Emergent has rolled out Wingman, an autonomous AI agent that lives in chat apps.

From vibe-coding to background work

Emergent started out as a tool for people to build software without writing code, and it's now trying to go one step further: let that software take action. The Bengaluru-based startup made its name with a so-called vibe-coding platform that lets non-developers assemble full-stack apps using natural language prompts. Over 8 million builders have tried the platform and roughly 1.5 million people use it every month, the company says.

This move makes sense. If people can describe an app and see it appear, the next ask is often: can the app actually do things for me? That query is exactly what prompted Emergent to design Wingman — a messaging-first agent that can run tasks across apps while you carry on with other work.

Emergent was founded in 2025 and raised $70m in January at a $300m valuation, backed by investors that include SoftBank, Khosla Ventures and Lightspeed Venture Partners. Those numbers give the startup a runway to push Wingman into an already crowded but fast-growing market for autonomous agents.

How Wingman works

Wingman is built to operate through chat interfaces such as WhatsApp, Telegram and Apple’s iMessage. Users tell Wingman what they want inside a conversation — assign a task, ask for a follow-up, or hand across context from a thread — and the agent acts.

It also links to email, calendars and workplace apps so it can run routine actions in the background while you keep using the tools you already have.

Point is, Emergent wants Wingman to feel like a colleague in chat rather than another dashboard to learn. The agent will complete small tasks, like scheduling, triaging messages or pulling together briefs, without a constant handover. But for actions that could have bigger consequences — sending a payment, deleting records, committing code — Wingman will pause and ask for explicit user approval.

Emergent calls these "trust boundaries," which separate tasks the agent can handle alone from those requiring human approval. That design aims to reduce friction while still keeping control where it matters.

Where Wingman sits in the agent race

Autonomous AI agents have exploded into the market this year. Projects such as OpenClaw and larger firms including Anthropic and Microsoft are all pushing variants of agents that can carry out tasks across services. Some of those systems are oriented toward power users and automation hobbyists, while others aim for enterprise workflows.

Emergent is picking a familiar route to differentiation: meet users where they already work. The company argues that much of daily coordination already happens through chat, voice and email — so agents that live there can be more useful and less disruptive than a separate product. Mukund Jha, co-founder and CEO of Emergent, told TechCrunch that the team wants to bridge creation and execution — helping people not just to build software but to run it more autonomously.

Mastheads

Wingman’s chat-first approach also plays to regional usage patterns. In many parts of Asia and beyond, WhatsApp and Telegram are core productivity channels, not just messaging apps. Embedding an agent into those platforms could speed adoption among teams that already rely on chat as the glue between tools.

Privacy, safety and limits

Autonomous agents bring up many questions, like how much access they should have, how credentials stay safe, and who’s responsible if something goes wrong. Emergent’s use of trust boundaries is a response to those concerns, but it’s only part of the picture.

Security architects will want to know how Wingman stores tokens, whether it can be compartmentalised by team or project, and what audit trails the system produces. Enterprises will also look for integration with single sign-on, enterprise-grade logging and policies that can be enforced centrally. Emergent says Wingman will require permission for consequential steps, but enterprises tend to demand more control than consumer users do.

But there are still technical limits. Many agent systems today still struggle with long-running tasks that need cross-system consistency, and with error handling when APIs or services change. Wingman will need solid orchestration and retry logic to avoid user frustration. And like other agents, it must manage ambiguity in natural-language instructions to avoid unintended outcomes.

Product strategy and go-to-market

Emergent plans to grow its vibe-coding user base and turn those users into Wingman customers. That could create a neat loop — someone builds an app in Emergent’s platform and then asks Wingman to operate that app or manage workflows around it.

Early adopters often come from teams that have already automated parts of their workflow with scripts or bots, and those groups could be the first to try Wingman. The chat integration lowers the barrier to trial: instead of having to deploy an app or train users on a new interface, a team can invite Wingman into an existing channel and start delegating.

Emergent will also have to decide how to price the product. Will Wingman be a free add-on for vibe-coding customers, a premium feature, or a standalone subscription aimed at enterprises? Each route has trade-offs: bundling can drive adoption among builders, while a dedicated enterprise product can fund the heavier engineering required for security and compliance.

Competition and differentiation

Competition is intense. Some projects target developers and power users who want full control over agents, while others aim at business users who want simple, safe automation. Emergent’s focus on chat and its existing user base give it a distinct angle, but it won't be easy to hold the ground.

Anthropic and Microsoft bring vast compute and enterprise reach; niche projects like OpenClaw appeal to automation-savvy early adopters. Emergent will need to show that Wingman is both easy to use and trustworthy — and that it can scale beyond simple tasks into multi-step processes that actually save time.

Basically, Emergent must prove Wingman improves productivity in measurable ways. If a team can shave hours off weekly admin, or reduce missed follow-ups, paying for an agent becomes much easier to justify.

What it means for builders

For the millions who've used Emergent to build apps, Wingman promises a new phase: apps that not only exist but act. That could change how organisations think about low-code and no-code tools — from “we built an app” to “we have software that helps run the business.”

Still, developers and IT teams will watch how much autonomy Wingman is allowed and how easily its behaviour can be audited and limited. Teams that rely on strict change control won't want an agent making decisions without clear approvals and visibility.

Right now, Wingman looks like an experiment in putting agents inside everyday workflows rather than asking users to learn new software. Whether that experiment scales will depend on trust, reliability and clear business benefits.

Related Articles

"The obvious next step for us was, can we help them not just build the software, but actually operate more autonomously through it?" said Mukund Jha, co-founder and CEO of Emergent.

This article was created with AI assistance.