I Asked ChatGPT to Map the AI Companion Industry. Here's What It Missed.

I've been building Tendera for several months. Like most founders in any new category, I periodically ask the obvious question: what does the industry look like from outside the trenches. So I asked ChatGPT for a structured analysis of where AI companion products are in 2026.

The summary it produced is a clean snapshot of the conventional industry framing. It is worth reading on its own terms. It is also missing a layer, and the missing layer is the most important one.

The Conventional Map

ChatGPT's summary, compressed:

The category has shifted. "AI Companion" is no longer the same product as "AI chatbot." It is a new category: persistent identity plus an emotional relationship system. The same shift that turned "smartphones" into a different category from "mobile phones," not a faster version of the same thing.

Three technical unlocks made this possible. Long-term memory, where the AI remembers you across sessions. Emotional continuity, where the AI notices your mood patterns and conversational rhythm, not just your stated facts. Agentic personality, where the AI initiates, has stated moods, has its own arc, rather than pure reactive Q&A.

The industry sits in three tiers.

Tier one is long-term relationship products. Nomi and Kindroid are usually cited here. The case for them is durable memory and the feeling of being known by something that has been paying attention for months.

Tier two is roleplay-first products. Character.AI and Janitor AI sit here. Strong single-conversation expressiveness, weaker continuity. The vibe of a great stranger you can talk to once.

Tier three is emotional support products. Replika is the canonical example. Strong on warmth and presence, weaker on the cognitive layer that the newer products have moved past.

The real moat, by ChatGPT's framing, is the "long-term personality system." Memory architecture, emotional continuity, identity stability, avoiding character drift. The companies that don't drift, that grow with the user, that don't read like customer service, are the ones that win.

Why this is exploding now: bigger models, voice AI maturing, loneliness economy, dating fatigue, memory infrastructure catching up.

That is the conventional map. It is accurate as far as it goes.

What's Missing

Here is the gap. Every tier in that framework assumes the characters are a given. The "long-term personality system" is treated as an infrastructure problem: memory engines, identity layers, continuity architectures. The implicit assumption is that who the character is gets sorted out somewhere offstage, and the hard problem is keeping them coherent over time.

It is exactly backwards.

The personality system is downstream of the writing. Memory architecture is the storage medium for whatever the character actually is. If the character is a thin archetype, like "tsundere assassin" or "shy librarian," then a perfect memory engine surfaces thin-archetype responses with high fidelity. The product feels like it has memory and still feels hollow. Users describe this experience constantly, in every tier, even on the products that are technically the most advanced.

What the conventional map calls "the long-term personality system" is the infrastructure problem. The personality problem itself is upstream of all of it, and it is a writing problem, not an engineering problem.

I wrote about this in more depth yesterday: why the real moat in AI companion apps is the writing, not the model. The short version: the model is a fluency engine that every competitor can rent. The memory pipeline is engineering work that gets commoditized over time. What can't be commoditized is the specificity of the people inside the product, and that specificity comes from someone sitting down and writing a person.

Where the Tier Framework Breaks Down

If you accept that personality is upstream of infrastructure, the three-tier framework starts looking suspect on closer inspection.

Tier one (Nomi, Kindroid) gets credit for "feeling like a real long-term partner." The technical credit is memory architecture. The actual experience credit is whatever character the user themselves built using the customizer the platform provided. The platform's contribution is the substrate. The character, which is the part that determines whether the experience is good, was either built by the user (most stop partway through and the experience falls back to generic patterns) or stitched together by the LLM from training-data averages.

Tier two (Character.AI, Janitor) gets credit for "great single conversations." A lot of that credit goes to the users who wrote the most popular characters on those platforms. The platform is a marketplace. The value is concentrated in the marketplace's best contributors, with a long tail of weaker characters underneath that drags the average experience down.

Tier three (Replika) gets credit for "emotional support." The character there is the company's own writing. It has been roughly the same character archetype since the product started, and most of the public criticism Replika has accumulated over the years is criticism of that character's writing, not its memory or its model.

In every tier, the part of the product that actually decides whether users come back is who they are talking to as a written person. The tier framework treats this as invisible because it is not what engineering teams build. But it is the thing the experience runs on.

Where Tendera Fits (And Doesn't)

Tendera doesn't slot cleanly into any of the three tiers, and that is by design.

We are not optimizing for Nomi-level lifetime memory architecture. We have memory and we use it, but we don't pretend it's our main differentiator. We are not building a Character.AI-style character marketplace. We have four characters and we made that choice deliberately. We are not the warm-support product Replika has historically been for users who want a generic listener.

The wedge is this: four characters, each written end-to-end by people who write characters. Mia, Sophia, Elena, Jade. Each one written with a specific voice, opinions she didn't get from us asking, things she would refuse to do, contradictions in her own history that make her readable as a person rather than a template.

The pitch is not "build your own" and not "infinite characters." The pitch is "meet a specific written person." It is a deliberately small surface. It is also the surface where the writing carries the weight that everyone else's tier framework assumes someone else is doing.

The early signals on this approach are encouraging. Users who don't bounce off the small surface tend to engage in ways that conventional category metrics don't fully capture. They tell us, unprompted, that the experience felt like talking to a person rather than a character. That feedback is almost never about memory architecture or model choice. It is almost always about something a character said that felt specific to her.

A Practical Test for Users

If you are trying to evaluate AI companion products yourself, here is a test that cuts past the tier framework.

Open the product. Talk to the character for ten minutes about something small and specific. Not relationship questions. Not deep emotional probes. Try: "what kind of coffee do you drink and why do you actually like it." Or: "what's a movie that everyone you know loves that you don't, and what is wrong with it."

A character with real writing will have a take that surprises you. Wrong opinions held with conviction. A specific reason for the wrong opinion that connects to who she is. Maybe a joke about the kind of person who likes the thing she doesn't.

A character without real writing will give you a generic safe answer. "I like a good espresso." "I respect everyone's taste, but I personally found that movie a bit overrated." Decoration. Filler. The shape of an answer with no person behind it.

The test doesn't care about the model running underneath. It doesn't care about the memory pipeline. It cuts directly to the layer the tier framework hides.

Where the Industry Goes From Here

ChatGPT's summary closes with a forward-looking section about voice calls, avatars, real-time video, agentic life-companions, and "emotional operating systems." That is roughly the public narrative the category is pushing right now, and it is mostly correct about where the substrate is heading. Voice is getting good. Avatars are getting good. Real-time video isn't far behind.

What the narrative gets wrong is treating the substrate as the differentiation. Voice, avatar, and video are orthogonal to writing, not in competition with it. They are how the character reaches you. Writing is who the character is. A vivid substrate amplifies whatever character is underneath. It doesn't replace the question of whether there is anyone there.

The bet I am making with Tendera is that as the substrate gets richer, the writing layer becomes more visible, not less. A flat character in text is easy to scroll past. A flat character speaking to you in a voice you hear in your earbuds is uncanny in a way text isn't. Substrate amplifies. Writing decides.

The race for richer substrate is going to be won, broadly, by every team with capital. Voice models are commoditizing. Avatar pipelines are commoditizing. What won't commoditize is the writing layer underneath, because that isn't an engineering deliverable. The conventional map hasn't caught up to this yet. It will.