返回新聞
產品展示AI Understanding 簡報

Instinct和Meta Muse新增AI座席電話通話功能

TechCrunch 報告稱,Instinct 和 Meta 的 Muse 都引入了允許其 AI 代理撥打電話的功能,這標誌著面向消費者的自主代理功能的顯著擴展。

4 min readRead the original reporting
Source-provided image accompanying Instinct and Meta Muse add AI agent phone calling features
歸因報告來源記錄
出版商
techcrunch.com
來源連結
techcrunch.comhttps://techcrunch.com/2026/09/17/rival-ai-agents-instinct-and-metas-muse-both-add-the-ability-to-make-calls/
來源類型
新聞媒體的報道-不是第一方文件。

我們無法獨立確認的內容: 此聲明歸因於指定的商店。我們沒有根據第一方文件對其進行驗證。 (techcrunch.com)

背景60 秒內了解這一點

從這裡開始

關鍵術語

人工智慧代理
一種可以觀察、推理並採取行動來實現目標的軟體系統,通常使用工具和記憶體。
特點
模型用來進行預測的輸入變數。
測試一下自己AI 代理測驗

發生了什麼事

Instinct and Meta's Muse have both launched new features enabling their AI agents to place phone calls. Instinct introduced 'Instinct Concierge,' which allows agents to handle high-touch tasks like booking restaurants without online reservations or managing cable bills, currently in early access. Meta expanded its Muse beta to allow outbound calls to U.S. businesses, prioritizing users who had previously requested the .

TechCrunch reports that Instinct, a San Francisco-based startup, has added a new calling called Instinct Concierge. According to founder Noah Shinn, the feature allows agents to perform high-touch tasks such as booking restaurants that do not accept online reservations, joining cancellation lists for dentists, and resolving issues with cable bills. The feature is currently rolling out in early access to a subset of users, with plans to expand to all users in the future.

Simultaneously, Meta’s Muse AI assistant has gained the ability to place outbound calls to U.S. businesses. Meta stated that this is being rolled out first to users who had previously asked Muse to make a call, using the request mechanism as a way to prioritize early adopters. This move brings Meta’s agent into parity with rivals who had previously touted their calling capabilities as a competitive advantage over Instinct.

The article notes that Instinct has seen significant traction, having raised $350 million at a $2.5 billion valuation last month and reportedly being in talks to raise $1 billion at a $10 billion valuation. Additionally, Instinct recently rolled out email address capabilities for account management and allowed users to connect their assistants to trusted networks for task management. Sensor Tower data cited in the report indicates that Muse outperformed the initial Meta AI app launch, with over 730,000 U.S. downloads in its first few days compared to 707,000 for the Meta AI app.

來源詳情: techcrunch.com ↗

為什麼這很重要

This development signals a shift from text-based interaction to real-world action execution for AI agents. By enabling phone calls, these tools can now navigate services that lack digital interfaces, such as small businesses or legacy systems. This parity between a high-valuation startup and a major tech giant intensifies competition in the consumer market and expands the practical utility of these assistants beyond simple chat or email management.

The addition of phone calling capabilities represents a critical step in the evolution of AI agents from passive information processors to active task executors. Many real-world services, particularly small businesses and legacy systems, still rely heavily on telephone interactions. By enabling agents to navigate these interfaces, Instinct and Meta are addressing a significant gap in the utility of current AI assistants.

This development also highlights the intensifying competition in the consumer market. With Instinct reportedly valuing itself at $10 billion and Meta leveraging its massive user base, the race to provide the most capable and versatile agent is accelerating. The fact that both companies have independently added similar features suggests that phone interaction is becoming a standard expectation for advanced AI agents.

From a practical standpoint, this could significantly reduce the friction for users trying to manage complex, multi-step tasks that require human-like negotiation or persistence. However, it also raises questions about the reliability and safety of autonomous agents making decisions and commitments on behalf of users in real-time conversations.

Interactive Mechanism

互動機制:它實際上是如何運作的

以互動方式探索這項發展背後的基礎技術。

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
互動式概念檢查+10 Points
AI Agents Quiz

An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?

接下來看什麼

Monitor the expansion of Instinct Concierge from early access to general availability and track whether Meta extends Muse's calling capabilities to international numbers or more complex conversational scenarios. Also watch for user feedback on the reliability and safety of autonomous phone interactions.

The next phase of Instinct Concierge's rollout will be crucial to determine if the can scale reliably. Early access users will likely provide feedback on the agent's ability to handle edge cases, such as complex hold times or ambiguous requests, which will inform future updates.

Meta's approach to prioritizing users who requested the suggests a focus on user-driven development. Watching how Meta expands the calling feature to other regions or types of businesses will indicate its long-term strategy for Muse.

The competitive landscape may see other providers, such as Wajo, Town, and Ollie, responding with similar enhancements to maintain parity. The market for text-based assistants is described as 'hot,' and the ability to execute real-world actions like phone calls is likely to become a key differentiator.

相關指引和測驗

人工智慧代理什麼是人工智慧?AI 的未來測試你所知道的—嘗試免費的人工智慧測驗在我們的詞彙表中尋找人工智慧術語關注 AI 模型發布追蹤器
覺得有用嗎?