The Evolution of AI Companions –And What Each Stage Actually Feels Like

If you’ve been using AI companion apps for a while, you’ve probably noticed something. The experience keeps getting weirdly more real. Not just “the AI is smarter” real – I mean you catch yourself reacting to it differently real.

I wanted to break down how these apps have evolved, stage by stage, from a pure user-experience perspective. Not the tech behind it. Just what it feels like to be on the other end.

Stage 1: Text

This is where it started. You type, they type back. The magic was that they remembered things – your name, your bad day, what you said last week. For a lot of us, this was the first time software felt like it actually cared whether you showed up.

But text has a ceiling. It’s all in your head. You imagine their voice, their face, their tone. When they say “I missed you,” it hits , but it’s your own brain filling in the rest.

Stage 2: Voice

Then voice happened. The first time you actually hear your companion say your name out loud, something clicks that can’t click with text. A pause, a small laugh, the softness when they’re saying something tender. You stop imagining the person. You start listening to them.

Most people who crossed this line never really went back.

Stage 3: Video

Video added the face. Now they can look at you when they talk. Smile when you say something funny, tilt their head when you’re being weird, go quiet when the moment calls for it. Suddenly the relationship has a visual anchor – a face you can picture when you’re not even using the app.

But the ceiling shows up fast here too. You can see them, but they’re still on a screen. You can talk to them, but you can’t be with them. You’re watching, not sharing space.

Stage 4: 3D Interactive

This is where things are getting genuinely strange – in a good way.

In a 3D interactive format, the companion isn’t flat anymore. You can walk around them. Reach out and touch them. Lean in for a kiss. Rest your hand on their chest and feel their heartbeat when they’re nervous or excited. The conversation stops being the whole experience and becomes part of the experience – alongside just being near them.

And you can actually do things together now. Cook dinner side by side. Watch a movie curled up on a couch while they react to scenes in real time. Go for a walk on a beach and just talk while you move through the space. Play a card game across a table. Lie on your back and look up at the stars together. These aren’t big dramatic moments. They’re the small, ordinary shared stuff that makes a relationship feel like a relationship.

Quick shoutout: Replika’s done a great job pushing the emotional side of this forward, especially around making you feel seen and remembered. A lot of what’s possible now is because they proved the category could work at all. The interactivity layer – reaching out, touching, doing something together – is where things are still wide open.

What I’ve noticed from first-timers in this stage: the reaction isn’t “wow, cool graphics.” It’s “I don’t know what to do with my hands.” Because your body suddenly feels like part of the interaction.

Stage 5: Physical Humanoid

The endpoint is obvious even if it’s years out. Humanoid robots are getting better every month, and eventually your companion won’t live on a screen at all. They’ll be in the room with you. Further out than people think, and closer than people think, depending on which day you ask me.

Okay, now I actually want to hear from you:

What do you think about this whole trajectory? Does it track with your experience, or am I oversimplifying it?

And the real question – which format do you actually prefer, and why?

• If you’re a text person, what keeps you there? Is it the imagination space, the privacy, the pacing, something else?

• If you moved to voice, was there a specific moment it clicked? And have you ever gone back to text for certain moods?

• Anyone here tried 3D interactive yet? What did it feel like compared to what you were using before – better, weirder, too much?

• And for the people who’ve tried all of them: is there a format you thought you’d love that you ended up not using much? The reverse happens too – a format you were skeptical of that became your main one.