To Launch Something Big, You Need an Enemy

01 /The Concept

Digipals is building an app where "AI makes us more social." That might sound strange at first – aren't we already plenty connected?

But the founders make a specific argument: social networks have mutated into glossy media platforms and stopped serving their original purpose – helping people stay in touch with friends.

"In my Facebook and Instagram feeds, I don't see friends anymore," the startup writes. "A third of it is ads. Another third is posts from strangers the algorithm decided to show me. The last third is AI-generated content produced purely for views and engagement. So I actually talk to my friends in real life, or in group chats where it's just us."

Digipals wants to build the app for those group chats – specifically, to help friends see each other more often in real life.

The core mechanism: an AI agent embedded in the chat that's aware of each participant's schedule, location, preferences, and relationship history. The agent helps coordinate meetups – suggesting times when everyone is free, finding a restaurant convenient for the group, booking a table, helping split the bill, and reminding people to post photos after the fact.

All of this is possible in any group chat today, of course – but it requires too many messages, too much back-and-forth, too much friction. Digipals wants to reduce that friction so that the activation energy for planning a meetup is low enough that more of them actually happen.

The product concept is an "AI-native social operating system" – a chat platform where specialized widgets surface at exactly the moment they're needed. Someone mentions dinner plans? A restaurant suggestion widget appears. Someone posts a receipt? A bill-splitting widget pops up.

First widgets in development: restaurant, movie, and concert recommendations; flights and hotels; bill splitting; and sports score notifications. Each appears as an inline block with images and buttons – you choose the time, choose the place, and book without leaving the chat.

Next-phase widgets will focus on maintaining relationships with old friends and building new ones: reminders about shared memories, alerts when a chat participant has news worth celebrating together, and suggestions to add a mutual friend of one of the members who the whole group might enjoy.

The widget ecosystem will be open to third-party developers, with its own app marketplace where any community can find and install what it needs.

Digipals is currently in Y Combinator and posted about its upcoming launch on the YC blog last week.

02 /Why It Matters

Technically, Digipals is close in concept to Continua ([related review](/review/a-tut-nuzhny-sovsem-drugie-ii-agenty)), which raised $8M in its first round in August – also for an AI agent embedded in group chats, also focused on helping friends coordinate real-world activities.

Continua's agent monitors conversation context and surfaces only when participants seem to need help – when, where, and how to get somewhere, for instance. Knowing when to intervene turns out to be a genuinely hard technical problem: almost every existing agent is built for individual question-and-answer interactions, not for reading an ongoing multi-party conversation and deciding whether to step in.

Digipals, though, is engaging two distinct trends rather than one.

The first is group chat chaos. Chats have become the default space for families, friend groups, colleagues, and interest communities. But their linear structure and the short memories of their participants create endless repetition – people asking the same questions, re-explaining the same context, re-confirming the same plans. And reaching any group decision – where to go, when to meet, what to do – becomes a coordination nightmare the larger the group gets. Paralysis is common; the meetup often just doesn't happen.

Interestingly, this problem isn't just a real-world phenomenon. Root ([related review](/review/boltovnjoj-polzovatelej-ne-uderzhish)) built a platform with an app store for gaming communities – including apps for planning raids, coordinating missions, and managing group objectives. Root raised $9M in its first round in June.

The second trend is what many startups are now framing as "too connected online to avoid loneliness." The more time spent online, the more followers you accumulate and the fewer friends you have. "Real life happens offline" – a phrase that keeps coming up across multiple startup pitches. Only offline interaction sustains and deepens actual human relationships.

Howbout ([related review](/review/a-teper-vse-poshli-obratno)) built a group calendar app for this same purpose – group chat, shared schedules, in-person meetup planning, photo sharing. Raised $13.7M, including $2.5M post-review.

Broader platforms like Plots ([related review](/review/pojdjom-potusuemsja)), Pie, and POSH took the concept further, enabling discovery of and participation in local in-person events. Raised $3.5M–$31M.

A separate segment has emerged specifically for older adults – apps setting minimum age thresholds of 40, 50, or 55. Meet5 ([related review](/review/dlja-teh-komu-za-50)) raised €8M in September. Hank and Rest Less have also raised substantial rounds.

Another niche: apps built specifically for adult women looking to make new female friends. RealRoots ([related review](/review/okazyvaetsja-ljudi-gotovy-platit-ne-tolko-za-seks-no-i-za-druzhbu)) went through Y Combinator this past summer; Les Amis appeared in the same review.

The underlying dynamic: most offline social events – parties, clubs, meetups – skew young by default. Adults who wander into those spaces often feel out of place. Hence the need for dedicated platforms – addressing the same loneliness problem, for a different audience.

03 /Opportunities

Two trends described, which suggests two directions. Actually three.

First: platforms for managing group chat chaos. This can take the form of AI agents like Digipals and Continua, or specialized apps like Root.

Second: platforms for real friendship – separate from social networks. Options here are numerous: smart group chat tools focused on in-person coordination (Digipals, Howbout), broader event discovery platforms (Plots, Pie, POSH), and segment-specific apps (Meet5, Hank, Rest Less, RealRoots, Les Amis).

Or – combine both directions in a single product, like Digipals is doing: reduce group chat friction specifically in service of helping people meet up in person.

Of the three directions, the combination play – reducing group chat friction specifically to drive real-world meetups – is the most defensible because it creates a habit loop: the app earns its place in the group chat by being genuinely useful for coordination, then the meetup happens, and the meetup reinforces the habit of using the app. That's a stickier acquisition path than either pure coordination tooling or pure social discovery. Which slice of that stack would you build?