Why I'm NOT Automating These 5 Things (Even Though I Probably Could)

I'm pretty bullish on AI automation. I've automated a bunch of stuff in my business, and it's been great.

But I'm also saying "no" to automating some things, and I think that's just as important to talk about.

Everyone's publishing articles about what you SHOULD automate. Cool. But nobody's talking about what you shouldn't, and honestly, that's the more interesting question.

So here are five things I could automate but won't. Your mileage may vary, but I figured I'd share my reasoning in case it helps you think through your own decisions.

1. Thank You Notes to Big Customers

Could I automate this? Absolutely. Set up a trigger: when deal closes over $X, send personalized thank you email. Easy.

Why I'm not doing it:

Because it would ruin the whole point.

I send handwritten thank-you notes to customers who spend over $5K with us. Not emails. Actual cards, with a pen, in the mail. Takes me about 15 minutes each.

Could I automate an email? Sure. But the entire value is in the fact that it's clearly not automated. The ink smudges a little. My handwriting isn't perfect. There's a real stamp.

Last month, a customer posted their card on LinkedIn. "Can't remember the last time I got a handwritten note from a CEO. This is why we chose them."

That post was seen by like 3,000 people. You know what an automated email doesn't get? Posted on LinkedIn.

Some things are valuable BECAUSE they don't scale. The moment I automate it, it loses the thing that made it special.

The principle: When the human touch is the whole point, keep it human.

2. Firing Clients

Could I automate this? Technically, yes. Autodetect problem clients, trigger a "we're no longer a good fit" email, done.

Why I'm not doing it:

Because I'm not a sociopath?

Look, sometimes you need to part ways with a client. Maybe they're not a good fit, maybe they're abusive to your team, maybe they're just not profitable. It happens.

But having that conversation - even when it's uncomfortable - is important. It's human. It's respectful.

I had to fire a client last year. Guy was treating my team like garbage. I got on a call, explained it wasn't working out, offered to help transition them to someone else. It wasn't fun. But it was right.

Could I have sent an automated email? Sure. Would I be able to look at myself in the mirror? Questionable.

There are conversations you need to have yourself, even when (especially when) they're hard.

The principle: Don't use automation to avoid being a human.

3. Initial Sales Calls with New Prospects

Could I automate this? Partially. AI could do the discovery call, ask qualifying questions, present solutions, maybe even close some deals.

Why I'm not doing it:

Because I learn too much from these conversations.

Here's the thing: when I talk to new prospects, I'm not just selling. I'm learning about their problems, their industry, their objections, what they actually care about vs. what they say they care about.

Last month, three different prospects mentioned the same pain point I hadn't really thought about. That turned into a new feature we built. That feature has now sold to like 15 more customers.

If I'd automated those calls, I would have missed that insight. The AI would have followed the script, answered the questions, maybe closed the deals. But I wouldn't have learned anything.

There's strategic value in being in the trenches sometimes.

Now, do I automate the scheduling of these calls? Hell yes. Do I have AI help me prepare for them? Absolutely. But the actual conversation? That's me.

The principle: Stay close enough to your customers to learn from them.

4. Apologizing for Screw-Ups

Could I automate this? Yeah, and some companies do. Trigger: complaint detected, send apology email, offer discount code.

Why I'm not doing it:

Because automated apologies feel worse than no apology at all.

We screwed up an order last month. Customer ordered blue, we sent red. Simple mistake, annoying for them.

I sent a personal email apologizing, explained what happened (new team member mixed up the labels), overnight shipped the correct item at our expense, and included a handwritten note in the package.

Customer wrote back: "Mistakes happen. Your response tells me everything I need to know about your company. I'll be ordering again."

That wouldn't have happened with an automated "sorry for the inconvenience" email and a 10% discount code.

When you mess up - and you will - own it personally. Don't hide behind automation.

The principle: When emotions are high, humans should show up.

5. Strategic Planning and Decision-Making

Could I automate this? Getting there. AI can analyze data, spot trends, even make recommendations.

Why I'm not doing it:

Because running my business is my job, not just a series of tactical decisions.

I use AI to help with analysis. Show me patterns I might miss. Crunch numbers faster than I could. That's great.

But the actual strategic decisions - where we're going, how we're getting there, what we're betting on - those are mine.

Not because AI couldn't make reasonable suggestions, but because I need to own those calls. When things go well, great. When they don't, I need to learn from that.

Also, honestly? The strategic thinking is the fun part. If I automate that away, what am I even doing here?

There's a difference between AI that helps me think better and AI that thinks for me. I want the first one.

The principle: Don't automate yourself out of doing what you're actually here to do.

The Pattern I'm Seeing

Looking at these five things, there's a common thread:

They're all situations where the human element IS the value, not something that gets in the way of the value.

With customer support emails, the human element (me typing at midnight) was getting in the way. The customer just wants their tracking number fast. Automating that makes it better.

But a thank you note? The human element is the entire point. Automating it makes it worse.

The question isn't "can I automate this?" It's "should I automate this?"

And "should" depends on whether automation makes it better or just faster.

What Other People Are Keeping Human

I asked some other business owners what they refuse to automate. Got some interesting answers:

Coffee shop owner: "I could automate our social media, but I post personally. People follow us for the personality, not generic cafe content."

Consultant: "I write all my own content. AI can draft, but my unique perspective is literally what people hire me for."

Wedding photographer: "I personally respond to every inquiry even though most are tire-kickers. The ones who become clients remember that first conversation."

Therapist: (Obviously everything, but specifically) "Even my booking reminders are personal. The therapeutic relationship starts with the first interaction."

Notice the pattern? They're all keeping human the things that make their business THEIR business, not just A business.

The Counterargument

I can already hear someone saying "you're leaving money on the table" or "that doesn't scale."

Fair points. Let me address them.

"You're leaving money on the table."

Maybe? But I'm building a business I want to run, not maximizing every possible dollar of revenue. If I wanted maximum revenue extraction, I'd automate everything and sell to a PE firm. That's not my goal.

I want a business that's profitable AND that I enjoy running. For me, that means keeping some stuff personal.

Your goals might be different. That's fine. But "more money" isn't automatically the right answer to every question.

"That doesn't scale."

Correct. By design.

I send maybe 20 handwritten notes a month. If I was sending 500, yeah, I'd need to rethink it. But I'm not trying to close 500 $5K+ deals monthly. I'm trying to build relationships with 20 great customers.

Not everything needs to scale. Some things should stay deliberately small and personal.

There's this assumption that scaling is always the goal. Sometimes maintaining the current scale but doing it really well is the goal.

How I Decide What to Keep Human

Here's my incredibly unscientific decision framework:

Ask yourself three questions:

1. Is the human element the product, or is it overhead?

If customers value the human touch specifically, keep it human. If they just want the result and the human element is incidental, automate it.

Example:

• Customers value that I wrote this article → keep human
• Customers just want their tracking number → automate

2. Would automating this make me uncomfortable?

Not "is it unfamiliar" uncomfortable (that goes away). But "does this feel wrong" uncomfortable.

If you'd feel shady automating something, trust that instinct. Even if you could get away with it, don't.

3. Am I automating this to improve the experience, or to avoid doing something hard?

Automating customer emails = improves experience (faster response) Automating difficult conversations = avoiding something hard (wrong reason)

There's no right answer that works for everyone. But these questions help me figure out what works for me.

The Stuff I Thought I'd Keep Human But Didn't

Plot twist: there's also stuff I thought I needed to do personally but turned out I was wrong.

Weekly Newsletter

I thought people subscribed for MY voice and my personal writing. Turns out they mostly subscribed for useful information about our industry.

I still write our newsletter, but I use AI to help with research, outlining, and editing. It's better than when I did it 100% myself because I was usually writing it at 11 PM and phoning it in.

Lesson: Sometimes what you think people value isn't what they actually value.

Social Media Responses

I used to personally respond to every comment and message. Felt important. Then I realized I was basically just saying "thanks!" or "glad it helped!" over and over.

Now AI handles the routine acknowledgments. I personally respond to actual questions or substantial comments. Everyone's happier - customers get faster responses to simple stuff, and I can give real attention to real conversations.

Lesson: Your ego about doing things personally might be getting in the way of actually serving people better.

Monthly Reports

I spent hours creating financial reports for myself. Like, I was the only person reading these. I thought doing them manually kept me "close to the numbers."

Automated the whole thing. Turns out I stay close to the numbers by reading the reports, not by making them. Who knew?

Lesson: Busywork disguised as diligence is still busywork.

The Future Stuff I'm Watching

There are some things I'm keeping human now but might change my mind about later:

Content Creation

Right now I write everything myself (with AI assistance). But AI is getting really good. At some point, maybe it'll make sense to have AI draft more and I just edit and add personality. Not sure when that line is, but I'm watching.

Product Recommendations

We currently have humans recommend products to customers because it feels more personal. But AI recommendations are getting scarily good. And honestly, they might be more helpful because they're not influenced by whatever's top of mind for the human at that moment.

Hiring Decisions

Will never fully automate this, but AI could probably screen resumes better than I do (I'm bad at it). Might automate early stages but keep final decisions human.

The line keeps moving. What seems crazy to automate today might seem obvious in a year.

My Actual Advice

Here's what I'd tell another business owner thinking about this:

Don't automate something just because you can. Ask if you should. Those are different questions.

Don't keep something manual just because you always have. Ask if there's actually value in keeping it human or if you're just being stubborn.

When in doubt, ask your customers. They'll tell you what they actually care about. Often it's not what you think.

It's okay to be selective. You don't have to automate everything OR automate nothing. Pick and choose thoughtfully.

Your answer doesn't have to be everyone else's answer. What's core to your business might be overhead for someone else. That's fine.

The Thing That Made Me Write This

I was talking to a friend who was feeling guilty about not automating more. Like she was falling behind or doing it wrong.

She runs a tutoring business. She could automate scheduling, reminder emails, progress reports, all kinds of stuff. And she does automate some of it.

But she refuses to automate the actual tutoring sessions (obviously) or the personal check-ins with parents. Those are the whole point of her business. And she felt guilty about it because there are AI tutoring platforms out there.

I told her: your parents hire you BECAUSE you're human. The AI platforms are competing on efficiency and scale. You're competing on actually giving a damn about their kid. That's not a weakness. That's your differentiator.

She seemed relieved.

So if you're feeling pressure to automate things you don't want to automate - as long as you're being thoughtful about it and not just resistant to change - it's okay to say no.

Where This Leaves Me

I've automated a ton of stuff. Customer emails, scheduling, data entry, reporting, a bunch of other things. Super happy about all of it.

But I'm also drawing some lines. And I'm comfortable with those lines.

My business works for me. I'm profitable, growing at a reasonable pace, and I actually enjoy what I do. If I automated the five things I listed above, I might make more money. But I'd enjoy it less, and I'd lose the parts that make it mine.

That's a trade-off I'm not willing to make.

You might make different choices. That's cool. Just make them intentionally, not out of guilt or peer pressure or because some guru said you should automate everything.

Your business, your rules.

For a different perspective on what you definitely should keep human, check out "When NOT to Use AI Automation: A Business Owner's Reality Check". And if you want the structured approach to figuring out what to automate, "How to Choose the Right AI Agent for Your Business Needs" is actually pretty helpful even though the title sounds boring.