Build in public
Rewriting the website around speed to lead and reactivating consented lists
Today was mostly about positioning, not code. The voice agent now calls people who have already consented, so the way I describe Dialessa had to catch up with what it actually does.
What I built
The website copy is now built around 2 ideas: speed to lead and lead reactivation. Speed to lead means calling someone soon after they ask to be contacted. Reactivation means going back to people already on a business's list who agreed to hear from them but have gone quiet. Both fit a voice agent that only calls consented contacts. The old copy didn't.
The updated website went live repeatedly through the day as I tightened the wording. I also built and rolled out new releases of the API, the web app and a background worker.
The day's numbers:
- working sessions with coding agents: 1
- prompts typed: 13
- commits: 0
- deploys: 10
- hours of active work: 2
- lines of code changed: 0
10 deploys with 0 commits and 0 lines of code changed looks strange. Today was rollouts and copy, not new product code.
What was hard
Changing the positioning doesn't stop at the headline. Once the website says "consented lists", everything else on the site has to agree with it. The integrations list was the clearest gap. It was written with cold lists in mind, so I've marked it for revision. I haven't rewritten it yet.
The bigger worry is that copy can get ahead of the product. When the story changes, it's easy to describe the product you plan to have instead of the product you have.
The feature audit
That's why I started a feature audit. It's a plain check of what the product promises against what is actually built. I go through each claim on the website and confirm there's a working feature behind it. Where there isn't, the claim gets fixed or the gap gets written down.
It isn't finished. Starting it on the same day as the copy rewrite felt right, because the rewrite is when new promises slip in.
Making the solo plan sellable
I also started work on making the solo plan properly sellable. Right now it exists, but it isn't in a state I'd be comfortable pointing someone to and saying "sign up for this". This work has only just begun.
What I learned
- Positioning changes ripple out. A new headline is easy. Making every page agree with it is the real work.
- Run an audit alongside a copy rewrite, not afterwards.
- A short session with few prompts can still move a lot if it's focused on shipping what's ready.
It was a short day, about 2 hours, but by the end the website described the product more honestly than it did that morning.
Written from that day's build log: the work sessions, commits and deploys recorded for it.