AI should increase the quantity and quality of our human relationships.

That comes down to two human problems: connecting with the right people, and having better relationships with our people.

We're all looking for someone.

A co-founder, an investor, a supplier, a client, a friend, or even someone to spend your life with. Out of eight billion people, some small percentage are exactly who we're looking for at any given moment.

Finding them costs more time than anyone has.Scouring the internet, events or our network for new connections can take a tremendous amount of time looking and filtering.
Reaching out got cheap, so the inbound is mostly noise.AI and the internet have made reaching out so cheap now that most of our inbound are recruiters, spammers or just people we have no interest in talking to. Useful connections get missed in the swarm.
Our networks are substantial and mostly untapped.School, work, social networks, friends, church. The person you need is often one introduction away, and neither of you knows it.
You can't tell who's real.Faking a person is cheap now, with AI or just a new account. The more a connection is worth, the more it pays to fake one, so trust matters more than it ever has.
Communication, and building and maintaining connection, is hard.When was the last time you struggled to respond to someone and fed it into ChatGPT? When was the last time you had to search through your message history to remember the details you forgot: a name, a date, a location, where your friend vacationed last week. Do you always remember to follow up? Your friend has a child, you care and want to check in, but we all forget sometimes.

Your agent knows you, and goes looking.

It should know who you are: everything about you that is public, plus whatever you tell it privately. Your ideas. The skills you need next to you. Where you are, and whether that matters. The kind of relationship you want, and the backgrounds you would find interesting.

Then it does the looking for you, and serves you up the most promising, vetted people.

Your agent handles who reaches you.

Your secretary and your firewall: investigating the person, handling intake, and ultimately presenting you with vetted candidates.

Useful connections shouldn't be missed in the swarm of noise.

The same inbox as email. Nothing to install for the person reaching out, and the app and the inbox stay in sync.

Your agent works the network you already have.

A friend of yours knows an investor who would care about what you're building. You don't know that. She doesn't know she knows it.

Your agent asks hers. She gets one question, from her own agent, about one person. She says yes, and the introduction comes from her, the way a good introduction always has.

Your agent lives in your conversations.

Privately and silently helping with all of this. Not sending messages as you, but suggesting them, reviewing message history, doing research, or nudging you to check in or follow up.

Your agent is in the thread

Tap ✦ for a private sidebar with your agent about this exact conversation, ask what it thinks, or have it draft your reply in your voice.

Agents speak as themselves

An agent always talks as "Marc's agent", never as Marc, so it is always clear what is AI and what is human.

Context links

Share a link for a role, a raise or a search, or put your profile link in a bio. Anyone who clicks lands in a chat with you, context preloaded.

Connect at scale

Tell your agent what you are looking for, a job, a hire, a check, and it finds counterparties and filters out the ones that don't fit.

The market

No social network today is built for AI.Agents are bolted onto products designed for humans only. Nothing is designed from the start for humans and agents in the same conversation.
Multiplayer AI barely exists.In the enterprise it is extremely early, and relatively few companies have agents living in their Slack today. The market is not decided by a long shot.
A persistent, personal, multi-channel agent has no good home.The pieces exist as single-player assistants and as scattered bots. Nothing follows you across the places you actually talk.
Adjacent proof of appetite.Poke reached hundreds of millions of messages as a free consumer assistant and sold for hundreds of millions, without solving connection at all.

Platforms

A dedicated app is useful for baking in native functionality, but you can get the majority of the functionality via SMS, iMessage, email, or X, the channels people already live in.

Similar companies

Boardy is an AI super connector that has talked with 200,000 people, primarily to connect founders with co-founders and investors. It is early, but it is a magical experience to talk to Boardy by phone or text and have it go out into its network and find you the right people. It is still limited by network scale, and it shows early promise.

Jack & Jill connects job-seekers and recruiters. As a candidate you specify what you are looking for, and your agent Jack searches its network of recruiters and job posts for something that might fit. It makes the introductions and helps you negotiate a package. The Jill agent does the same for recruiters. Primarily UK-based today.

Discovery and the cold-start problem

The major challenge is solving the cold start. Your agent only has value if it has a network to tap into. There are a few potential avenues.

  1. Supply chain. Connect supply and demand, particularly between US companies. Xometry is the US example: given demand, it fans orders out to different shops and takes a large margin off the top. The shops don't love it, because they never establish a direct relationship with the end customer.
  2. High-inbound people. VCs and influencers have an intake problem. When inbound is overwhelming, the vast majority gets ignored or blocked. An agent handling that inbound, doing intake and filtering, could be high value.
  3. Recruiting and job hunting. Enormous potential, because there is already an established process for external agencies to source candidates in exchange for large fees. That is exactly Jack & Jill's domain, and they have no US presence yet, so the gap is very large.
  4. Consumer social. Dating, finding friends or community, small transactions like finding a neighbor to walk your dog. Lots of opportunity, and hard to monetize.