Group Chat Agent Wabi Secures Unicorn Status With $55M Series B

TL;DR San Francisco startup Wabi has raised $55 million at a $1.25 billion valuation, betting that practical AI assistants should live directly inside messy group threads rather than isolated consumer dashboards.

Every consumer has suffered through the friction of the group chat: seven people debating dinner reservations across two time zones, three calendar conflicts, and an abandoned plan because nobody wanted to open OpenTable. Wabi wants to handle that entire exchange without anyone switching apps.

The company closed a $55 million Series B funding round this week that pushes its valuation to $1.25 billion. The round underscores an emerging pivot among startups backing software agents that operate collaboratively among multiple humans, rather than treating artificial intelligence as a private one-on-one sounding board.

group of friends looking at smartphone screen outdoors — Photo by Vitaly Gariev on Unsplash

Engineering Consensus Among Humans

Wabi functions as an ambient participant across platforms like WhatsApp, Apple’s Messages via RCS, Slack, and Telegram. Instead of waiting for a single prompt, the software parses natural banter between multiple users, tracks divergent preferences, and interjects only when it detects an actionable consensus or an explicit request.

Building an autonomous agent capable of negotiating with a single person is straightforward. Doing so in a room of five people with conflicting schedules, variable payment preferences, and disjointed conversational styles requires a different software architecture altogether.

Functional Layer	Standard 1-on-1 Chatbot	Wabi Group Chat Agent
Context Window	Single user prompt stream	Multi-speaker attribution & sentiment parsing
Trigger Mechanism	Direct query / invocation	Ambient consensus detection or direct @mention
Execution Rails	Single API call via user token	Multi-party authorization & split-billing hooks
Conflict Resolution	Asks user to restate prompt	Polls participants or proposes compromise slots

Once a group reaches agreement—say, booking a table for four in downtown Austin at 8:00 PM—Wabi executes the booking using integrated APIs, syncs the event to the participants’ calendars, and drops the confirmation directly into the thread.

The Multi-Party Authorization Problem

The biggest hurdle for social agents is not natural language processing; it is authorization and security. If an automated assistant has access to payment rails and booking systems, how does it prevent one malicious or prank-minded friend from spoofing an order?

Wabi relies on decentralized cryptographic consent flows built on modern OAuth 2.0 authorization frameworks to verify transactions. If the agent needs to reserve flights or split an expensive rental deposit, it generates biometric authorization prompts sent individually to each payer’s registered device.

The approach sidesteps the prompt-injection vulnerabilities that plague naive web-browsing agents. By treating every participant as an unverified entity until confirmed via hardware-backed security keys, the platform prevents conversational manipulation from triggering unwanted financial commitments. Rapid maturation across ai apps has made action execution possible, but Wabi’s team argues that safety boundaries in multi-user environments require purpose-built state machines rather than open-ended model prompts.

modern server room with glowing blue status lights — Photo by Tyler on Unsplash

Competing on Neutral Ground

The valuation reflects fierce investor appetite for horizontal layers that bridge walled gardens. Meta already offers consumer agents within WhatsApp and Messenger, while Apple continues expanding Apple Intelligence across iOS system apps. Yet platform owners face inherent structural limitations: Apple cannot easily arbitrate a thread containing Android devices, and Meta lacks native integration into enterprise calendars or competitor messaging silos.

By positioning itself as an agnostic service that functions identically across operating systems, Wabi is attempting to capture the connective tissue between platforms. The capital influx will fund expanded merchant integrations across ticketing, travel, and healthcare scheduling, where multi-stakeholder coordination typically falls apart.

Winning consumer attention has always been expensive, but capturing consumer coordination represents far stickier territory. If Wabi can reliably eliminate the administrative friction of daily social life, it will turn messaging threads into the primary execution environment for consumer commerce.