Writing follow-up emails is one of the most time-consuming, low-creativity tasks in freelance work. The words matter, but the process of staring at a blank email to a client you haven’t heard from in a week is mostly friction. ChatGPT removes that friction. Here’s exactly how to use it.
What ChatGPT can write for you
The free version of ChatGPT handles all of these competently with the right prompt:
- Proposal follow-up emails (day 3, day 10, final close message)
- Proposal cover emails (the message you send with the actual proposal)
- Invoice reminder emails (friendly reminder, overdue notice, firm final notice)
- Post-project check-ins (“how’s the work holding up?”)
- Cold outreach emails to prospective clients
- Meeting follow-up summaries
The pattern is the same for all of them: give ChatGPT the context, tell it what you want to happen, specify the tone.
The prompt formula
Bad prompt (produces generic output):
“Write a follow-up email to a client about a proposal.”
Better prompt (produces usable output):
“Write a follow-up email to a client named [Name] who received my brand design proposal five days ago and hasn’t responded. The proposal was for a brand identity system — logo, color palette, and type system — priced at $3,200. I want to ask if they have questions and keep the door open without being pushy. Keep it under 4 sentences. Professional but human.”
That prompt tells the model: who it’s to, what the context is, what the goal is, the length constraint, and the tone. The output will be significantly more usable.
Templates to copy and modify
Proposal follow-up — day 5:
Prompt: “Write a 3-sentence follow-up email to [Name] who received my [type of project] proposal 5 days ago. I want to check in, ask if they have questions, and not sound pushy. Professional tone.”
Proposal follow-up — day 10 with reframe:
Prompt: “Write a follow-up email to [Name] who hasn’t responded to my [type of project] proposal. I want to acknowledge they might be busy, mention that I’m happy to adjust the scope if the current version doesn’t fit perfectly, and ask a specific question to make it easy to reply. Keep it under 5 sentences.”
Final close message — day 21:
Prompt: “Write a final follow-up email to a prospect who hasn’t responded to a proposal. I want to give them an easy out, close the loop from my end, and leave the door open for the future. Warm, not guilt-tripping, under 4 sentences.”
Invoice reminder — 3 days past due:
Prompt: “Write a polite invoice reminder email to [Name] for a $[amount] invoice that was due 3 days ago for [project]. Assume it was an oversight. Keep it friendly and under 3 sentences. Include a note to reach out if there are any issues with the invoice.”
Invoice reminder — 10+ days past due:
Prompt: “Write a firmer invoice reminder to [Name] for a $[amount] invoice that is now 12 days overdue. Keep it professional, not aggressive. State clearly that the invoice needs to be paid and ask for an expected payment date.”
Editing the output
The 2-minute edit is what separates a message that sounds like ChatGPT from one that sounds like you.
After generating a draft:
- Replace any generic opener (“I hope this message finds you well”) with something direct
- Add one specific detail — their project name, something you discussed, their company name
- Make sure the ask is one concrete thing, not a vague “let me know”
- Read it aloud. If any sentence sounds like a robot wrote it, rewrite that sentence
Common phrases to delete: “I wanted to reach out,” “I hope you’re doing well,” “as per my last email,” “I look forward to hearing from you” (replace with a specific next step).
Pairing AI drafts with tracking
Writing the email is only half of effective follow-up. Knowing when to send it matters too. If you sent a proposal and your tracking shows the client opened it twice yesterday, that’s the moment to follow up — not day 5 or day 10 by default.
Tools like Waco3 surface proposal open activity so you can time your follow-up to when the client is actively thinking about your work. The email ChatGPT helps you write in 3 minutes lands at exactly the right moment instead of a random one.
What ChatGPT can’t do
It can’t know your voice exactly without examples. If you give it a sample of your past emails, the output will be closer to your style. If you don’t, it will produce competent but generic professional language.
It can’t anticipate specific objections based on your actual conversation. That’s where the editing step matters — you add the context it can’t know.
It also won’t follow up for you automatically. The draft is still a draft. You still make the decision about when to send, what to tweak, and whether the moment is right.
For the draft itself, though — the blank page problem — ChatGPT solves it every time.
Ready to send stronger proposals?
Build, send, and track proposals in one place so follow-up is easier.
Start your free trial →
