Skip to content

Meta lets its agent Muse make phone calls: automating queue lists sparks a clash between software and reception staff

Share
Meta lets its agent Muse make phone calls: automating queue lists sparks a clash between software and reception staff

Listen to this article

Read by Anchor

Meta has opened the feature that allows its intelligent agent Muse to make direct phone calls to U.S. companies and institutions, limited to paid-plan subscribers, after an initial test period that began in September and involved a small group of early users. The move, announced by Alexander Wang, head of AI affairs at the company and head of its superintelligence labs, is part of an effort to turn the agent, whose user base exceeded ten million last month, into a daily consumer product that handles arduous routine tasks such as calling insurance companies, cable providers, vehicle registration offices, and reserving tables at restaurants that do not offer digital booking.

Unlimited software patience against queuesThis is the core promise the tool offers to ease the accumulated frustration of consumers with tangled automated-response trees and long hold times while waiting for a human representative who may ultimately transfer the caller to another department. However, the feature comes with direct pricing that reflects inference and call costs, with Muse subscriptions starting at $20 per month for 500 million tokens weekly and rising to $100 per month for power users for 3 billion tokens, while the option to create an email address for a Muse agent remains a feature that will be offered soon to free-tier users.

Meta’s agent is not the only one that can call commercial establishments; the startup Instinct and Google’s Gemini system also provide similar services. Yet field experience shows mixed reactions, as many reception staff and companies rush to end the call and hang up once they realize the speaker is not human. As voice-simulation technologies continue to become more accurate and convincing, distinguishing a human caller from the generative agent representing it becomes extremely difficult for front-line staff.

The impact of this shift directly touches the architecture of call centers and support services in the regionIn Egypt and the Levant, where the outsourcing and business-process-outsourcing sector is a vital engine serving global and regional markets, customer-service employees will gradually face automated agents with virtually unlimited negotiating and inquiry capabilities, prompting updates to work policies and the development of tools to verify caller identities to avoid draining human staff time. Conversely, this scenario places the banking, telecommunications, and services sectors in the Gulf, who have heavily invested in automating responses to client inquiries, in front of a new equation: automation is no longer merely a cost-cutting tool for institutions, as consumers now possess a counterpart tool to breach waiting-line channels, turning the customer-service space into a direct meeting point between two software programs negotiating on behalf of both parties.

Limiting the voice-call feature to paid subscriptions reveals that turning AI agents into telephone-task executors remains subject to computing and network costs, making it a paid-for utility rather than a universally available tool, pending the pressures of usage that may shape corporate communication standards.

Don't miss the next story

Subscribe for updates