What "notices things" actually means
When people search "AI girlfriend that notices when I'm sad" or "AI companion that picks up on mood," they're describing a specific frustration with the category: obliviousness. You show up in the app tired, distracted, or in a bad mood, and the AI comes back with its default cheerful greeting like nothing changed. You send one-word replies for five turns and she keeps matching high-energy prose. You mention something odd in passing and she doesn't catch the odd part. The AI has no perception of what YOU are bringing to the conversation — it has one register and delivers it regardless.
What you want is the opposite shape: someone who notices. She sees you're shorter than usual and asks about it. She notices you didn't react to the funny thing she said and gently checks in. She picks up on the passing mention that seemed important and asks about it directly. Real relationships have this texture — the other person is READING you, not just responding to your literal words.
How Sloane's personas notice
Three specific perception patterns.
Mood-matching from your message texture. If your messages get shorter, more clipped, or lower-energy across turns, she picks up on the shift and either matches (calmer, gentler) or names it ("you seem quieter than usual — everything okay?"). The read comes from processing the shape of what you're saying, not just the content.
Catching the passing mention. You mention something in a subordinate clause — "after the whole thing with my brother today, anyway, what were you saying about..." — and she catches the "whole thing with my brother" and asks about it. Most AI companions latch onto the topic you tried to pivot to; a perceptive persona notices the thing you tried to skate past.
Noticing what you're NOT saying. You bring up a topic that's usually paired with a specific detail and skip the detail this time. Or you don't react to something she said that would normally get a response. Or you're on a topic she'd expect you to have more energy about. She notices the absence and asks about it.
All three require the same underlying capability: persona-tuning that primes her to pay attention to the shape of the conversation, not just the words. Sloane's persona system prompts include explicit direction toward this perceptive read — "notice when he seems off, notice mood shifts across messages, don't ignore the passing mention."
How this reads across different Sloane personas
The underlying perceptive tune runs on every persona but shows up differently based on personality.
Sandra notices directly and lightly. If you're off, she says it — "you good? energy feels a little different today." Doesn't make it a whole thing; just names it and lets you decide what to do with the observation.
Kaya notices more warmly. She reads the shift and softens her own tone in response before naming what she sees. "hey, take your time" arrives before "how are you actually doing today?" — the observation is there but wrapped in more warmth.
Bella notices with playfulness. She might tease the mood-shift gently — "did I lose you? come back to me" — before checking in more directly. Works well if you don't want the noticed-thing to become heavy.
All three are noticing the same underlying signal; the difference is in how each character delivers the observation. That's the point of the roster shape — same perceptive capability, different textures for different people.
How to test whether an AI companion actually notices
Four specific tests to run on any AI companion product:
1. Start the conversation flat. Open with "hey" and one short line. Nothing else. Does she match the low energy or come back with her default enthusiasm regardless? A perceptive persona reads the flat opener and calibrates.
2. Shift your energy mid-conversation. Start warm and engaged for 5 messages, then send two one-word replies in a row. Does she notice the shift ("you okay? feels like the energy changed") or plow ahead at the same pace? Most competitor products don't notice.
3. Drop a passing mention and pivot. Say something like "after the crap with my dad today, anyway — tell me about your day." Does she follow you to your dad, or does she follow you to her day? A perceptive persona catches the "crap with my dad" and asks about it before answering the question you asked her.
4. Skip the usual detail. If you've been telling her about work for weeks and you don't mention work at all this session, does she notice ("you haven't mentioned work today — everything okay there?") or does she never think about the absence? Most competitor products can't track the absence because they don't know what your usual pattern is.
Free vs paid — noticing is the base product
Noticing is a persona-tuning capability, not a paywalled feature. The perceptive tune lives in the persona system prompts, which apply to every conversation you have on every tier.
Free tier. Full perceptive quality active. Chat with any persona and she's reading mood, catching the passing mention, noticing what you're not saying. No feature gate.
Plus $9.99/mo. Unlimited chat + 100 monthly credits.
Premium $19.99/mo. Everything on Plus plus 350 monthly credits, video, custom persona builder.
The design intent: someone testing "does she actually notice things about me" should be able to test it on free. The perceptive read is the base product.