Emergent, an innovative Indian startup based in Bengaluru, has recently launched “Wingman,” an autonomous AI agent focused on messaging. This launch marks Emergent’s expansion from its popular vibe-coding platform to a software category that performs background tasks, similar to platforms like OpenClaw and Anthropic’s Claude.
Previously, Emergent garnered attention with its vibe-coding technology that allows non-technical users to create full-stack applications using natural language prompts, competing with products like Cursor and Replit. With Wingman, the startup aims to advance from merely software creation to task execution, enabling AI agents to undertake routine activities across various tools and workflows.
Mukund Jha, the co-founder and CEO of Emergent, explained that the natural progression for the company was to enhance operational autonomy, moving from software that supports business processes to software that actively contributes to managing them. To date, more than eight million users have engaged with Emergent’s vibe-coding platform, with about 1.5 million active monthly users. Established in 2025, the company successfully raised $70 million in January at a valuation of $300 million, securing investments from notable firms including SoftBank, Khosla Ventures, and Lightspeed Venture Partners.
Wingman operates through widely-used messaging platforms such as WhatsApp and Telegram, enabling users to assign and track tasks via chat. It functions in the background across integrated tools like email and calendar apps, autonomously executing routine tasks while seeking user approval for more significant actions. This dual approach aims to balance user control with efficiency.
The emergence of autonomous AI agents is becoming increasingly competitive, with various firms racing to develop similar task-completing tools. Projects like OpenClaw have garnered interest from early adopters, and industry giants like Anthropic and Microsoft are also entering the space with their agent-based solutions.
Emergent differentiates Wingman by embedding it in familiar messaging platforms, allowing users to interact with the agent naturally, without needing to navigate new interfaces. They have also introduced “trust boundaries,” which enable the agent to autonomously handle everyday tasks while requiring user consent for more critical operations. This feature addresses prevalent concerns regarding the risks of fully autonomous systems.
Jha emphasised that the choice to integrate Wingman within existing messaging frameworks stems from their understanding of routine workflows, noting that effective communication often occurs via chat, voice, and email. As a result, he believes that these channels will increasingly serve as the primary means for users to engage with AI agents.
However, like many emerging AI solutions, Wingman is not without challenges. Jha acknowledged that the AI may struggle with ambiguous situations, complex edge cases, and scenarios requiring significant human judgement. The rollout of Wingman includes a limited free trial, transitioning to a paid model, with existing users of Emergent able to access the agent through their current accounts.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
