Phonic unveils Phonic-1, a voice AI that starts processing before the user finishes speaking
IA en un minuto newsroom · Editor: Jon Elgezabal
In 30 seconds
Phonic is unveiling Phonic-1, a speech-to-speech model that starts running tools as the speaker keeps on talking. In the company's demo it answers 1.1 seconds after the user stops, against 15.7 seconds for GPT-Live in the same comparison. Phonic itself, which says it is available in more than 50 languages, also announces it has raised $27.5 million in a Series A.
Phonic is unveiling Phonic-1, a voice model that begins working while the person is still talking.
The company calls it a "speech-to-speech" model (one that handles voice directly, without converting it to text) built for agent tasks. In its demo, Phonic-1 makes five calls to outside tools as the user keeps talking and answers 1.1 seconds after they stop. In the same comparison, GPT-Live answers after 15.7 seconds.
Phonic says it is trained on more than 50 languages and is available now. These figures come from the company itself: in the preference test carried out by the firm Podonos, Phonic-1 gets 61.5% against 32.0% for GPT-Live-1, and human voice actors beat it (57.0% against 32.0%).
Phonic also announced a $27.5 million Series A round, led by First Round.
Why it matters · analysis and opinion
In voice, the delay between someone falling silent and the system answering is what makes a conversation awkward, and Phonic goes straight at that by starting work before the sentence ends. For a company with phone support, appointments or voice orders, the idea is interesting because the assistant could check a calendar or an inventory without long silences. But every comparison comes from the company itself, including the preference test, and the fact that human actors still win is a reminder that naturalness is not solved yet. Before signing up for anything, it is worth trying it with real calls in the customers' language, measuring how often it makes mistakes when acting on data and deciding which tasks should never be left to the machine alone.
Source: Phonic · Written with the help of AI: how we make the news
