with the work
A computer built around this environment would be especially useful to the person at the keyboard. I could ask the assistant to organize and edit audio or video, explain a software problem, help build an application, or work through a business plan. It could help manage appointments, inventory, documents, and spreadsheets while showing me exactly what it proposes to change. The user stays in charge of the final work.
How the system could fit together
• Home hub. A high-quality television with substantial computing, speakers, and a clear visual interface becomes the large screen for conversation, entertainment, maps, instructions, and home controls. The same environment could also run on an existing TV or computer for people who do not want new hardware.
• Room units. Compact microphones and speakers provide voice access, music, and nearby device control. Where practical, they could also extend a home Wi-Fi mesh network. All units use the same assistant account instead of behaving like unrelated speakers.
• Connected products. The assistant works with compatible lights, thermostats, locks, cameras, solar equipment, and business applications. OpenAI could build the common experience and partner with manufacturers rather than make every accessory.
• Access everywhere. Phone, computer, vehicle, and room units give the user a consistent way to reach the assistant. The interface changes to suit the device, but conversations and permitted tasks remain available.
An alternative in the car
I would also like to see an OpenAI alternative to CarPlay: a simple dashboard interface that brings the same assistant into a compatible vehicle. By voice, I could pick up a conversation, play exactly the music I ask for, navigate, make a call, or check on home and work. The driving interface should keep eyes on the road, limit distracting tasks, and work with automakers and phone platforms rather than assume every vehicle can be replaced at once.
An ecosystem for software companies
Software makers could offer their applications inside this environment, letting customers ask the assistant to help with real tasks in the tools they already use. I believe companies would pay for useful distribution and integration if the system brings them customers and respects their products. OpenAI could explore developer tools, a marketplace, and commercial terms with partners; that is a potential business model, not a requirement for the first version.
Trust has to be part of the design
Home and work need distinct spaces with clear rules about what can cross between them. People should be able to see and change what the assistant remembers, which devices it can reach, and which actions require confirmation. A room unit needs an obvious microphone mute. Everyday commands such as turning on a light should respond quickly and, where supported by the hardware, keep working during an internet outage. High-impact actions such as unlocking a door, disarming an alarm, or moving money should require stronger authorization.
A practical place to start
OpenAI could first make the continuous assistant work well across existing phones, computers, and TVs, then add a few excellent room units. Home controls and optional mesh networking could follow through compatible partners. A dedicated TV computer could come later if it clearly improves the experience. The first test should be simple: can I begin a task at work, continue it by voice in another room, and see the right result on a screen without setting everything up again?
Why I am sharing this
I am an everyday user, and I have found that a conversation with an assistant becomes much more useful when it can pick up where we left off. I would love to see that continuity become part of the home and the workday. I hope your product and hardware teams will consider this as a direction worth exploring. Seeing an idea like this become real would make me happy, whether or not I receive anything for suggesting it.