I posted a boring intro in a 456k-person community. One line got all the reaction
My team and I are building Mavibot, a platform for AI agents, workflow automations, a CRM, a site builder and a store builder, all in one place. Last week I posted an intro in AI Automation Society, a Skool community with about 456k members. Nothing special, the usual "here's what I'm building" post.
Most of it was a feature rundown. One paragraph mentioned an onboarding agent, the thing that walks a non-technical user through a conversation and gets them to a working funnel, site or CRM setup without them touching any settings. I wrote that line almost as an aside.
Within a day it had 30 likes and 21 comments. All but a couple of them were about that one line. Nobody asked about the store builder, the CRM, or the automations, even though I'd described all of it.
People asked things like: how does the agent handle a vague request? What happens once the thing is live and a customer needs a human? How do you keep that "just talk to it" experience reliable across different businesses? Same three or four questions, from different people, all pointed at the same feature.
None of them were asking whether AI-builds-your-funnel is a good idea. They already believed that. What they wanted to know was whether it breaks, and who's watching when it does.
That's the part I didn't expect. I assumed the pitch was "everything in one place instead of ten tools." Turns out the pitch that actually lands is one specific, easy to picture moment: someone who doesn't know how to build anything, typing what they want, and getting a working thing back. A feature list makes the reader do the work of imagining the value. One concrete scenario does the imagining for them.
We're now putting together a deeper piece just on that one mechanism instead of spreading the story across the whole platform.
Curious if others have had this happen, one line in a bigger pitch accidentally becomes the whole thing people care about. What did you do once you noticed it?
The tell isn't that they asked about the onboarding agent, it's that every question was about failure modes. Running Microsoft migrations for two decades, the feature never closed the deal. What closed it was answering who picks up the phone when it breaks at 2am. I'd write your deeper piece about the handoff and the failure path, not the magic moment.
Yes, almost exactly this.
In June I posted a feature list here for my lead tool: signal detection, enrichment, a growing dataset, outreach drafts, mobile research. Two people wrote long replies and both skipped the list and went straight for one phrase. One put it as the difference between who exists and who is moving right now. The other told me I would land in the "another lead tool" bucket unless I led with that.
What I did was take it literally. I stopped describing features and asked what would have to be true for "moving right now" to hold up. It needs a date. So I measured where dates actually come from, on 60 companies: RSS, the WordPress API, JSON-LD, og:modified, the Last-Modified header, news search, job ads. Six of the seven were close to empty. The official business register was the one that worked, with 7 percent of leads carrying a dated event in the last 30 days and 20 percent within 90.
So the aside did more than change the pitch. It decided what I built for the next three months.
It happened again this morning on Reddit, at a smaller scale. I posted new-company numbers for Norway and mentioned in passing that only 36 of 2,589 new companies had listed a website. Every question in the thread ended up being about contact details. How many put in a mobile number? 67.
One thing I would add to your approach: when one line takes over, check whether the thing people are excited about exists in the product at the level they are picturing. For me it did not, not fully. The excitement was a spec.
Really good addition, thanks for sharing that. The "was it a spec" question is the one I hadn't thought to ask myself yet.
In our case the onboarding agent line was already a real, shipped feature when people reacted to it, not something on a roadmap. So that particular risk didn't bite us this time. But your point stands as a general check I'll actually start running: before treating a reaction as validation, ask whether what people are picturing matches what exists, not just whether the feature exists at all.
The "moving right now" story is the sharper example though. That's not just people liking a phrase, it's people telling you what would make the phrase true, which is way more useful feedback than a feature request. Curious how long it took you to go from "I need a date" to actually mapping which of those seven sources held up. Feels like the kind of thing that could've eaten months if the business register hadn't turned out to work.
You already have a useful outline for the follow-up piece in those three recurring questions. Keeping them in the readers' own words as headings would make it easy for the same people to find their answer. Hope the deeper piece goes well.
Good call, hadn't thought of just lifting their exact wording for headings. Easy fix that'll probably make the piece land better too. Thanks
This is a good reminder that distribution often surfaces the user’s job-to-be-done before the founder’s category label does. I’d cluster the questions by underlying trigger, then rewrite the lead around the most repeated trigger and test whether it attracts the same quality of replies in a second community.
That's basically the test I want to run next. Same lead, different community, see if it pulls the same kind of reaction or if this one was specific to that group.
Good test. Keep the lead line frozen and only change the room. If it pulls again, you learned something about the line. If it dies, you learned something about that community’s attention, not that the line was lucky noise.
One constraint that keeps the result clean: same timing window, same format (no extra polish), and write down what “reaction” means before you post — replies, DMs, or just upvotes. Otherwise the second run is a different experiment.
Would be interesting whether the second community is closer to builders or to general career chat. That usually moves the result more than vibe.
The repeated questions might be as useful as the feature people reacted to. They seem less worried about whether the agent can build something and more about what happens when it gets things wrong. That feels like a pretty strong signal that reliability and human fallback should be part of the pitch too.
Yeah, that's the part that stood out to me too. Almost nobody asked whether it can build funnels, they all asked what happens when it's wrong or doesn't have enough to go on. Reliability and the human fallback might end up being a bigger part of the pitch than the building part itself.
The reaction is strong enough to treat as more than a copywriting signal, especially with the same questions repeating. Have any of the people engaging with that onboarding-agent message actually tried the product or asked to use it, or is the evidence still limited to interest in the idea?
Fair question, and I want to be straight about it: right now it's mostly interest and questions, not people actually using it. A couple asked about connecting or exploring further, but nobody's come back with "I tried this and here's what happened." So it's real engagement, just not usage proof yet, worth keeping that distinction clear instead of treating comments as adoption.
The “asked to connect” signal is still worth following, even though usage proof isn’t there yet. If you’re open to it, what’s the best email to reach you on?