ChatGPT Can Write Your Emails. Should It?

A client forwarded me an email last month that one of their competitors had sent to a prospect. It opened with “In today’s rapidly evolving landscape…” and included the phrase “leverage synergies” in the second paragraph. We both knew immediately it was AI-generated. Not because we’re particularly clever — because it sounded like every other AI-generated email flooding inboxes right now.

The irony is that the sender probably thought they were saving time. And they were. But they were also broadcasting to every recipient that they didn’t care enough to write a real email.

This is the tension at the heart of AI in marketing right now. The tools are genuinely useful. They’re also making a lot of business communication worse.

Where AI earns its keep

I use AI tools every day. I’m not going to pretend otherwise, and I’d be suspicious of anyone in marketing who claims they don’t. The question isn’t whether to use AI — it’s where.

First drafts are where AI shines brightest. Staring at a blank page is the slowest part of writing. Having a tool generate a rough structure, a first pass at an email sequence, or a framework for a blog post eliminates that cold-start problem. I can take a mediocre AI draft and turn it into something good in a quarter of the time it would take me to write from scratch.

Research and synthesis is another strong suit. Need to understand a new industry before a client meeting? AI can digest ten articles in minutes and give you a useful summary. Need to pull together data points for a report? Same thing. The output isn’t publishable as-is, but it’s a solid foundation.

Repurposing existing content is genuinely transformative. Take a 2,000-word blog post and need a LinkedIn summary, an email teaser, and three social posts? AI handles that well because the thinking has already been done — you’re just reformatting.

Brainstorming is underrated. I regularly use AI to generate twenty subject lines, ten headline options, or a dozen angles on a topic. Most are average. Two or three are genuinely good starting points. That’s a better hit rate than staring at a whiteboard.

Where it quietly kills trust

The problems start when AI writes the final version of anything customer-facing without meaningful human editing.

Emails are the biggest risk area. Your email list is people who have given you permission to show up in their inbox. That’s a privilege. When you fill that inbox with content that reads like it was generated by a machine — generic, slightly too polished, devoid of personality — you’re spending down the trust that got you there.

I can usually spot AI copy within the first sentence. Not because I’m looking for it, but because it has a particular cadence: confident, comprehensive, slightly hollow. It covers all the bases without saying anything specific. It uses “leverage” and “navigate” and “landscape” without irony. It’s technically correct and emotionally empty.

Your customers can spot it too. Maybe not consciously. But they feel it — that subtle sense that nobody real is on the other end of this communication. And when everything in their inbox feels that way, they stop opening emails.

The uncanny valley of business writing

There’s a concept in robotics called the uncanny valley — the point where something looks almost human but not quite, and it becomes deeply unsettling. AI-generated marketing copy lives in its own uncanny valley. It’s almost like a real person wrote it. Almost.

The giveaways are small but consistent. AI tends to be relentlessly positive — every challenge is an opportunity, every problem has a solution, every paragraph ends on an upbeat note. Real business writing has edges. It admits things are hard. It has opinions. It occasionally makes jokes that don’t land perfectly.

AI also struggles with specificity. It will tell you that “many businesses face this challenge” when a human would say “I worked with a construction company in Auckland last year that had exactly this problem.” The difference between generic and specific is the difference between content people scroll past and content that makes them stop.

The practical framework

Here’s how I think about it with my clients:

Use AI for structure and speed. Outlines, first drafts, research summaries, content repurposing, brainstorming. Let it handle the mechanical parts of content creation so you can focus on the parts that require judgment and personality.

Write the human parts yourself. Opening lines, stories, opinions, anything that requires your specific voice or experience. If a sentence could have been written by anyone, AI is fine. If it needs to sound like you — your company, your perspective, your relationship with the reader — write it yourself.

Read everything out loud before sending. This is the simplest quality test. If it sounds like a press release when you read it aloud, it needs more human editing. If it sounds like something you’d actually say to a customer over coffee, it’s probably ready.

Never send a first draft. This applies to human writing too, but it’s especially important with AI. The first draft is a starting point, not a finished product. Edit it. Add your examples. Remove the corporate filler. Make it shorter.

The real risk isn’t bad copy

The deeper risk of over-relying on AI for customer communication isn’t that you’ll produce bad content. AI content is rarely bad — it’s mediocre. And mediocre is harder to fix than bad, because bad content gets caught and rewritten. Mediocre content gets published because it’s “fine” and nobody has time to make it better.

Over time, mediocre content erodes your brand in ways that are hard to measure. Your emails get lower open rates. Your blog gets less engagement. Your social posts stop generating conversations. Nothing dramatic happens — you just slowly become background noise.

The businesses that will stand out over the next few years aren’t the ones that use AI the most or the least. They’re the ones that use it deliberately — as a tool that makes their human communication faster and better, not as a replacement for having something real to say.

That’s the question worth sitting with: not whether ChatGPT can write your emails, but whether your customers can tell when it did.