Follow ZDNET: Add us as a favorite source On Google.
Key takeaways of zdnet
- The AT&T’s digital receptionist will use AI to screen your call.
- The goal is to determine whether a call is valid or spam.
- This feature will roll as a test for select AT and T customers this year.
How often do you receive calls from an unknown number and argue whether to answer it? Spam calls continue to plague all of us, often hesitate to answer our phone until we know who is making calls.
Now AT & T is turning to AI in an attempt to save us from spamors and scammers who tie our phone.
Too: I unlocked the menu hidden with 14 secret codes on my iPhone and Android
Roll as a test to select customers in 2025, the new AT and T digital receptionist will serve as an automatic call screener.
Instead you are trying to decide what to answer, this assistant will pick up the call itself. Using Voice-to-Voice and Agentic AI skills, the receptionist will then ask questions to determine whether the collar is human or bot, if the call is immediate, and what it meets the criteria you install.
How call screening works
It is mentioned here how such a scenario will play, as described by AT&T’s Chief Data Officer Andy Marcus. Tuesday blog post,
The digital receptionist will answer the call coming from you, presenting such questions like “Who can I say?” Or “what is this?” If the collar examines the right box, the call passes through you. If the call worries something that AI is equipped to handle itself, such as taking a message or accepting delivery, it may be able to complete the conversation automatically.
Too: How to clean your Android phone cache (and hidden problem it really recover)
Okay, but what if the collar refuses to identify itself, it is a wrong number, or the call does not meet your criteria? In those cases, the receptionist will either hang or take a message. As the initial conversation plays out, you can see a live transcript of the conversation and choose to take a call at any point. You can also finish the call, see a summary when it ends, and then decide if you want to return the call.
Naturally, you do not want every call to be displayed in this way, especially from family, friends and other contacts. For that, you are able to add people to the “Do Not Screen” list so that their calls are automatically kept without AI screening.
Testing phase
For now, the receptionist is definitely in the test phase, which means that the AT&T is still changing its abilities and adding new features. But if all goes well, the carrier becomes even more AI-managed below the road.
Too: How to clean your iPhone cash (and why you should do it before updating iOS 26)
As an example presented by Marcus, a future AI agent can automatically connect you to a restaurant to make reservation on your behalf. To shut down this, you will tell AI the name and time for the restaurant’s name and reservation. The agent will not only connect you, but can also book reservation on its own.
Of course, we know that today’s AI bots are perfect. They are prone to mistakes. They can confuse or wrong the directions. They often fail to understand the context or meaning of our words. So this is a test, to see if and how an AI-based receptionist can be helpful without running into too much hiccups.
AI anxiety
To solve concerns about such AI, Marcus also explained the technique behind it.
For the receptionist, the AT & T is using several large language models (LLM) to process, create reactions and then convert those reactions into speech. The prevention of fraud and spam-fighting algorithms, designed to seek spam or fraud pattern, is designed to explore the call if such signs have been detected if it is detected. The company also promises that your personal information is safe and it is used only to help the receptionist determine how to handle the call.
Too: In 2 years, half of all service calls will be resolved by AI – Survey
Spam call filters and features are already available from OS and app developers along with AT&T and other carriers. Some of them are also exploiting in AI. But with such filters, the burden is often to find out whether to answer the call. When successful, the digital receptionist of AT & T can automatically be a useful next step in determining how each call is handled.
