We used em dashes before it was uncool

The short version

We use AI to help write nearly every reply we send you. A qualified human reads every message before it goes out. We never let these tools train on your data. And anything sensitive is handled by local AI models that aren't connected to the internet at all. That's the whole story. The rest of this post is the why, plus one small confession about punctuation.

First, the punctuation

Long before ChatGPT taught the internet to be suspicious of a little horizontal line, Greg was an em dash guy (thanks to macOS for making it easy to add them!). We run this business almost entirely over email. When nearly everything you say to a client is text on a screen, you look for ways to point at the important part. The em dash did that job nicely. It's a pause with a bit of emphasis built in.

Then large language models showed up, developed a fondness for the exact same mark, and now a stray em dash reads like a confession that a robot wrote your email. We never got the memo that our favorite piece of punctuation had been repossessed. So yes — if one sneaks into a reply, we were probably going to use it anyway.

Yes, we use AI. Here's exactly how.

We think you deserve a straight answer about how the work gets done, so here it is.

In plain terms: AI helps us write faster and more thoroughly, but it doesn't get the final say, and it doesn't get to keep or learn from your information.

Why we do it this way

We're a lean agency. We keep our team small on purpose, and that means we lean on good tools to move fast. If we hand-wrote every reply from scratch, our response times would stretch from minutes to hours, and sometimes days. The only way to keep quick turnaround without AI would be to raise prices, a lot, on everyone.

We don't want to do that. A managed WordPress site with us runs many clients somewhere between ten and twenty dollars a month, and that includes a pile of extras most hosts charge separately for. In 2026, that price is frankly a little ridiculous, in your favor. AI is a big part of how we keep it that way, so consider this our polite apology for the occasional AI-drafted reply or help article.

What this actually looks like for you

When you write in with a problem, we aim to answer in the same shape every time:

• A quick summary of the problem, so you know we understood it.
• The short answer, for when you're busy and just need the fix.
• The details, for when you want to understand what happened and why.

There's a practical reason for that structure. A clear written answer is one you can search and re-read later, which beats leaving us a third voicemail about a question we already answered back in April. We're not dodging the phone. We just think you'll get a better, faster, more complete answer in writing, and you'll still have it next month.

An apology, sort of

Sometimes an AI-assisted reply comes out sounding a little flat, or a little too polished, like it was written by something without a personality. We get it. If a message ever reads like a robot, we're sorry. But we'll also be honest: the answer the AI helps us write is usually more accurate and more thorough than the snarky one-liner Greg would have fired off at eleven at night. You're getting our expertise and judgment, delivered by a tool that doesn't get tired or cranky. On balance, that's a trade we think is worth it, and it's the same reason many of our help articles are AI-drafted and human-checked.

So: em dashes, AI, and all. The people behind your website are still very much people. We're just people with good tools and a policy of telling you the truth about how the work gets done.

— Greg

Em dashes in this post: two. We're showing restraint.

Still have questions?

If you'd like to know more about how we handle your data or your account, send us a note and a real human will get back to you.