You sent it. They never replied. You forgot. Ten minutes now saves the deal, the favor, or the offer that quietly died because nobody followed up.
| Follow up window | 10 days |
|---|---|
| Weekly review time | About 5 minutes |
| Recommended schedule | Monday mornings |
| Gmail send approval | Required by default before sending |
1. Connect your inbox as a tool, not a chat window
Most AI assistants can now connect to your email through a secure connector, sometimes called MCP, which is just a plain way of giving the AI a specific job it can do, like read my sent folder. You are not pasting emails in one by one, you are giving it standing access to check on its own. In Claude or ChatGPT, look for the option to connect Gmail or Outlook, and connect it as read only if you can. This takes about five minutes to set up once.
2. Ask it to find every unanswered thread
Give it one clear instruction: Look at every email I sent in the last 10 days where I was the last person to write, and nobody has replied. List sender, subject, and how many days it has been silent. Be specific about the window, ten days works well for most work emails. The AI hands you a short list instead of you scrolling your sent folder from memory.
3. Have it draft the nudge, not send it
For each name on the list, ask it to draft a two line follow up in your voice, polite, not pushy, no guilt. Give it two or three follow ups you have actually sent before so it matches how you sound. You read, tweak if needed, and send yourself, nothing goes out without you seeing it first.
4. Put it on a Monday schedule
Set this to run once a week, Monday morning before your day starts, either through your assistant’s built in scheduling or a simple reminder that prompts you to run it. The point is you stop relying on memory to catch these, the check happens automatically. Five minutes on a Monday now covers ten days of sent mail.
5. Review and send in five minutes
Read the list, drop anyone you already know is fine, send the rest. Over a few weeks you will notice which threads tend to go quiet, and you can decide whether to follow up sooner next time. This turns I should really follow up on that into an actual Monday habit.
A worked example
A reader sent five proposals to potential clients over two weeks and lost track of who had replied.
Before
Sent Aug 12 to a client about a proposal, sent Aug 15 to another about pricing, sent Aug 18 with a contract draft. No system for tracking any of it, just a vague mental note to check back sometime, and two of the three forgotten entirely.
After
Sent Aug 12 to Priya Sharma, proposal, no reply, 15 days silent. Sent Aug 15 to Ahmed Al Farsi, pricing question, no reply, 12 days silent. Sent Aug 18 to Ravi Iyer, contract draft, no reply, 9 days silent. Draft nudge for Priya: Hi Priya, following up on the proposal I sent last week, wanted to check if you had a chance to look it over.
Three ways to bend it
For recruiters and hiring managers
After an interview, ask the AI to flag if the recruiter or hiring manager has not replied within the timeframe they gave you. It drafts a short, professional check in instead of you refreshing your inbox out of anxiety.
For managers waiting on their team
Use the same check on requests you sent to colleagues, an approval, a document, a sign off. It surfaces who is quietly sitting on something you asked for last week.
For freelancers and consultants
Point it at client threads specifically, invoices sent, drafts shared for feedback, scope questions. A weekly list keeps you from chasing payment or approval three weeks too late.
When it goes wrong
It flags emails that were actually answered on a call or in a different thread
Treat the list as a draft to glance over, not gospel. A quick scan before sending nudges catches the false positives in seconds.
I do not want AI reading everything in my inbox
Scope its access to just the sent folder or a specific label instead of the full inbox, most connectors let you restrict this.
The follow up drafts sound stiff or robotic
Feed it two or three real follow ups you have written before so it learns your actual tone, not a generic polite template.
When not to use this
If you rarely send emails that need a reply, or you already flag and track everything by hand, this is extra machinery for a problem you do not have. And for anything sensitive, legal threads, HR matters, personal messages, do not hand over broad inbox access just to save five minutes, do that follow up yourself.
What you can now do
You stop losing deals, favors, and offers to plain forgetting, and following up becomes a five minute Monday habit instead of a guilty memory.
Part of Making Life Easier With AI, a weekly walk-through of one small thing AI can take off your plate.
Take this further
Act as a blunt hiring manager who has read ten thousand resumes, not a career coach. I will paste my full resume and the job description I want. Rewrite the whole resume for that role, section by section, in this order: summary, experience, skills, education. Rules: every experience line leads with impact, not duty. Use bracketed placeholders like [8 percent] for any number I did not give you, and list at the end every placeholder I need to replace with a real figure. Keep it to one page of text. Plain formatting only, no tables or columns, so screening software can parse it. After the rewrite, tell me the three weakest claims that need evidence before I send this anywhere. My resume: [paste resume]. The role: [paste job description].
Takeaways
- Connect Gmail or Outlook to Claude or ChatGPT with read only access scoped to your sent folder before running any follow up check.
- Ask the assistant to list every email you sent in the last 10 days with no reply, showing sender, subject, and days of silence.
- Paste in two or three of your own past follow up emails so the AI’s drafts match your actual tone before you send anything.
- Set the check to run every Monday morning so it becomes a fixed habit instead of something you have to remember.
- Keep legal, HR, and personal email threads out of the connector’s scope entirely.
Questions people actually ask
How do I get AI to check which of my emails never got a reply?
Connect your inbox to Claude or ChatGPT through a secure, read only connector if one is offered, then ask it to scan your sent folder for messages where you wrote last and nobody answered within a set window, such as 10 days. It returns a short list showing sender, subject, and how many days each thread has gone silent.
Is it safe to let AI read my email for something like this?
Scope the connection narrowly, to your sent folder or a single label instead of the whole inbox, and pick read only access whenever the connector offers it. Keep the AI limited to drafting nudges rather than sending them, so every message still passes through you first. Skip this setup entirely for legal, HR, or personal threads.
Can the AI send emails on its own without me checking first?
With Claude’s Gmail connector, sending, replying to, and forwarding a message all require your approval by default before anything goes out, and admins on Team or Enterprise plans can decide whether members can skip that approval step. Treat the approval prompt as a safeguard, not a guarantee, and always glance at a draft before confirming it.
How do I stop the follow up drafts from sounding robotic?
Give the AI two or three follow up emails you have actually sent before so it can copy your real phrasing, tone, and typical length. Without an example to work from, most assistants default to a stiffer, more formal style than how people actually write to colleagues or clients.
What happens if the AI flags an email that was already answered on a call?
These lists are a starting draft, not a verified record, since the AI only sees what happened inside your inbox and has no way to know about a reply that came by phone, text, or a separate thread. A quick scan before sending any nudges catches those false positives in seconds.
Sources
What happens next
Expect connector permissions to keep tightening, with more granular read only and per folder scopes becoming the default rather than an option users have to dig for. Native scheduling inside these assistants is also likely to replace external reminders, so a weekly check like this becomes a built in setting. Availability will stay uneven across regions in the near term, since OpenAI has already excluded the EEA, Switzerland, and the UK from some of its connector rollouts.
