Shenzhen Startup Seeks Feedback on Phone-Call 'Assisted Driving' Device

The proposed product combines hardware and software. A magnetic device attaches to the back of a phone and connects via Bluetooth to obtain call audio. The hardware has buttons, a screen and vibration functions for quick operations, status checks and feedback. A companion app is used to view real-time conversations, translations and reference prompts. The team said it is still refining the product and does not name itself in the post.

In one scenario, a user is in a meeting when a delivery worker calls. The device can answer on the user's behalf, and AI voice can relay a pre-recorded instruction such as asking the worker to leave the delivery on a table near the security room. If that spot is full, the app can suggest quick replies based on the current conversation, which the user confirms before they are sent.

For a one-person company, the team proposes an "AI front desk" that receives calls from customers, answers common questions using pre-supplied business materials, and records the caller's needs and contact details for later follow-up. A marketing specialist inviting dozens of clients to a launch event could hand basic event introductions and attendance-intent registration to AI, leaving more time for clients who need personal communication. During these AI-handled calls, the user can follow a live text transcript in the app and take over at any time.

The team is also considering more difficult calls, such as business and cross-language conversations. In an online meeting, if a client asks about delivery terms, AI could find a previous agreement. If the client raises an objection, the system could suggest what concerns to clarify, how to explain a position more clearly, or what alternatives to discuss. If the conversation turns to scheduling a next meeting, it could check the calendar for free slots so the user can settle the time on the spot. These prompts would appear on the phone screen, and the user decides how to respond.

In cross-language calls, the app could display real-time translations so the user can check terminology, numbers and specific requirements. The user could also speak in a familiar language, with AI translating and conveying the message by voice to the other party.

The team says its ideas are not limited to the call itself. After a call ends, it wants to bring raw recordings, summaries and to-do items together, extracting confirmed arrangements and unresolved issues. It said further functions, such as reviewing new requirements against project materials or flagging possible omissions, depend on learning how users actually work before deciding what is worth developing. It is also considering whether the product should extend to WeChat voice calls, Feishu meetings and face-to-face conversations, and which of those should be supported first.

A practical question remains: whether users are willing to carry an extra device on the back of their phone to get these functions. The team noted that extra hardware means additional weight, thickness, connection and charging burdens, while different phones, calling apps and earphones may affect the experience. To gather demand, it has prepared a questionnaire and will select 40 thoughtful responses, each receiving a 50-yuan JD card.