Although algorithms mostly operated in the background, artificial intelligence improved that system by flagging content, targeting advertisements, and ranking feeds. We can say, however, that age is coming to an end at the start of 2026.

With the silent acquisition of Manus AI in late 2025, Meta crossed a strategic threshold. This was not another internal research upgrade or a smarter recommendation engine. Manus represents something more ambitious: AI agents that don’t just respond, but act inside the apps billions already use, starting with WhatsApp.

Let’s start at the beginning.

Initially, Manus AI was a specialised start-up that concentrated on self-governing AI agents. Manus was designed to independently plan, reason, and carry out multi-step operations, in contrast to traditional chatbots that wait for commands and reply with text.

With a single conversational command, the system could produce documents, make presentations, edit media, manage schedules, and sequentially engage with several tools.

In practical terms, Manus was closer to a digital operator than a digital assistant.

For an estimated $2–3 billion, Meta purchased the business; this amount indicated a long-term strategic goal rather than immediate product improvement. In contrast to

Meta’s more spectacular statements about VR or the metaverse, the acquisition was mainly overlooked, but inside it marked a turning point: Manus would serve as the foundation for Meta’s upcoming AI phase.

The reasoning was obvious. Meta already had a large following on Facebook, Instagram, WhatsApp, and Messenger. An AI layer that could take significant action within those platforms was what it lacked.
Manus filled that gap.

A structural shift

Meta’s earlier AI efforts, branded simply as Meta AI, focused on conversational help: answering questions, generating images, editing photos, and summarising posts. Useful, but limited.

Manus changes the equation by introducing agency.

Instead of asking an AI what to do, users can ask it to do something.

Draft a report. Prepare slides. Create a short video. Organise messages. Summarise weeks of conversation. Plan a campaign. The AI doesn’t just respond, it executes, often invisibly, across systems.

This is why messaging apps matter so much to Meta’s strategy.

Why WhatsApp is central to Meta’s AI play

WhatsApp is Meta’s most strategically sensitive product. It is deeply personal, encrypted, and used daily by more than two billion people, often for far more than casual chat. Families plan logistics. Small businesses take orders. Teams coordinate work.

Embedding Manus-powered AI agents directly into WhatsApp turns the app into something closer to a personal operating system.

According to multiple reports, Manus-style agents are being rolled out first in controlled environments, with broader WhatsApp integration planned next. These agents can live inside chats, understand context, remember preferences, and assist quietly in the background.

This is not about adding a chatbot button. It is about rewiring how messaging works.

For Meta, this approach avoids a critical risk: asking users to download yet another app. Instead, AI arrives where people already are.

The Business model: Beyond ads

Meta insists AI is not just a cost centre, it’s a revenue engine in waiting.

Three overlapping business strategies are emerging.

Premium subscribers come first.
According to reports, Meta is working on premium tiers for WhatsApp, Instagram, and Facebook that will enable more sophisticated AI features. Basic help might stay free, but it’s anticipated that deeper memory, media creation, productivity tools, and autonomous task execution would all require a subscription.

For Meta, whose earnings have always come from advertising rather than direct user payments, this would represent a substantial shift.

Second, enterprise AI services.
Manus is expected to remain available as a standalone product for businesses and developers. This positions Meta to compete quietly but directly with enterprise AI offerings from cloud providers and AI labs.

Engagement comes in third.
It becomes challenging to replace an AI that is familiar with your routine, conversations, and habits. Meta raises switching costs without requiring users to enter overt walled gardens by integrating intelligence into routine interactions.

The stated vision

Mark Zuckerberg has framed Meta’s AI ambition in unusually expansive terms, describing a future of personal superintelligence, AI systems tailored to individuals, not corporations.

In this framing, Manus is not a productivity gimmick. It is infrastructure: a layer of intelligence that understands users over time and helps them navigate work, creativity, and communication.

Whether that vision is achievable or desirable remains open to debate. But it explains why Meta is willing to invest billions before a clear payoff is visible.

The Friction: Privacy, trust, and user pushback

Not everyone is convinced.

Users of WhatsApp have already expressed their annoyance about visible AI features that aren’t completely deactivable. Critics argue that AI agents, especially ones with context memory, conflict with WhatsApp’s image of valuing privacy.

Additionally, there are still unanswered concerns regarding permission, transparency, and data borders. The importance of information, behavioural signals, and usage trends is enormous, even if the message content is encrypted.

For a company with Meta’s history, trust will be as important as technical capability.

The goal of Meta’s acquisition of Manus AI is not to dominate the chatbot competition. It involves rethinking the interface between digital systems and people in a large, silent, and gradual manner.

WhatsApp will be more than simply a messaging app if it succeeds. Instagram and Facebook will be more than just social media platforms. These settings will be AI-mediated, with intelligence running constantly in the background.

This is Meta’s most consequential bet since the mobile transition, and unlike the metaverse, it is already embedded in how people live.