Bengaluru-based Emergent — the vibe-coding startup that hit 1.5 million monthly active users by letting non-technical founders build full-stack apps via natural language — is now trying to run those apps for you. Its new product, Wingman, is an autonomous AI agent that operates through WhatsApp, Telegram, and iMessage, handling routine tasks across email, calendars, and workplace tools without requiring users to touch a new interface.
What's new
Wingman is a messaging-first AI agent designed to sit inside chat platforms users already live in. It runs background tasks autonomously but introduces what Emergent calls "trust boundaries" — a tiered approval system where low-stakes actions execute automatically and higher-consequence steps require explicit user sign-off. The launch positions Emergent directly against OpenClaw and Anthropic's Claude-based agent systems, expanding the company's footprint from software creation into software operation. Emergent raised $70 million in January at a $300 million valuation, backed by SoftBank, Khosla Ventures, and Lightspeed.
Why it matters
The agentic AI space is getting crowded fast, but most players are building dedicated interfaces — new dashboards, new inboxes, new surfaces to manage. Emergent is betting the opposite: that the interface layer is already won by WhatsApp and iMessage, and the right play is to embed agents there. That's a meaningful UX wedge, especially in markets like India where WhatsApp is effectively the operating system for business communication. The trust boundaries approach also signals awareness of the liability problem that fully autonomous agents carry — a practical concession that the field hasn't fully solved autonomous action yet.
What to watch
Wingman's real test is whether task execution holds up outside demos. Vibe-coding platforms generate code; agents that touch live calendars, emails, and business tools carry far higher failure costs. Emergent has 8 million builders on its platform — converting even a fraction of them to Wingman users would be a strong proof point. Watch whether the WhatsApp-native approach gains traction in enterprise contexts where IT controls messaging infrastructure, and whether Anthropic or Microsoft respond with tighter messaging integrations of their own.