We shipped a feature that does nothing. It's our highest-rated one.
Last quarter, we built a feature at Murror that our engineering team jokingly called "the empty room." After a user finishes a journal entry, the AI doesn't immediately respond. It waits. For 30 seconds, the screen shows nothing but the user's own words and a gentle prompt: "Sit with what you just wrote."
No analysis. No reframe. No pattern recognition. Just silence.
We almost didn't ship it. Our product team argued it would feel broken. Our investors asked why we were spending engineering time on a feature that literally does nothing. Even I had doubts -- in a world where every AI product competes on speed and capability, deliberately slowing down felt like a liability.
But we had a hypothesis: the moment right after someone writes something vulnerable is the most important moment in the entire session. It's when the insight is forming. And every time we rushed in with AI analysis, we were interrupting that process.
So we ran the experiment. Half our users got the immediate AI response. Half got "the empty room" first.
The results were striking:
Users in the silence group were 2.4x more likely to add to their journal entry unprompted -- they'd keep writing, going deeper on their own
2. When the AI response did come (after 30 seconds), users rated it as more helpful, even though it was the exact same response
3. Session satisfaction scores were 35% higher in the silence group
4. And here's the one that surprised us most: users in the silence group were more likely to share their experience with friends
The silence wasn't empty. It was productive. People were processing. And when you give someone space to arrive at their own understanding before the AI speaks, the AI's words land completely differently.
This changed how we think about every feature we build now. Before adding any new AI capability, we ask: "Are we helping the user think, or are we thinking for them?" If it's the latter, we either redesign it or we don't ship it.
The hardest part of building an AI product isn't making the AI smarter. It's having the discipline to let it be quiet.
Anyone else building products where the absence of a feature became the feature?
Replies
I've definitely had moments where the first thing I write isn't the real thing I'm trying to say. That extra pause probably helps with that.
Murror
@trevor_nicholas2 So true -- we call that the "draft vs. truth" gap. The first thing you write is often a shield, and the real feeling lives underneath. The pause gives people time to get past the draft.
It's kind of refreshing to see a team test "less AI" instead of adding another layer of automation.
Murror
@paige_lauren2 Thanks Paige! The default in AI right now is "more is better" -- more features, more automation, more intelligence. We wanted to challenge that. Sometimes the most valuable thing a product can do is create space rather than fill it.
The 30-second wait sounds risky on paper. I wouldn't have guessed it would improve satisfaction that much.
Murror
@kyle_bennett6 We didn't expect it either! Our hypothesis was that it might improve depth of entries, but the satisfaction bump caught us off guard. Turns out that when people feel like the AI "listened" before responding (even though it's the same response), the whole interaction feels more human.
Great insight, designing less instead of more is actually harder, I like the discipline of asking whether the product is helping users think or thinking for them
Murror
@bryan_williamson3 Totally agree -- restraint is genuinely the harder design challenge. It's easy to add more features, harder to ask whether each one is earning its place. That question has saved us from shipping a lot of things that would have looked impressive in a demo but diluted the actual experience.
That ques " are we helping people think or thinking for them?" is going to stick with me.
Murror
@hailey_brianna1 Glad it resonated! It's become our team's north star for every feature decision. The moment we catch ourselves building something that thinks for the user instead of with them, we know we've drifted.
interesting experiment, the fact that the same AI response felt more helpful after 30 seconds of silence shows how timing can completely change perception. 😊
Murror
@malani_willa That's exactly what surprised us too! Same words, same analysis -- but the context of having sat with your own thoughts first completely changed how users received it. It's making us rethink a lot of assumptions about what "better AI" actually means in practice.
Hey Mona! This is interesting. I think it depends on the purpose of your product & what value it provides to users. Users go to Murror in order to take time to reflect. The empty room is an impactful feature that helps the product fulfill (or even exceed) users' expectations of it.
I just read your bio - I'm also typical Gen Z and building a product to help people feel connected. It seems like we have similar goals. I'd love to connect further!
The product I'm building is Tabster. It's a free utility app to help make it as easy for people to split the bill after dinner, or divvy up group trip expenses as possible. Each tab or trip added to split becomes a card on the homepage, so the user has a digital journal of their memories with friends. On each split tab or trip memory, users can add photos, comments, likes, etc. for the group to share.
Murror
Really appreciate you framing it that way -- the empty room works precisely because it aligns with why people come to Murror in the first place. When the product's core promise is reflection, silence isn't a gap, it's the feature delivering on that promise.
Tabster sounds like a great concept! Turning a utility moment (splitting bills) into something social with shared memories is a smart angle. Would love to connect -- feel free to DM me here or find me on LinkedIn!