1. Headline & intro
From ‘vibe coding’ to invisible staff: Emergent’s chat‑native AI agents raise the stakes
Software that quietly works in the background is moving from hype to reality, and an India‑born player just placed a big bet. Bengaluru startup Emergent, best known for its “vibe‑coding” platform that lets non‑developers build apps via chat, is now trying to operate those apps for you as well. Its new product, Wingman, lives inside WhatsApp, Telegram and other messengers and promises to do your routine work, not just help you click through it faster. In this piece, we’ll unpack what Emergent is really attempting, why the messaging-first strategy matters, and what this means for European businesses staring down the coming wave of AI agents.
2. The news in brief
According to TechCrunch, Indian startup Emergent has launched Wingman, a messaging‑first autonomous AI agent meant to execute tasks across users’ tools and workflows.
Emergent originally gained traction with a “vibe‑coding” platform that competes with Cursor and Replit, allowing people without formal programming skills to build full‑stack applications using natural‑language prompts. The company says more than 8 million “builders” have used the platform, with around 1.5 million monthly active users.
Founded in 2025 and headquartered in Bengaluru, Emergent raised about $70 million in January at a reported $300 million valuation, with investors including SoftBank, Khosla Ventures and Lightspeed.
Wingman connects to messaging apps such as WhatsApp, Telegram and Apple’s iMessage. Users assign and monitor tasks in chat while the agent runs in the background across email, calendars and workplace software. Emergent highlights “trust boundaries”: Wingman can perform routine actions autonomously, but it asks for approval before more consequential steps. The product is launching with a limited free trial, after which it will be paid, especially for existing Emergent users.
3. Why this matters
Emergent is not just adding a chatbot. It is attempting a full stack for small businesses and teams: AI to build the tools, and AI to run the tools. That’s a different ambition from most SaaS vendors bolting an “AI assistant” onto existing products.
The winners, if this works, are obvious: solo founders, micro‑SMEs and operations teams drowning in repetitive digital work. Wingman’s promise is that instead of hiring a junior ops person or paying a BPO firm, you delegate mundane tasks to an AI that lives where you already spend your time – WhatsApp or Telegram.
The losers are less obvious but equally real. Traditional business software that depends on users manually clicking through dashboards risks being commoditised into “back‑end APIs” for agents like Wingman. Outsourcing providers that handle routine email, data entry or follow‑ups could also feel pressure if a chat‑native agent can do 60–70% of that work cheaply and 24/7.
Strategically, the messaging‑first choice is key. OpenClaw‑style agent tools and Anthropic’s Claude experiments have mostly targeted developers and power users who are comfortable with terminals, IDEs or dedicated web apps. Emergent is going after everyone who already lives in chat – which, in high‑WhatsApp‑usage markets from India to Southern Europe to Latin America, is almost every knowledge worker.
Finally, “trust boundaries” are a smart attempt to address the biggest blocker for autonomous agents: fear. People might happily let an AI rearrange their calendar, but sending invoices or approving payments without oversight is another matter. Building granular control into the product from day one is not just UX; it’s risk management.
4. The bigger picture
Wingman lands in the middle of the industry’s pivot from assistants to agents. If 2023 was the year of chatbots in every product, 2024–2026 is about software that can chain actions together and do actual work.
We’ve already seen early versions of this trajectory:
- Developer‑focused tools like Cognition’s Devin pitched themselves as “AI software engineers” capable of reading tickets, editing code and running tests end‑to‑end.
- Open‑source experiments such as Auto‑GPT showed what happens when you let language models call tools repeatedly to pursue long‑running goals.
- Big platforms like Microsoft Copilot and Google’s Duet/Workspace assistants began to stitch together email, documents and calendars into semi‑autonomous workflows rather than one‑off prompts.
Emergent’s twist is to treat the messaging client as the control plane. Historically, we’ve been here before: Slack bots and early chat‑ops in DevOps circles tried to turn chat into the command centre for infrastructure. The difference now is that foundation models can understand natural language with far more nuance, plan multi‑step tasks and adapt to ambiguous instructions.
Compared with US‑centric players, Emergent is optimising for markets where WhatsApp is the default work tool, not an add‑on. That’s a blind spot for many Silicon Valley products that still assume email plus proprietary SaaS as the centre of gravity.
The competitive landscape is also being shaped by governance concerns. TechCrunch’s own coverage notes that Anthropic temporarily banned OpenClaw’s creator from accessing Claude, illustrating how uncomfortable model providers are with fully unleashed agents. Emergent’s pre‑emptive focus on “trust boundaries” is partly a product lesson, partly a hedge against being cut off by upstream model APIs if something goes wrong.
In short, Wingman is an early sign that the next wave of AI agents will be embedded in the tools we already use, not introduced as yet another SaaS tab.
5. The European / regional angle
For Europe, Wingman is interesting precisely because it does not come from a US or EU giant. An Indian startup building chat‑native agents on top of WhatsApp and iMessage will inevitably run into Europe’s regulatory wall – and that makes it a test case.
Under the EU’s AI Act, systems that help make decisions about hiring, credit, health or education can fall into “high‑risk” territory. If European SMEs start using Wingman to triage CVs, negotiate invoices or interact with customers, Emergent will need to demonstrate transparency, logging and human oversight. Its “trust boundaries” could map quite neatly onto the AI Act’s requirement for human control, but only if implemented with proper audit trails.
Then there is GDPR and the Digital Services Act (DSA). A chat‑embedded agent means:
- Personal data from WhatsApp, email and internal tools is being processed and possibly transferred internationally.
- Consent, purpose limitation and data‑minimisation suddenly become much harder to manage if the AI is dynamically pulling in context from multiple sources.
European customers – especially in Germany, Austria and France – will press for clear data‑processing agreements, EU data residency and the ability to self‑host or use EU‑based model providers.
For Europe’s own AI startups, Emergent’s move is a wake‑up call. Many EU players are focused on foundation models or vertical solutions (health, manufacturing, legal). Fewer are building general‑purpose, chat‑native agents that run across tools. If Indian and US companies occupy that layer, European vendors risk being relegated to compliance wrappers or niche add‑ons.
6. Looking ahead
Expect the next 12–24 months around Wingman – and AI agents more broadly – to revolve around four questions.
1. Reliability and guardrails. Emergent openly acknowledges that Wingman struggles with messy edge cases, ambiguous goals and judgment‑heavy workflows. The real test will be whether it can carve out a stable subset of tasks (follow‑ups, reminders, simple workflow orchestration) where it consistently performs better than a human assistant, without catastrophic mistakes.
2. Depth of integrations. A messaging shell is useless without deep hooks into CRMs, ERPs, ticketing systems and custom internal tools. Emergent’s vibe‑coding heritage is an advantage: it can, in theory, generate those integrations on the fly. But enterprises – and many mid‑market European firms – will demand robust, certified connectors rather than auto‑generated glue code.
3. Monetisation and positioning. Does Wingman become a paid add‑on for existing Emergent customers, a standalone “AI staffer” for SMEs, or a white‑label agent that European consultancies embed into their own offerings? Each path implies a different regulatory and support burden.
4. Platform dependence. Building on WhatsApp, Telegram and iMessage is powerful but risky. Policy changes by Meta or Apple around business messaging, bots or automation could break key flows overnight. To serve regulated European industries, Emergent may need a parallel, web‑based and email‑centric control surface that does not depend on consumer messaging apps.
For European readers, the practical takeaway is simple: start experimenting, but keep humans in the loop and insist on exportable logs, clear permissions and contractual guarantees about data location. The AI agent era is coming; the question is whether you shape how it enters your organisation or let it arrive piecemeal via employees’ phones.
7. The bottom line
Emergent’s Wingman is more than another AI chatbot. It’s an early attempt to turn WhatsApp‑style messaging into the command centre for autonomous software agents – a model that fits India, Europe and Latin America far better than US‑centric email workflows. If the company can prove reliability and navigate EU rules on data and AI, it could become a serious contender in the global “AI staffer” race. The real question for European businesses is: will your first AI colleague speak through a US, European or Indian platform?
