Rival AI agents, Instinct and Meta’s Muse, both add the ability to make calls

Instinct launched an early-access Concierge calling feature that can book offline restaurant reservations, manage dental cancellation lists and resolve service issues, matching a newly introduced call-to-U.S.-businesses capability at Meta's Muse. Instinct raised $350 million at a $2.5 billion valuation last month and is reportedly discussing a $1 billion round at a $10 billion valuation. Meta's Muse also showed strong initial adoption, with more than 730,000 U.S. downloads versus 707,000 for Meta AI over their respective first five days.

Analysis

For META, agentic calling is strategically more relevant as a commerce-intent and local-advertising data layer than as a near-term revenue product. Successful completion of offline tasks can create proprietary signals around purchase intent, merchant responsiveness and service availability, potentially improving click-to-message and WhatsApp business monetization over 6-18 months. The direct P&L effect remains immaterial versus META's ad base; the market should not capitalize early download velocity without evidence of retained weekly users, task-completion rates and incremental business messaging revenue.

The competitive bottleneck is shifting from model quality to permissioning, identity, liability and unit economics. Calling workflows require consent architecture, fraud controls, escalation to humans and sufficiently low inference/telephony cost per completed task; failures here could turn a consumer-growth feature into a trust and regulatory expense. This favors platforms with authenticated social graphs, business communication endpoints and distribution, but also raises scrutiny around automated-call rules, recording disclosures and deceptive-agent behavior.

Consensus may overstate feature parity: availability of a calling capability does not establish a durable moat. The defensible advantage will accrue to the platform that converts agents into a closed-loop transaction surface with merchant payments, customer service resolution and measurable advertiser ROI. META has the best path to distribution-led monetization, while venture-backed standalone assistants face escalating customer-acquisition and compute costs if large platforms bundle equivalent functionality at zero price.

Key Decisions for Investors

• Maintain/accumulate META on broad AI-product pullbacks rather than chase feature headlines; assess the thesis over 6-18 months through WhatsApp Business revenue acceleration, click-to-message pricing and disclosed AI engagement. A material rise in AI infrastructure capex without corresponding ad pricing or engagement improvement would falsify the monetization case.
• Do not establish a standalone META trade solely on this launch over the next 1-3 months: the likely earnings impact is below reporting materiality, while regulatory or reliability incidents can dominate near-term sentiment.
• Monitor META's incremental capex and depreciation guidance at the next earnings release. If management frames agent usage as meaningfully increasing inference demand while keeping monetization timelines vague, reduce AI-premium exposure; multiple compression risk is greater than feature-driven revenue upside in that scenario.
• For private-market exposure, avoid underwriting standalone consumer-agent valuations on download growth alone. Require cohort retention, completed-task cost, paid conversion and merchant-side revenue evidence; bundled competition from META and other platform incumbents can compress standalone monetization before scale is achieved.