Cut email time by 75% with an AI voice bank (5-step system)
The time saving from AI email tools is real. But most people skip the one step that makes it work, and end up with replies that every reader can immediately identify as machine-written. Here is what the system actually looks like when it is built correctly.
One consultant tracked every email she sent over a month using a spreadsheet with start and stop timestamps. The result: 11 hours a week, just on replies. Not proposals, not client calls, just answering the same six questions over and over. What do you charge, can we get on a call, do you work with e-commerce, are you available in March, can you send a case study, is this right for a team of four.
Eleven hours is one and a half working days. For a solo operator or a small team, that gap is the difference between hitting a monthly revenue number and missing it.
The split that actually matters
The first instinct is to automate the emails that feel most important, the warm leads, the press enquiries, the existing client questions. That instinct is backwards.
The right split is this: AI drafts the replies you have written a hundred times before. That covers roughly 60 percent of a typical inbox by volume: pricing questions, availability checks, industry fit questions, requests for a speaker one-sheet, unsubscribe and out-of-office style replies.
The other 40 percent, emails from existing clients, press contacts with a genuine story angle, anyone who has already paid you money, get written by hand. Every time. No exceptions. Those are the relationships where a slightly off tone costs real money, and the risk of an AI tone mismatch is not worth it on the emails that matter most.
️ The 5-step system
This took one afternoon to set up and about 20 minutes a week to maintain after that.
- Pull your last 100 sent emails. Look for repeated patterns. The consultant found 11 distinct question types in hers, from day rate questions to internal conference speaking requests.
- Write one gold-standard reply for each pattern, by hand. Not with AI. This is your voice bank, and it needs to sound like you on a good day, not like a chatbot’s approximation of you.
- Feed those replies into a custom instructions or project file in ChatGPT or Claude. Give it a direct brief: here are 11 examples of how I reply to emails, match this tone exactly, short sentences, no corporate padding, British spelling (or whatever your defaults are).
- For every new email that matches a pattern, paste it in and ask for a draft. Read it, edit it. The consultant typically changes one or two sentences per email and never sends a draft unread.
- Once a month, add any new recurring question to the voice bank. After eight months, the original 11 patterns had grown to 19.
Result: 11 hours a week became roughly 2.5 hours. Not zero, because every draft still gets read and edited before sending. But the typing-from-scratch time disappeared entirely.
⚠️ The problem with AI’s default tone
ChatGPT has a house style. You probably recognise it even if you have never named it: three short paragraphs, an em dash where a full stop would do, the word “delve” doing heavy lifting, and a closing line like “let me know if you’d like me to expand on any of this.”
Last year, the consultant read 40 client emails that all had that exact rhythm. She could tell within one sentence that the sender had not touched their own draft. It did not make her think less of them for using AI. It made her think less of the message, because it had stopped sounding like a person.
This is the part most AI productivity posts skip: using AI badly is now more obvious than not using it at all. A slightly clunky, human reply with a typo reads as more trustworthy than a polished paragraph that sounds like it came from a template. The voice bank is not a nice-to-have step. It is the entire thing that stops your emails reading like everyone else’s.
A real example
A bookkeeping firm founder near Leeds emailed asking about a speaking fee, whether half-day workshops were available, and whether there was experience in financial services. Three questions, all in the voice bank. The consultant pasted the email into her project, got a draft back in under a minute, changed one line because the draft had not referenced a specific event date the client mentioned, and sent it. Total time: four minutes, including reading the original email.
The old process: open a blank reply, type the day rate from memory, second-guess the number, check the calendar tab, spend 15 minutes on something that should take four. The client replied within the hour, booked a call, and became a paying client six weeks later.
Common pitfalls
- Starting with ChatGPT before building the voice bank. You will get something generic and it will take longer to fix than to write from scratch.
- Automating the high-value emails. Existing clients, anyone negotiating price, and press contacts with a non-generic story angle should never go through the automated list.
- Never editing the draft. Every AI draft should be read before sending. One or two edits per email is normal and expected.
- Not maintaining the voice bank. New recurring questions appear every month. Adding them takes 20 minutes and keeps the system accurate.
What to do instead of automating some emails
Some inbox volume does not need a reply at all. It needs a self-service tool. If 40 people a month ask what something would cost for their business, that is not really 40 emails. That is one missing calculator on your website. Removing those questions from your inbox entirely reduces volume before AI is even involved.
Quick answers
- Is it obvious when an email was written by AI?
- Often yes, with default ChatGPT phrasing. Feeding AI your own past replies as examples removes most of that tell, because it copies your rhythm instead of its own defaults.
- How long does setup take?
- About one afternoon for the initial voice bank from your last 100 sent emails, plus 20 minutes a week to add new recurring questions as they come up.
- Which emails should never be automated?
- Anything from an existing paying client, anyone negotiating a contract or price change, and press or partnership enquiries with a specific, non-generic story angle.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
