The first time one of us tried the voice coach, it felt strange. Talking to a phone, in the kitchen, about a fight we'd had that morning. She kept wanting to stop and type instead — typing felt more controlled, more deliberate, less exposed.
That's exactly why talking is better.
Not because voice is more convenient (it isn't always). Because the part of you that wants to type everything out carefully and control every word is the same part of you that can't quite get to the real thing in a hard relationship conversation. Voice bypasses it. You say more than you meant to, and what comes out is usually more honest.
Quick Answer
Talking out loud gets you to the unedited version of a feeling faster than typing does. Speech is about three times faster than typing on a phone, so the inner editor has less time to rework what you mean, and putting a feeling into words is linked to lower emotional reactivity. We haven't found research showing that voice beats writing for relationship work, so treat this as our experience backed by adjacent findings. Use voice for the raw version and text for drafting what you'll say next.
Key Facts
- Speech is roughly three times faster than typing on a phone. In a Stanford-led experiment of 32 people, English speech input was 3.0 times faster than the iPhone keyboard. People were copying short phrases, so the gap is likely smaller for long, emotional messages.
- Naming a feeling can lower its intensity. In a 2007 brain-imaging study, Lieberman and colleagues found that labeling the emotion in a face reduced amygdala activity compared with other ways of processing it. The study tested labeling, not speaking versus typing.
- Our reading of those two findings: faster output leaves less time for the inner editor, and naming the feeling does some of the calming. That is a hypothesis we act on, not a proven effect of voice.
- The Growing Us voice coach is built for sub-500ms response times, which keeps the exchange conversational instead of turn by turn.
- Text chat with an AI is still useful, particularly for drafting, revisiting transcripts, or when speaking aloud isn't practical.
Why is the inner editor the problem?
Most people have a well-developed inner editor for difficult emotional content. It runs automatically, particularly in high-stakes conversations. It turns "I'm scared you don't want this anymore" into "you've been distant lately." It turns "I need you to notice me" into "you never pay attention to what I do."
This isn't dishonesty. It's protection. The real thing feels too exposed, too much, too likely to land badly.
When you type — especially into an AI, which feels like writing rather than talking — the inner editor stays fully engaged. You backspace. You rephrase. You craft the version that sounds reasonable. The result is polished and incomplete.
When you speak, particularly at conversational pace, the editor can't keep up. Things come out before they've been checked. And the things that come out first, unedited, are often truer.
We noticed this the first time the voice coach asked a follow-up question and one of us answered before thinking. The answer was noticeably more honest than the typed version would have been.
Try it: Think of something you've been trying to say to your partner and write it out. Then close what you wrote, speak the same thing aloud (to yourself, to your phone), and notice what's different. The spoken version often has something the written version doesn't: a feeling, a specific moment, the thing underneath the thing.
Why does the rhythm of real conversation matter?
There's a second reason voice matters for relationship work: it approximates the rhythm of actual conversation in a way that text doesn't.
Your relationship problems don't live in documents. They live in conversations — spoken, real-time, with silences and interruptions and the way someone's voice goes flat when they're actually hurt. Practising relationship communication through text is a bit like training for a marathon on a stationary bike. Related, not the same.
When you talk through something with the voice coach, you're rehearsing the actual modality — speaking, listening, responding — that you'll use with your partner. You're finding out what it feels like to say the real thing out loud, before the stakes are high.
The pause before a difficult sentence, the way your voice changes when you get to the part that actually hurts — these are things you only discover by speaking. And discovering them in a lower-stakes space means they're less likely to ambush you when you're in the actual conversation.
When is text the right tool?
None of this is to say voice is always better. Text has real advantages.
Reviewing what you said. After a voice session, reading the transcript is genuinely useful. You see your own words from a slight distance and often notice things you missed while saying them. Bring the one sentence that surprised you to your weekly check-in.
Drafting messages. If you need to send something to your partner (a hard message, a repair attempt, something you've been putting off), text is the right medium. Our repair scripts and prompts can help with the wording. You want deliberateness there, and a general chatbot can help with the wording, as we explain in why ChatGPT works as a rehearsal partner.
When you can't speak aloud. At work, in public, anywhere privacy matters. A text fallback is essential for real-world use.
Processing slowly. Some people genuinely think better in writing. For structured reflection — listing what happened, what you felt, what you want — text can be more useful than rambling voice.
The practical approach: start with voice for the emotional content, use text for the output. Talk it through; then read what you said and draft what you want to do next.
Try it: After your next voice session, read the transcript and highlight one sentence that surprised you — something you said that you didn't know you were going to say. That sentence is usually where the actual work is.
Why does typing keep you out of the moment?
There's one more thing worth naming: when you're typing at a relationship problem, you're looking at a screen. When you're talking through it, you're in your body, in the room, with the feeling.
This sounds small, and we think it isn't. Psychologists call naming a feeling "affect labeling." In a 2007 brain-imaging study, Matthew Lieberman and colleagues found that labeling the emotion in a face quieted the amygdala more than other ways of processing it. That study didn't compare speaking with writing, and expressive writing has its own research behind it, so our case for voice is partly lived experience.
What we can say is that "I felt dismissed" said out loud is a clear example of labeling a feeling, and doing it before the real conversation is useful preparation. Working out what you mean before you say it is also the idea behind our post on whether AI can help you feel heard. The goal isn't only to understand something intellectually. It's to stay in the conversation with your partner without flooding, and practicing that by speaking is the closest rehearsal there is.
FAQ
I feel silly talking to my phone about my feelings. Is that normal?
Yes, and it usually passes after about a minute. The strangeness is your inner editor objecting to the loss of control. The useful signal is: if it feels vulnerable, you're probably getting at something real. If it still feels odd to use AI for this at all, we wrote about whether that is weird.
What if I say something wrong or embarrassing?
A relationship coach isn't keeping score. Say the wrong thing, then hear yourself say it, and you'll know it was wrong. That's not embarrassing; that's how you find out what you actually think.
Can my partner and I use voice together?
Yes. In couple mode of Growing Us, the two of you talk while the app listens, and it identifies each speaker so the post-session insights can separate what each of you said and brought. If you'd rather prepare alone first, our guide to solo modes and rehearsing a conversation covers that.
Is there a transcript after a voice session?
Yes, you get a full transcript you can review, along with the coach's insights about what happened in the conversation. You can revisit it later. So the voice session gives you the honest content, and the transcript gives you the distance to see it clearly.
What happens in a first voice session?
Sessions run up to 10 minutes. Expect the first minute to feel strange, then the coach asks follow-up questions and you answer out loud. After the session you can read a full transcript along with the coach's insights, and pick the one sentence that surprised you to bring into your real conversation.
Typing keeps you in control. That's useful sometimes, and exactly the wrong thing for processing hard feelings.
Say it out loud. You'll get to the real thing faster, and you'll be better prepared for the conversation that matters.
Growing Us is built voice-first — talk through what's happening with the coach, read what you said, and go into your real conversation knowing what you mean. Start with the free Pulse Check.