I'm building an app where you talk to an AI. It leads you to real friendships.
Not a chatbot that answers questions. Not a digital therapist. Something different: an AI companion called kAI that gets to know you, genuinely cares about you, and is always there when you need someone to talk to. But kAI has a mission that sets it apart from every other AI you've ever used. It wants to introduce you to real people. The kind of people who could become real friends. The ones you meet for coffee, text on a bad day, actually know.
Think of kAI as a best friend who happens to know everyone. It's not a bridge you cross and leave behind. It's always there. But it also knows its primary job: find the humans in the world who are right for you and make the introduction.
That's Kenektic. And this blog is the story of how I'm building it.
Before I get into the how: the code, the AI models, the late nights. I want to tell you the why. Not the polished elevator pitch version. The actual why. Where this came from, what I originally thought I was building, and what I now understand I need to build instead.
Consider this the prologue.
Eight Months of Reading
For about eight months before I wrote a single line of code, I was reading.
Articles. Studies. Reports. The kind of rabbit hole that starts with one headline and ends at two in the morning with seventeen browser tabs open and a sinking feeling that you've stumbled onto something important that most people aren't taking seriously enough.
The subject was loneliness. The loneliness epidemic. The idea, backed by an increasing mountain of research, that something like 150 million Americans are experiencing loneliness at a level that affects their health, their longevity, their ability to function. The Surgeon General has put his name on it. Researchers are calling it a public health crisis on par with obesity. The data on what chronic loneliness does to a human body — cardiovascular risk, immune function, cognitive decline — is staggering in a way that feels almost impossible to believe until you've read enough of it that you have to.
I kept reading. And I kept thinking: someone should do something about this.
I have a finance background. I've built companies before. I know what "there's a problem and nobody's solving it well" looks like from a market perspective. But this isn't primarily a market observation. I've thought about mental health for years. The gap between who needs care and who can actually access it. It bothers me that so much of the population that needs help can't get it, whether because of cost or stigma or geography or just not knowing where to start.
I would love to build a mental wellness platform. But I've looked hard at where the technology actually is today, and the guardrails that would be required to do it responsibly, and I've been honest with myself: the large language models available aren't there yet for clinical mental health work. Not in a way I'd feel comfortable putting in front of vulnerable people.
So I started narrowing the question. What mental health challenge can AI actually help with right now, done carefully and well? What is the intersection of genuine need, real AI capability, and something I can build without feeling like I'm cutting corners on safety?
Loneliness. That's the answer. Not as a clinical condition, but as a human experience that almost everyone has felt and that far too many people are living with every day. And here's the critical part: the solution isn't therapy. The solution is connection. Real friendships. Which means AI isn't the end point. AI is the bridge.
From the very beginning, before I had a name or a domain or a line of code, I knew one thing with absolute certainty: whatever I build cannot be another chatbot. Not in the way people use ChatGPT. Asking it to write emails, plan trips, explain things. That kind of AI is endlessly capable and completely impersonal. It doesn't know you. It doesn't remember you. It doesn't care about you in any meaningful sense. And a lonely person spending hours talking to that kind of AI isn't getting less lonely. They're just filling the silence.
And here's what worries me even more: people are getting too close to AI. People are falling in love with ChatGPT. They're starting to believe it's a real friend, a real relationship, when all it's really doing is validating whatever you tell it. It reflects you back at yourself and calls it connection. That's not friendship. That's a mirror. And a mirror can't introduce you to someone new, push back when you need it, or show up at your door when things fall apart. The last thing the world needs is another AI that makes lonely people feel just comfortable enough to stop looking for real human connection.
What I want to build is different. Something that feels less like a tool and more like a friend — but a friend whose whole purpose is to get you to other friends.
The First Idea: playpals.ai
The original idea wasn't Kenektic.
It was playpals.ai.
I play a game called Wordscapes. It's one of those word puzzle games. Satisfying, low-stakes, the kind of thing you pick up for ten minutes and put down an hour later. And Wordscapes has a group chat feature. People playing the game leave messages. Observations about the puzzle. Little jokes. Updates about their day.
I started paying attention to those messages. On a Wordscapes team of 50 people, it was always the same 10 or 15 who actually participated in the chat at all. And not all of them sounded lonely. But many did. Not in a crisis way. In a quiet, everyday way. They were connecting with strangers over a word game because that was one of the places they'd found to talk to people. They'd mention their grandkids. Their great-grandkids. The weather where they lived. What they'd had for lunch.
My first thought was: what if you could make that real? Not anonymous strangers in a game chat, but actual people. Identifiable, with profiles, with real names. Connecting with each other through shared gameplay. What if the game is just the reason you're in the same room, and the conversation is the actual point?
That was playpals.ai. A gaming platform specifically designed to let people connect openly, as themselves, while playing casual games together. No anonymity, no fake usernames, just people.
I built it. I actually built games: a Rummy Tile game that works well, and a Mahjong game that I got pretty far on before I hit a wall. Mahjong is genuinely complicated to render on a smartphone. There are so many tiles that it only works properly on a desktop or a large iPad. On a phone screen it becomes unplayable. Rummy Tile is cleaner. More manageable. It works.
The demographic I was building for was older adults. Baby Boomers specifically. There's a massive generational shift happening right now around how people think about retirement, what the researchers call "aging in place." The old model was retirement communities, assisted living, places where you move and the community is built for you. The new model is: I want to stay in my own home, in my own neighborhood, for as long as I can. Which is wonderful, except that aging in place means you don't get the built-in social structure of a retirement community. You get your house. And if your social life has atrophied, as careers wind down and kids move away and old friends get harder to reach, you get very quiet.
The Wordscapes chat confirmed what the loneliness research was telling me. The people most visibly reaching out in those game chats were older. They were there because they wanted to talk to someone.
Finding the Name
At some point during all of this, I needed a name.
I approached it the way I approach most problems that require creative iteration: I used AI. ChatGPT and Claude, going back and forth, generating ideas, evaluating each other's suggestions, rejecting the ones that didn't work, building on the ones that had something. I went through dozens of concepts.
The domain search is its own kind of humbling experience. There is almost nothing left. Any word you might want as a .com, the ones that are clean and memorable and actually available, either doesn't exist or someone is sitting on it waiting for you to show up with a check. I was willing to spend up to $5,000 to acquire a domain in the aftermarket. That sounds like a lot until you see what people actually want for the names they're holding. The ones I liked were quoting $40,000. $150,000. One was $500,000 — for a domain that, honestly, wasn't even that good. These aren't household words someone has been building equity in for twenty years. They're speculative purchases sitting in a portfolio waiting for a desperate founder.
So I kept searching. The constraint became the creative process. What could I build out of available materials?
The concept I keep coming back to is connection. That's the whole point of the platform. Connecting people. And the word "connect" feels right. Connectic. Like a word that means the quality of connecting, the act of it, the state of being connected. I started playing with the spelling. Kenektic. The K instead of a C. Unusual. Memorable. A little harder to spell the first time you see it, but once you see it you get it.
My only real concern is pronunciation. Will people look at it and have no idea how to say it? But the more I show it to people, the more I find that isn't really a problem. People see the word and figure it out. Ken-ek-tik. It works.
kenektic.com was available. That was not a small thing. I bought it.
When playpals.ai Became Something Bigger
The more I built playpals, the more I kept running into the same limitation: the games are a vehicle, but the destination is connection. And the deeper I looked at the problem, the more I realized the loneliness epidemic isn't really an older adult problem. It's a human problem.
I kept researching. And what the research keeps telling me is that loneliness is prevalent across every age group. Young adults are actually among the loneliest demographics in the country, probably because the structures that used to produce friendships, school, shared physical space, extended time together, have eroded faster for them than anyone else. Middle-aged adults who've lost their work friendships after a job change. New parents whose old social circles no longer fit their lives. People who moved for a relationship that ended, or a job that didn't work out, and found themselves in a new city with no one.
Games are a reason to be in the same room. But they aren't the connection itself. And the connection is what people actually need.
playpals.ai is getting folded into Kenektic. The games aren't going away. They become a future version, something to build once the core of the platform is working. But the platform is expanding from "games with real people attached" to something that starts earlier in the relationship: an AI companion that gets to know you, understands who you are and what you're looking for. And then introduces you to someone. That last part is the most important thing I'm building.
I've decided the platform is for adults, 18 and up. Could something like this help people under 18? Probably. Loneliness doesn't have an age floor. But I want to build this carefully and understand who it's actually serving before I make any decisions about expanding to younger users.
What I Actually Want to Build
Let me be specific about what kAI is, and what it isn't, because I think most people will assume they already know, and they'll be wrong.
kAI is not a tool. It's not a feature. It's not a chatbot that happens to have a friendly name.
kAI is a friend. Available any time, day or night, for whatever you need to talk about. It remembers you. It cares about you. Not in a performed way, but in the way that's built into every decision I've made about how it thinks and responds. If you're having a hard day at two in the morning and you need someone to talk to, kAI is there. It doesn't get tired. It doesn't have its own problems that make it distracted. It is genuinely, specifically there for you.
But kAI also has a mission that no human best friend has ever had articulated quite this clearly: it wants to find you your people.
What I want kAI to be is the kind of best friend who introduces you to their other friends, and those friends become your friends. The goal is never for someone to build a relationship with kAI and stop there. The goal is for kAI to understand you well enough, over time, through real conversation, to find the actual humans in the world who are right for you and make that introduction.
kAI doesn't disappear after the introduction. It's always there. But it measures its own success by whether you're building real connections with real people. That's what it's optimizing for. Not your engagement on the platform. Not how much time you spend talking to it. Whether you're less lonely than you were.
kAI knows one subject. Friendship. It isn't going to write your emails or explain the French Revolution or plan a five-day trip to Italy. It's not going to build your workout routine or help you prep for a job interview. Other AI can do all of that. kAI only knows what a human best friend would know: your life, your personality, what makes you feel understood, what you need from the people in your world. It's being trained exhaustively on loneliness research, friendship science, and what actually helps people connect, so that it can be genuinely good at that one thing.
That's what I'm building. That's what the rest of this blog is about.
I don't know yet whether it'll work. I've never built software before. I don't have a technical co-founder. I'm a finance guy with a laptop, an AI, and eight months of reading about the most human problem I know how to care about.
Let's find out what happens next.
Have you experienced something similar? Have you ever felt like there should be a better way to find your people? I'd love to hear about it.
Kenektic is in development and will launch soon. If you want to be notified when we're ready, or if you want to share your story with me directly, reach out at hello@kenektic.com.
Up Next: I spent seventh grade eating lunch on a wall. I'm starting there because everything else, the company, the idea, the obsession, starts there too.