# Get your morning briefing before you sit down

> Every morning starts the same: check the calendar, check email, dig up context on who you're meeting. This agent does it on its own, before you sit down.

**For:** Founders and account leads who want calendar, email, and meeting prep waiting for them each morning  
**Format:** pdf  
**Video:** https://youtu.be/bvwF7-PKlLE

## What you'll be able to do

- Build a scheduled Superhuman Go agent that runs on its own every weekday morning at a set time, no manual trigger needed.
- Give it standing context about you once, your role, your company, so it never has to be re-explained on every run.
- Have it check your calendar for external meetings, pull past meeting history from Granola, triage your inbox, and write the result into an existing doc instead of a new one every day.

## What's inside

- The doc ID and table ID format so a scheduled agent writes to one existing page instead of creating a new document every run
- A ready-to-copy prompt for a Monday status-email agent, the same reference pattern used to point an agent at a live Projects table
- A ready-to-copy prompt for an onboarding agent that runs against an Onboarding table and books session invites automatically
- Notes on Always allow versus Always ask, the setting that decides whether a scheduled agent can run unattended
- SDK snippets for adding a skill or MCP server to a custom pack

## Questions

### Do I have to trigger this every day myself?

No. Once it's saved with a scheduled trigger, it runs on its own on the days and time you set, no chat message needed.

### What if I use Outlook or Gmail instead of Superhuman Mail?

Swap the connector. Add the Outlook or Gmail calendar connection in the trigger's connector list instead of Superhuman Mail, everything else in the setup stays the same.

### Why does the agent need Always allow instead of Always ask?

A scheduled agent runs while you're not there to approve anything. Set write permissions to Always allow, or the run pauses and waits for someone to manually approve it.

## Transcript

All right, today I'm going to show you how to create scheduled Superhuman Go agents. If you're unsure of what Superhuman Go is, you can look at some of my other videos for some introductions, but Superhuman Go agents are these powerful AI agents that can collaborate with you to do your work and just get your work done for you. So today we're going to build an agent that has a scheduled routine. It's going to go off every Monday, or every Tuesday, without you needing to do anything.

Let's go with advanced setup first, and this is what you're going to see: this is where you can manually build an agent. First off, what you have is a name. We're going to build a daily briefing agent, which should be applicable to any of you that have a job, you come in and sit down for work first thing in the day and you want to know what the heck should I do today, what emails did I get, and what's on my calendar. So we're going to call it a Daily Briefing.

And next we're going to give it a description. The description here actually does matter. You may think it's just quick for you to visualize and know what it is, but it's actually going to be used later for general Go agents to find and trigger this agent. So I'm going to copy this down, you'll see I put "runs my daily sales briefing, use it when I ask for a morning brief, a daily briefing, what my day looks like, or prep for today's meetings."

So the description, if I am inside of this new task and I generally just say "run my daily briefing," Go has to figure out should I use an agent for this, what should I actually do, and it's going to determine if any of the agents that you own here are relevant to what you're asking it to do. That's what the description does. The description will help Go determine which agent it should actually use. If you're in daily briefing you can just ask and it's already going to know, but of course if you're using your new task it needs that. So let's go back to our daily briefing and keep moving on.

We have a good quality description here already. Now we need the instructions. The instructions are the larger prompt that is actually going to tell your agent what to do every time that it executes. So I have a larger prompt saved here, and let's paste it in and talk about the different aspects of it.

So first off, I'm giving it context about me. Every time an agent runs, it runs fresh. It doesn't actually have a memory like we think of human memory, and so I'm going to give it some information it needs every single time. Luckily I don't have to ever retype this, I type it once and it's saved forever. Lets them know that I'm the founder of SimplaDocs and we do enterprise Coda or Superhuman Docs slash Go implementations, custom workflow, and more.

Okay, then I'm going to start telling it what to do. Say Go, basically, you know, look at today's calendar for me, do some external meeting prep. I'm saying specifically if you find meeting invites that have external invitations on it, people not from my company, this is what I want you to do. I'm also going to use our meeting recorder, Granola, to search and determine if we have met with this individual company in the past. So I'm going to be doing a lot of things: searching the web, looking at my calendar, looking at past meeting invites. It's going to do a lot of actions.

Last thing I'm going to have it do is an email triage, understand what my email is. It's going to build a suggested task list. And then lastly, I think this part is really important: it's going to do all this research, it's going to look at all these things, and now we have to determine what's the actual output, what do you want to see in the morning. There's a lot of different options. What I have right now is, I'm saying a new page in a Superhuman Docs document. I don't want to create a new Superhuman Docs document every single time, so I've actually given it an ID to an existing document, and then I told it what specifically I want in it.

So let's go ahead, now that we have our instructions laid out and we're using IDs to give it an ID, I'm going to add a trigger, and this is where that scheduled trigger comes in. If you go into docs, you'll have different trigger options here. These are always growing at the time of creating this, this is what is available, but I'm going to say repeat, actually, on weekdays at, let's do it at 8am. Let's do it a little earlier in case I start work, I want it to be ready for me by the time I step into work. I'm going to add this. There, now we're actually not ready to run this yet, we have to connect the appropriate connectors.

It will not be able to use Granola. You'll see here I said "use my meeting recording Granola," it's not going to just be able to use it just because I asked it to. I have to authenticate and tell it to use that. So first off, I know it needs to use Superhuman Docs, right, I know that it needs to input information to a document. You'll notice when I first turn this on it says write is always ask. If I want this to trigger automatically every day, I can't give it manual approval, so you'll see it gives you a warning here as well: some tools are set to ask before running and it will pause until someone manually approves.

I don't want that, so I'm going to allow it to work by itself. I'm also going to allow it to search my mail and my calendar. This is if you do use Superhuman Mail, then you can turn that on. If you don't, let's say you use Outlook as a calendar, right, you can go ahead and you can search for Outlook Calendar and add those. So if you use a different calendar, you're going to add those connections here. But the connection that I want to use is our Granola connection. I'm going to add it, see that it is added and connected, it is set to always allow, and I might need to go to its settings to re-authenticate. I can test that if I need to.

But now I've got it connected to my mail, to my calendar, to docs, and to Granola. If you want to test it, you can press run this agent, but as it is, if I press save, this agent will run every weekday at 7:30am and provide that briefing to me in a new document. So let's press run agent and see what happens.

All right, so you're going to see here that it actually did create the daily briefing. We've got a lot of information here available for me as we go. So that is just a quick look at scheduled Go agents. I'm going to have a lot of other videos on how to use Go agents side by side with docs, by themselves, with proactive triggers as well. So take a look at my channel for those other videos, and let me know if you have any questions in the comments below. Adios.

---
Source: https://simpladocs.com/resources/scheduled-superhuman-go-daily-briefing