Bring Me the Work You Hate Doing: A Better Way to Find AI Use Cases

You don’t need to start with an AI tool. Start with the work that frustrates you, slows you down, or keeps stealing time from the work that matters more.

When people ask me where they should start with AI, I usually don’t want to start with AI.

I want to start with the work.

More specifically:

Bring me the work you hate doing.

The thing you keep putting off.

The thing that takes too long.

The thing you repeat every week.

The thing you know needs to get done, but every time it shows up on your list, you think:

Not this again.

That’s where I want to look first.

Because one of the easiest ways to make AI feel overwhelming is to start with the technology.

What tools should I use?

Do I need an agent?

Should I automate everything?

What’s the best AI platform?

What’s the newest thing?

Those questions can come later.

First, show me the work.

Start with friction

Every business has friction.

Sometimes it’s obvious.

You spend hours answering the same kinds of emails.

You copy information from one system into another.

You rewrite the same explanation for every new customer.

You keep forgetting to follow up with leads.

You finish a meeting and then still have to create the notes, action items, emails, and next steps.

Sometimes the friction is less obvious.

It’s five minutes here.

Ten minutes there.

A task you do twenty times a week.

A decision you keep making from scratch.

A process that only works because one person remembers how to do it.

A document you rebuild every single time instead of starting from something reusable.

None of those things may feel dramatic on their own.

But together, they can eat a huge part of your day.

That’s where AI can get interesting.

Not because AI is exciting.

Because getting that time back is.

Ask better questions about the work

Instead of asking:

“What can AI do for my business?”

Try asking:

What takes too long?

What do I repeat?

What do I keep doing manually?

What do I avoid because I hate doing it?

What falls through the cracks?

What depends too much on me remembering something?

What do customers keep asking me?

What do I keep explaining to my team?

What do I wish I had more time for?

Those questions tell you much more than a list of AI features ever will.

Because now you’re not looking for somewhere to force AI.

You’re looking for a real problem.

Repetitive doesn’t always mean simple

Here’s where it gets interesting.

Some repetitive work is completely mechanical.

Move this information here.

Rename this file.

Send this notification.

Create this task.

That may not need AI at all.

A normal automation might handle it perfectly.

Other repetitive work requires a little interpretation.

Read this message and figure out what the person is asking.

Summarize these notes.

Turn this meeting transcript into action items.

Draft a response based on the customer’s situation.

Pull the important information out of this document.

That’s where AI may be useful.

And then there are tasks that involve judgment, relationships, money, legal decisions, health, safety, or something else where getting it wrong matters.

Those need more care.

The point isn’t to ask:

“Can AI do this?”

The better question is:

“What part of this could AI help with?”

Sometimes that’s the whole task.

Sometimes it’s only the first 30 percent.

Sometimes AI drafts and a person decides.

Sometimes AI organizes the information and a person handles the conversation.

Sometimes AI shouldn’t touch it at all.

That’s still a good answer.

Don’t automate a mess

This one matters.

If a process is confusing, inconsistent, or broken, adding AI doesn’t magically make it better.

It can just make the confusion happen faster.

Before you automate something, understand what’s actually happening.

What starts the process?

What information comes in?

What happens next?

Where does somebody make a decision?

Where does the process get stuck?

Where do mistakes happen?

Who needs to approve something?

What should the final result be?

You don’t need a giant flowchart for every little task.

But you do need enough understanding to know what you’re improving.

Otherwise, it’s very easy to build an impressive system around a bad process.

And now you’ve got a faster bad process.

Congratulations. 😂

Not exactly the goal.

Go smaller than you think

A lot of people look at their work at too high a level.

They say:

“I need AI to handle my marketing.”

That’s huge.

Marketing might include:

Research.

Content ideas.

Writing.

Email.

Social media.

Scheduling.

Lead follow-up.

Offers.

Campaign planning.

Customer research.

Analytics.

Design.

Reporting.

That isn’t one job.

It’s a pile of smaller jobs.

So go smaller.

Maybe the actual problem is:

Every Friday, I need to look at everything that happened this week and figure out what to post next week.

Okay.

That’s a much better starting point.

Or:

After every sales call, I spend twenty minutes writing the same kind of follow-up email.

Good.

Now we have something.

Or:

Every new customer asks the same eight questions, and I answer them one by one.

There it is.

Small tasks are easier to understand.

They’re easier to improve.

They’re easier to test.

And if something goes wrong, the damage is smaller too.

Find the work that steals your best energy

This is the part I care about most.

Not every annoying task needs to disappear.

But I think it’s worth noticing when the work around your actual work starts taking over.

You might be an incredible baker.

But somehow you’re spending your evenings figuring out email marketing.

You might be an amazing coach.

But half your week disappears into scheduling, notes, follow-up, reminders, and admin.

You might run a great local business.

But now you’re also expected to be a content creator, copywriter, marketer, CRM manager, review manager, website person, and technology expert.

At some point, the support work becomes the job.

That’s backwards.

Technology should help carry some of that weight.

Not so you can cram more work into the empty space.

So you can have some space back.

The goal isn’t to replace the human part

There are parts of work I don’t want AI to erase.

Judgment.

Taste.

Relationships.

Trust.

Care.

Creativity.

Craft.

Experience.

Knowing when something feels wrong even if you can’t immediately explain why.

Understanding a customer because you’ve worked with people like them for twenty years.

Knowing when somebody needs reassurance instead of another automated message.

That’s valuable.

That’s human.

If AI can take care of some of the repetitive weight around that work, great.

But the goal isn’t to automate away the reason people wanted to work with you in the first place.

The goal is to protect more of it.

Look for patterns

Once you start paying attention, you’ll see patterns everywhere.

Maybe you answer the same kinds of questions repeatedly.

Maybe every project starts with the same research.

Maybe every customer needs the same information, just slightly customized.

Maybe every Monday you gather information from five different places before you can make a decision.

Maybe every meeting creates the same chain of follow-up tasks.

Maybe every proposal starts from the same structure.

Maybe every month you rebuild the same report.

Those patterns are useful.

Repeated work gives you something to study.

What stays the same?

What changes?

What information is needed every time?

What decisions require a person?

What could be prepared automatically?

What could be drafted?

What could be organized?

What could be routed to the right person?

That’s how an AI use case starts becoming clear.

You don’t need the fanciest solution

Sometimes the answer is a saved prompt.

Sometimes it’s a reusable project with your business context.

Sometimes it’s a template.

Sometimes it’s an automation.

Sometimes AI needs to sit inside that automation.

And sometimes, yes, an agent may eventually make sense.

But bigger isn’t automatically better.

I’d much rather have a simple system that quietly saves you five hours every week than an impressive AI agent that needs constant babysitting.

The smallest solution that actually solves the problem is usually a very good place to start.

You can always build more later.

Start a “work I hate doing” list

Here’s something you can do today.

For one week, pay attention to the moments when you think:

I hate doing this.

Why am I doing this again?

There has to be a better way.

Why does this take so long?

I already answered this.

I’m going to forget this.

I wish someone else could handle this.

Write those moments down.

Don’t worry yet about whether AI can solve them.

Just collect the friction.

At the end of the week, look at the list.

Circle the things that:

Happen repeatedly.

Take meaningful time.

Follow a pattern.

Use information you already have.

Don’t require your full personal attention every single time.

Those are excellent places to start exploring.

Then take one.

Not ten.

One.

Understand how it works now.

Figure out which part causes the most friction.

Decide what you’d like to improve.

Then see whether AI can help.

AI should give something back

For me, this is the bigger point.

AI shouldn’t just help us produce more.

More emails.