We'll tell you when to click the button

This essay explores recurring work built on meeting context. For currently available features and setup, see the product guide.

Post 5 of 5 — The Handoff series

Across this series we’ve built up a stack: Meeting recall you can check against the source, shared projects, and a one-click handoff to the agent you already use. There’s one piece left, and it’s the one that ties the whole thing into a loop.

So far, everything we’ve described still requires you to start it. You have to remember to ask for the recap. You have to think to draft the email. You have to decide it’s time to create the tickets. Even with the blank-page problem solved, there’s still a blank-intention problem - you have to remember the work exists before you can ask for it.

What if you didn’t?

The work comes to you

The idea is to let someone define recurring work once: a Friday project update, an end-of-day feedback brief, or a proposed ticket set after a standup. A useful implementation would gather the relevant context, prepare the draft, and show enough evidence for the person to review it before handing it to another tool.

This is the difference between a tool you have to operate and a tool that works for you.

This is the difference between a tool you have to operate and a tool that works for you. We do the gathering, the loops, the first pass - the things you run weekly, the report you always owe someone, the tickets you always forget. Then we hand you the finished thing and let you decide where it goes.

The way we think about it: we’ll tell you when to click the button.

Why we still own the loop

You might wonder why this lives in Earmark at all. Couldn’t you just open your agent and ask it directly?

Any recurring task should use only the meetings the person is allowed to access. Private and temporary conversations should remain outside that history.

And because Earmark holds the context, these loops get smarter as your memory grows. The same standing prompt produces a better answer next month than it did today, because there’s more in the record.

Pull the thread on what this becomes

Now stack it with everything else in the series:

• A PM commits to talking to 25 customers about a new feature. They stand up a Project, run the calls over two weeks, then ask Earmark for a clean research report across all of it - and hand it off to their agent to kick off the next round of work, pulling in whatever other tools they need.

• A CS team adds one question to every customer call. Their PM, who joined none of those calls, gets an end-of-day chat with exactly the feedback they were listening for - then pushes it to Slack, or to a doc, or to a backlog.

• You’re heads-down in your agent all day. An email comes in. “Remind me to raise this at the next eng standup.” A week later, it’s on your agenda - and you never left the tool you were already in.

Each of these is the same machine: memory captures the context, loops do the work on a cadence, the handoff routes it to the agent you trust. None of it asks you to switch tools, connect your stack, or wait on IT.

Where this leaves us

The goal is a useful first pass with a clear review step and a chosen destination.

You use a couple of different agents on any given day and a handful of meeting platforms in any given week. Everyone else is trying to lock you into one of each. We’re just going to meet you where you are, do the work, and let you choose where it lands.

That’s the whole pitch. Your meeting is the prompt. Your agent does the work. We make the two talk to each other - and increasingly, we start the conversation for you.

Want early access? waitlist@tryearmark.com.