Templates have an honest appeal. Write a good message once, fill in the name and one or two details, send it to fifty people, done. For certain kinds of outreach at certain volumes, templates make sense. The question is whether they make sense for the kind of relationship maintenance that actually moves professional relationships forward, and the evidence from our early-access cohort suggests they mostly do not.
This post covers what we observed when we compared template-driven and AI-draft-driven outreach across users in our cohort over a three-month period. The sample is small, the methodology is not academic, and we are being careful not to overstate the findings. But the pattern was consistent enough that it is worth sharing honestly.
What We Mean by Templates vs. AI Drafts
For this comparison, we defined templates as pre-written messages with variable fields that the user fills in before sending. A typical template might look like: "Hi [Name], I've been thinking about our conversation at [Event] and wanted to share [Relevant Article]. Would love to hear how things are going at [Company]. Let's catch up soon." The user fills in the brackets and sends.
AI drafts, in contrast, are messages generated from the contact's relationship record in Vinly: the specific notes from past interactions, the person's current role and recent professional activity, any context the user has added over time. The draft is not filling in a template. It is producing a first draft of a specific message to a specific person based on what is actually known about the relationship.
Response Rate Observations
Across the early-access cohort, users who primarily used AI drafts saw meaningfully better response rates on their outreach than users who primarily used templates, with a few important caveats that are worth being direct about.
The gap was largest for contacts where the relationship had a rich history in the system: multiple past interactions logged, specific notes about conversations, clear context about the person's current situation. In those cases, the AI draft could draw on genuinely specific material, and the resulting message read like it came from someone who actually knew the recipient.
The gap was much smaller for contacts with thin relationship records. When the system does not have much to work with, the AI draft is not meaningfully better than a decent template, because both are drawing from a limited pool of context. The draft's advantage is context-dependent.
Where Templates Still Win
Templates are not bad for all outreach. They have a legitimate role in contexts where personalization is not the primary goal: initial cold outreach where you have no relationship to draw on, mass communications where volume matters more than response rate, or event invitations and announcements where the message is the same for everyone by design.
The problem is that founders often use templates for contexts where personalization would make a significant difference, because templates are faster. The speed is real but so is the cost. A template sent to someone who knows you well and who you have a good relationship with says, implicitly, that you did not bother to think about them specifically. That signal often overrides whatever the template says explicitly.
The Edit Rate Finding
One of the more interesting data points from the cohort was the edit rate on AI drafts. Users edited approximately 70 percent of the drafts they used before sending, and in follow-up conversations, they reported that the editing process itself was valuable. Reading the draft, recognizing what was right and what was off, and adjusting it to sound exactly like them produced messages they felt more ownership over than template-filled messages.
This is a somewhat counterintuitive finding. A feature that requires editing to work well should, by simple time math, produce more friction than a template that requires only filling in a few fields. But founders did not report it that way. The editing felt generative rather than laborious, because the draft gave them something specific to react to. Editing a draft is a different cognitive task from writing from scratch, and apparently a more tractable one for most people.
The Context Quality Problem
The single most important variable in draft quality was the richness of the relationship record. Users who logged notes immediately after conversations, who updated their contact records with relevant professional changes, and who used Vinly's note fields consistently got drafts that they described as "already close to what I would have written myself." Users who had sparse records got drafts that were more generic and required more work to make sound personal.
This is not a surprise, but it is worth stating directly: AI-assisted outreach is only as good as the input. The relationship intelligence has to come from somewhere. If you have been logging context over time, the system can produce something genuinely useful. If you have not, it cannot, and a template will produce comparable results at lower effort.
What This Means Practically
The practical takeaway is that the choice between templates and AI drafts is not as binary as it might seem. The right tool depends on the context, the relationship, and how much history you have logged about the contact. For your most important relationships, with rich history in the system, AI drafts with editing are worth the extra few minutes. For bulk outreach or contacts with thin records, templates are a defensible choice.
The broader takeaway is that the system you use to record relationship context is ultimately more important than the drafting mechanism. Both templates and AI drafts improve when you have better inputs. Investing in keeping your relationship records current is the most impactful thing you can do to improve the quality of your outreach, regardless of which approach you use.