How to set up follow-up reminders for your AI agents

If more of your work happens inside agent sessions, your follow-ups start there too. You are in Claude, Codex, OpenCode, or some terminal agent. A client thing comes up. You think, "I need to remember this on Friday." Then you move on and trust your brain, which is usually where the whole plan falls apart.

I wanted a tiny escape hatch for that. Not a new task manager. Not another dashboard. Just: tell the agent to remind me later, get an email when later arrives.

Prefer to see the exact flow? Watch the YouTube tutorial here.

What are AI agent follow-up reminders?

They are plain email reminders created from the agent session you are already in. If the agent helped you draft a proposal, debug a client issue, write a handoff, or plan a next step, it can also help you remember when to come back to it.

That is the whole trick. The reminder gets captured where the work appeared instead of asking you to switch tools and hope you still care enough to write it down.

Why do agent follow-up reminders matter?

Follow-up is where a lot of service work leaks. Not because anyone is careless, but because the next step often shows up in the middle of something else. You are fixing a bug, reviewing copy, or writing the email, and suddenly there is a future task attached to it.

If that future task starts inside an agent session, I want the agent to catch it before I leave. Otherwise it becomes one more loose thread in a notebook, Slack DM, or browser tab I swear I will revisit.

What does the follow-up reminder skill do?

The follow-up reminder skill lets an agent schedule a one-time email for later. That is it. It is small on purpose.

Use it for reminders like:

• Remind me tomorrow morning to follow up on this proposal.
• Email me in two hours to check whether the deploy worked.
• Remind me Friday to check this client handoff.
• Nudge me next week to revisit this estimate.

Which agents can use this setup?

Use this with agents that can run commands or install skills. OpenCode is the one I use in the video. Claude, Codex, and other terminal-style agent tools can use the same basic pattern if they can follow the setup prompt.

The screens will look different, but the flow is not complicated: copy the prompt, paste it into the agent, answer the setup questions, then test it before you trust it.

How do you set up follow-up reminders for an AI agent?

Start from the Pipa tools page and let the agent do the boring parts.

1. Go to usepipa.com/tools.
2. Find the follow-up reminders tool.
3. Copy the setup prompt.
4. Paste the setup prompt into Claude, Codex, OpenCode, or your agent tool.
5. Tell the agent which email should receive reminders.
6. Check your inbox for the verification code.
7. Paste the verification code back into the agent session.
8. Let the agent finish the one-time setup.
9. Send a test reminder for a few minutes in the future.
10. Confirm the reminder email arrives.

Step 1. Copy the setup prompt from the Pipa tools page

Open usepipa.com/tools and choose the follow-up reminders tool. Copy the prompt from that page.

You should not have to wire everything up by hand. The prompt gives the agent the setup instructions so it can walk through the install from the session.

Step 2. Paste the setup prompt into your agent

Paste the prompt into the agent you want to use. In the video, I do this in OpenCode. The same idea applies to Claude, Codex, or another agent that can run the setup.

Once the prompt is in, the agent will ask for what it needs next.

Step 3. Add your reminder email

The agent will ask where reminders should go. Use the inbox you already check for client work. If reminders go somewhere you ignore, the system is technically working and practically useless.

For most studio and service work, that means the same email where client replies, invoices, and project loose ends already land.

Step 4. Verify your email

The setup sends a verification code to that inbox. Copy the code from your email and paste it back into the agent session.

This is just to prove you control the address before the agent starts sending reminders there.

Step 5. Let the agent complete the one-time setup

After verification, the agent finishes the setup and stores what it needs for future reminders. You should not have to repeat the full email verification every time.

Future reminders should feel much lighter: ask for the reminder, choose the timing, move on.

Step 6. Test a reminder

Before you use it for real client work, test it with something low stakes. In the video, I schedule a reminder for two minutes later so I can check the inbox right away.

For example, ask:

Send me a reminder in 2 minutes to check that this worked.

Then check your inbox. If the email arrives, you are done.

When should you use agent reminders?

Use them for the weird middle layer of work: too small for the project plan, too important to leave in your head.

• Following up with a client after sending a proposal
• Checking whether a deploy or automation worked
• Coming back to a bug, PR, or implementation note later
• Remembering to nudge someone after a handoff
• Revisiting an idea when you have time to act on it

What is the full workflow?

The short version:

1. Copy the follow-up reminder prompt from Pipa tools.
2. Paste it into your agent.
3. Add your reminder email.
4. Verify the email.
5. Let the agent finish setup.
6. Send a short test reminder.
7. Check that the email arrives.

Once that works, you can ask in normal language whenever a follow-up pops up during agent work.

FAQ

Can Claude send follow-up reminders?

Yes, if you are using Claude in an environment that can install or run the skill. Copy the setup prompt from the Pipa tools page and paste it into that Claude session.

Can Codex or OpenCode use the follow-up reminder skill?

Yes. OpenCode is shown in the walkthrough, and Codex should work in the same kind of command-running setup. The important part is that the agent can follow the install instructions.

Do I need to set up reminders every time?

No. The email verification is part of the first setup. After that, you should be able to ask for reminders without doing the whole flow again.

Where do the reminders go?

They go to the email address you verify during setup. Use an inbox you actually check.

Is this a task manager?

No. It is a reminder pipe. Use it for one-off nudges that come up while you are working with an agent. Keep your CRM or project management tool for the bigger stuff.

What should I test first?

Ask for a reminder two minutes from now. It is boring, which is exactly what you want for the first test.

Set up the reminder skill

If you want this in your own agent setup, copy the prompt from usepipa.com/tools and paste it into Claude, Codex, OpenCode, or whatever terminal-style agent you use. Run the two-minute test first. Then use it the next time a follow-up shows up in the middle of the work.