Digipals Wants AI to Fix Your Friend Group, and That Is Either Genius or Delusional

The Macro: Social Media Stopped Being Social a Long Time Ago

I want you to think about the last time a social media app helped you actually see a friend in person. Not watch their story. Not react to their post. Not comment on their photo from a trip you were not invited to. Actually facilitated a real-world hangout. For most people, the answer ranges from “rarely” to “never.”

This is the paradox of modern social platforms. We are more connected than ever and hanging out less than ever. The average American adult has fewer close friendships than at any point in recorded history. Surgeon General Vivek Murthy declared a loneliness epidemic. People have plenty of group chats. What they do not have is a seamless way to turn those group chats into actual plans.

The logistics problem is real and underappreciated. Getting four adults together for dinner requires coordinating calendars, picking a restaurant everyone likes, making a reservation, figuring out who is driving, and eventually splitting the bill. Each step introduces friction. Each friction point is an opportunity for the plan to die. “Let’s do something this weekend” turns into a week of back-and-forth texts that end with “maybe next time.” Everyone has experienced this. Nobody has fixed it.

The existing tools are fragmented. iMessage handles the chat. Google Calendar handles scheduling. OpenTable handles reservations. Splitwise handles bill splitting. Yelp handles restaurant discovery. You need five apps to do one thing. And none of them talk to each other in any meaningful way.

The Micro: Garry Tan Saw Something Here

Digipals was founded by Peggy Wang (CEO) and Mathew Matakovic (CTO). They came through Y Combinator’s Fall 2025 batch with their YC partner being Garry Tan himself, which is worth noting because Tan does not attach his name casually. The team is three people in San Francisco, building what they describe as the first AI-native social operating system.

The product is a group chat app with ambient AI agents baked into every conversation. These agents have access to your calendar, your location, and your photo library. When your group chat starts talking about getting dinner, the AI jumps in and handles logistics. It can call restaurants to make reservations. It can scan everyone’s calendars to find a time that works. It can split bills by analyzing receipt photos. It can even surface memories from past hangouts to help the group decide where to go.

The “call restaurants” feature is interesting because it implies voice AI integration for real-world tasks. This is not just matching algorithms and push notifications. This is an AI agent picking up the phone and booking a table for six at the place your group liked last month. That level of proactive coordination is something no existing messaging app even attempts.

The longer-term vision involves an expandable widget ecosystem where developers can build AI-powered tools that plug into group chats. Think of it like apps within the app, but each one is designed to facilitate human connection rather than capture attention. That is architecturally ambitious and potentially very sticky if they can attract third-party developers.

The competitive landscape is dominated by platforms that are not really trying to solve this problem. iMessage is the default group chat for iPhone users but has zero coordination features. WhatsApp owns the global messaging market but treats groups as chat rooms, not planning tools. Discord built sophisticated server infrastructure for communities but is not optimized for small friend groups making dinner plans. Partiful does event pages for parties. IRL tried to be a social calendar and imploded spectacularly. The space has a graveyard of failed attempts, but most of those failures were trying to be social networks rather than coordination tools.

What makes Digipals different, at least in theory, is the AI-native approach. Previous attempts at social coordination required users to manually enter information, create events, and manage logistics. That is just moving friction from one app to another. By embedding AI agents that proactively handle coordination tasks, Digipals is trying to reduce the friction to near zero. You talk about plans in your group chat. The AI makes them happen.

The Verdict

I go back and forth on this one. On one hand, the problem is painfully real. Social coordination is genuinely broken and nobody has fixed it. The AI-native angle is the first approach that might actually reduce friction enough to change behavior. On the other hand, consumer social is where startups go to die. The switching cost from iMessage is enormous. The chicken-and-egg problem is brutal: the app is only useful if your friends are on it, and your friends will only join if it is useful.

The IRL comparison looms large. That company raised hundreds of millions, claimed 20 million users, and turned out to be largely fraudulent. The social coordination category has trust issues. Digipals will need to be transparent and measured in its claims, and patient about growth.

What gives me some optimism is the messaging-first approach. They are not trying to be a social network. They are trying to be a better group chat. That is a more contained product with a clearer value proposition. If the AI coordination actually works, word of mouth could drive adoption one friend group at a time. You do not need viral growth. You need one person in each group to drag their friends onto the platform, which is exactly how WhatsApp and iMessage grew.

At thirty days, I want to see how many group chats have used the AI coordination features at least once. Downloads are meaningless. Feature activation is everything. At sixty days, I want to see whether groups that use the AI features have higher message volume and retention than groups that just chat. At ninety days, the question is whether the AI restaurant calling and calendar coordination work reliably enough that users trust them. One botched reservation and you lose a group forever. The idea is strong. The timing is right. The graveyard of predecessors is large. Digipals needs to be better, not just different.