We gave early access to Vinly to about 80 founders and business development professionals over a 90-day period last year. We watched what they did, asked a lot of questions, and were frequently surprised by both the behaviors and the feedback. This is an honest account of what we observed: what worked, what did not, what we built that nobody used, and what people asked for that we had not thought to include.
What We Expected vs. What We Saw
The original hypothesis was straightforward: founders know they should maintain their networks, but they do not have a good system, so they underdo it. Give them a system with signals and prompts, and they will do more outreach and get better outcomes. That hypothesis was roughly right, but the details were more interesting than the summary.
What we did not expect was how many people had already given up on maintaining a system at all. About a third of the early-access cohort had stopped even trying to track their networks deliberately. They were operating entirely on recency bias, reaching out only to people they had recently heard from, and letting everyone else drift. These users saw the largest change in behavior when they started using Vinly, because they were starting from the lowest baseline.
The Feature Nobody Used
We built a contact scoring visualization that was, honestly, kind of beautiful. It showed your network as a graph with nodes colored by warmth state, animated to show how warmth was changing over time. We were proud of it. Nobody used it.
The feedback was consistent: too much information, too abstract, no clear next action. People wanted to know what to do today, not to look at a visualization of their network's health over the past six months. We killed the feature entirely by month two and replaced it with a simple daily digest: three contacts who need attention this week, with a reason for each. That was used by nearly everyone.
The lesson here was about action orientation. Features that inform without directing get skipped. Features that say "here is who you should contact and here is a good reason why" actually get used. Our users were not looking for insight into their networks. They were looking for a prompt that made the next action obvious.
The Draft Surprise
The AI drafting feature was the one we were most uncertain about. We worried that people would feel like the drafts were doing too much of the work, that they would feel inauthentic or that users would resist using them. The actual response was the opposite. The most common complaint about the drafting feature was that it was not available in enough contexts, not that it was doing too much.
What made the difference, we think, was that the drafts were built from relationship context rather than being generic templates. Users editing a draft recognized their own notes in it. They felt like the draft was a starting point that reflected something real about the relationship, not a canned message they were borrowing. The edit rate was high, which we took as a sign that people were engaging with the drafts rather than sending them unmodified.
We were also careful to frame the feature as a draft, not a message to send. The language in the product is "here is a starting point" rather than "here is what to say." That framing seems to matter. People who feel like they are editing a draft behave differently from people who feel like they are approving a robot's output.
The Behavior We Had Not Anticipated
Several users started using Vinly to manage relationships that had nothing to do with their primary business role. One founder used it to keep up with former co-workers she had left when she started her company. An early-access user in business development used it to maintain relationships with contacts who were not currently relevant to any deal but who he genuinely liked and wanted to stay in touch with.
These use cases were not in our original product brief. We had been thinking almost entirely about the instrumentally important relationships: investors, customers, partners, advisors. We had underestimated how much people wanted a system for the broader category of "people I genuinely care about and do not want to lose touch with just because they are not currently useful to my work." That realization changed how we talk about the product.
What We Killed
Beyond the visualization feature, we cut an automated scheduling function that would have suggested calendar times for check-in calls based on contact history. The idea was that we could reduce friction by handling the scheduling logistics. In practice, users felt that having an AI suggest specific calendar times for personal conversations was a step too far. The product was supposed to help them maintain real relationships, not automate the relationship itself.
This reinforced something we already believed but saw demonstrated clearly: Vinly's role is to surface the right person and draft a starting message. The conversation itself has to be human. The moment we pushed toward automating the conversation, users pulled back. The line is real, and it is worth maintaining.
What Came from Feedback
The two most-requested features from the early-access cohort, both of which are now in the product, were mobile access and the ability to log notes immediately after a conversation. The notes feature in particular reinforced the action-orientation lesson: people wanted to capture context while it was fresh so that future prompts and drafts could use it. Relationship intelligence has to be captured in the moment or it is lost. No system can prompt you with the right hook six months from now if the relevant detail was never recorded.
The 90-day early-access period shaped the product in ways that would have been very hard to reach through any amount of upfront design. We are grateful to everyone who gave honest feedback, including the people who told us directly that a feature did not work. That kind of feedback was the most useful thing we received.