Shapes Raises $24M to Build AI Companions for Group Chats

Shapes, Inc. has secured $24 million in funding to develop a social platform where AI companions and human users chat together in groups. The startup, founded by Anushk and Noorie, sits at the intersection of conversational AI and social messaging — targeting people who want to add personalized AI characters to their group conversations.

How the Platform Works

The company's core technology, called S.H.A.P.E.S. (Sentient Human-like Artificial Personalization and Enhancement System), lets users create customized AI companions. These can take almost any form — a pet, a celebrity, an anime character, or anything else. The key difference from today's chatbots is that these AI entities join group conversations alongside real people, rather than sitting alone in one-on-one chats.

Users configure what their AI companion is like — its personality, how it responds, what it acts like — then invite it into ongoing group chats with friends or colleagues. The AI participates in the same conversation thread as everyone else.

What Problems Does It Solve?

Shapes frames itself as AI that enhances human interaction rather than replacing it. The platform targets moments where an AI participant could add value: entertainment in gaming communities, creative brainstorming for work teams, or conversation when no humans are available to chat.

Investors backing the round seem confident in this "hybrid" approach — mixing AI and humans in the same social space. Large language models (the AI technology behind ChatGPT and similar tools) have become good enough now to hold conversations in character over long stretches. Early users mention applications ranging from gaming groups wanting persistent AI characters to professional teams using AI to help brainstorm ideas.

The Technical Challenge

Making AI work naturally in group chats is not trivial. The system must follow a conversation with multiple people talking at once, keep the AI's personality consistent over time, and add to the chat without dominating it or feeling awkward.

There are real engineering hurdles. The platform needs to generate responses in real time — group chat users expect quick replies, similar to what another human would do. That differs from applications where you ask an AI a question and wait minutes for an answer.

Creating AI companions as customizable as Shapes describes likely requires sophisticated techniques for shaping how the AI behaves (called prompt engineering) and possibly training custom AI models for different character types. The system has to keep each AI consistent in personality while adapting to different conversations and group dynamics.

Why This Matters Now

The funding reflects a shift in how startups are thinking about AI. Most AI assistants today work one-on-one — you and the chatbot, alone. Shapes is part of a wave exploring how AI could fit into existing social structures people already use, rather than asking people to adopt entirely new platforms.

This pattern echoes something we saw in the early days of social media. Facebook, Twitter, and others had to figure out how to mix automated or algorithmic content with what real people posted, without making things feel fake or turning people away. The central question was the same then as now: how do you add powerful technology to a social space without breaking what makes that space valuable to users.

Over three decades of covering these shifts, I have watched several attempts at this balance. The companies that succeeded — Instagram adding algorithmic feeds, Slack adding automation, Discord adding bots — found ways to enhance what people were already doing rather than forcing them into something entirely new. Those that failed often pushed too hard, too fast.

How Does It Make Money?

Shapes has not announced a specific business model yet. Platforms like this typically explore a few paths: charging subscriptions for advanced customization options, selling premium AI characters, or letting users buy and sell AI personality configurations to each other.

The group chat angle may give Shapes an advantage over single-user AI assistants when it comes to keeping people engaged. When you chat with friends or colleagues, you tend to come back — the social pull keeps you using a platform. But that same social element creates challenges around moderation: the company will need to ensure AI participants behave appropriately across many different group contexts and cultures.

What Comes Next

The $24 million gives Shapes runway to grow its engineering team and refine the platform before it tries to reach a mass audience. The crucial test will be whether AI companions add genuine value to group conversations or just feel gimmicky and distracting.

The harder problem is balance. The AI needs to be capable enough to contribute something real to a chat, but restrained enough not to overwhelm the humans or trigger the uncanny valley — that creepy feeling when something artificial becomes too lifelike and starts to feel wrong.

The opportunity extends beyond games and entertainment. As remote work settles into how people operate, AI tools that can help with team meetings, knowledge sharing, and collaborative work might find a real home in professional settings.

For now, Shapes is testing one approach to a question the entire AI industry is grappling with: how to weave artificial intelligence into the social structures people already use, in ways that make life better rather than replace human connection. The money buys time to explore that balance. Whether it works will come down to what actual users do.