AT&T is piloting a new AI-driven assistant aimed at screening incoming phone calls. This innovative feature, which the company describes as a “digital receptionist,” is set to be available to a limited number of customers this year. The assistant utilizes data from AT&T’s network, including call history, to identify and filter out spam calls.
Andy Markus, AT&T’s chief data officer, shared with Technology News that the current functionality is merely the starting point. He envisions that the assistant could evolve to assist users with tasks such as making reservations or scheduling appointments, akin to the capabilities of Google’s AI assistant.
The AI assistant will answer calls from unknown numbers and inquire about the caller’s identity and purpose. If the assistant determines that the caller is human and expresses urgency, it will connect the call. Otherwise, it can either record a message or terminate the call.
The AI system will utilize network data to recognize frequent contacts. Users will also have the option to create a “Do Not Screen” list, which will enable calls from important but less frequently used numbers—such as a doctor’s office or a child’s school—to bypass the screening process.
Markus emphasized that the integration of the AI assistant within the carrier’s network distinguishes it from other call-screening services like those offered by Google and Apple. He noted, “Using the network, it understands your calling patterns as a consumer.” For instance, if a user’s sibling frequently calls, that call will be prioritized based on established interactions.
In contrast, Google’s Call Screen feature relies on the user’s contact list and analyzes the content of calls to identify known spam. It compares incoming numbers against a database of recognized spammers and robocallers.
AT&T’s assistant is designed to function independently of specific devices. Users can view a real-time transcript of the conversation through the AT&T app, allowing them to monitor the dialogue or intervene at any moment. If the assistant handles the call, it can provide a summary of its findings afterward.
