Why We Built Earmark: We Were Tired of Working Two Shifts

We met the way a lot of product and engineering people do - inside organized chaos. Sprints with rough edges, roadmaps with gaps, and a team that still needed something real to ship by Friday. We knew the rhythm too well. Days were for meetings where good decisions actually got made. Nights were for turning those decisions into the things a team can act on - PRD changes, Jira tickets, status reporting, and a clear update for stakeholders. It was a second shift no one officially acknowledged. You leave a great conversation and then… you start over at a blank page.

One Thursday, after another “quick sync” turned into a 90-minute rewrite, we Slacked each other the line we’d repeated for years: there has to be a better way to close this loop. We’d tried the meeting bots. They showed up as guests, the room got quiet, and the result was a long summary email nobody read. The problem wasn’t collecting more information. The problem was converting alignment into shippable artifacts before the context evaporated.

That’s when we started measuring our own gap - the space between “thanks everyone” and the moment the team has the actual output in hand. We called it Time-to-Shippable Artifact. Not time to transcript. Not time to a tidy summary. Time to the PRD change in our house style. Time to tickets with owners and acceptance criteria. Time to a stakeholder note that makes people nod and move. At most companies, that number lives in hours or days or weeks. Sometimes it lives in “we revisited that decision three times and still didn’t ship.” We asked a simpler question: what would it take to make that number minutes?

We drew a line in the sand and built Earmark as the anti-notetaker. No guest bots - ever. Capture runs on your device, quietly. And most importantly: finished artifacts before you hang up. The point isn’t more words about the meeting; it’s the work you agreed to in the meeting, finished enough to ship while the story is still warm.

Those constraints shaped everything. We bias toward deliverables over descriptions - summaries can be helpful for context, but the product is the PRD delta, the ticket with acceptance criteria, the update people will actually read. We use Product Management-grade structure - problem, rationale, risks, decisions, next steps - so engineering isn’t translating prose into work. And we treat privacy as a product choice: zero-storage by default, opt-in retention when a team asks, and transparency about how data moves.

What does that feel like in real life? Before: the meeting ends, a transcript and AI summary arrive, and someone (often you) spends 60–90 minutes turning it into real artifacts - usually later that night. After: the meeting ends and the PRD delta already exists in your voice; tickets are in Jira or Linear with the right labels, owners, and acceptance criteria; a concise stakeholder update is ready to send. You make a few surgical edits and move on. The loop closes while the story is still fresh. The calendar stops writing your evenings.

You don’t need Earmark to start changing this. Pick one recurring meeting - the one that creates the most follow-up work. Decide what “done” actually means. Is it a PRD update? A set of tickets? A brief for stakeholders? Write down the structure once. Then make a small rule: the meeting doesn’t end until a draft artifact exists where it lives - Jira, Linear, Confluence, Notion, Slack. Track your Time-to-Shippable Artifact for two weeks. You’ll see exactly where the friction hides - where context fades, where decisions get re-litigated, where evenings disappear.

If you want help, Earmark sits quietly on your device, maps the conversation into your templates, so you can post finished artifacts in the right places. No guests. No theater. Just the pieces your team needs to move. We built it because we were tired of working two shifts, and because the best part of our jobs is helping people ship.

Meetings in → deliverables out. No bots. No after-hours. Help us end the infinite workday.