Everyone is trying to sell AI services. The real opportunity is solving problems nobody wants to deal with.

A few years ago, learning a new digital skill felt like discovering a shortcut.

Learn copywriting. Build a portfolio. Find clients.

Learn SEO. Offer monthly packages. Build a business.

Then AI arrived, and suddenly everyone became an AI expert.

LinkedIn filled with automation specialists. Freelance platforms exploded with ChatGPT consultants. People who discovered prompts last Tuesday started selling AI transformation services by Friday.

And honestly?

I understand the excitement.

AI has created opportunities for freelancers that barely existed a few years ago.

But there’s a problem.

Everyone seems to be chasing the same opportunities.

AI content creation. Chatbot development. Prompt engineering. Social media automation.

These services aren’t necessarily bad. They’re just becoming increasingly difficult to differentiate.

The interesting opportunities are somewhere else.

They’re hiding inside boring business problems.

And that’s exactly where I’d start looking.

1. AI Workflow Auditing for Small Businesses

Most small businesses don’t need another AI tool.

They need someone to explain why the seven tools they’re already paying for aren’t saving them any time.

That’s an opportunity.

An AI workflow auditor reviews how a business operates, identifies repetitive tasks, and recommends practical improvements.

Think customer inquiries, appointment scheduling, reporting, document management, and internal communication.

You don’t need to build revolutionary software.

You need to understand how businesses waste time.

Your deliverable could be a simple audit showing what to automate, what to leave alone, and which existing tools could work better together.

Businesses don’t buy automation because it’s exciting.

They buy back their time.

2. AI Knowledge Base Organization

Here’s a problem nobody seems particularly excited about.

Company information is a mess.

Documents live in Google Drive. Procedures hide inside Slack conversations. Important answers exist exclusively inside someone’s head.

Then businesses introduce AI assistants and wonder why the answers are terrible.

The problem often isn’t the AI.

It’s the information feeding it.

Freelancers who can organize internal documentation, remove outdated information, structure knowledge bases, and prepare content for AI-powered search have a useful service to offer.

It’s not glamorous.

Neither is spending three hours looking for a document someone created in 2022.

That’s precisely why businesses might pay to fix it.

3. AI Quality Control and Human Review

Everyone wants AI to produce more content.

Far fewer people want to check whether that content is actually correct.

That’s a business opportunity.

Companies using AI for customer support, product descriptions, educational materials, and internal documentation still need human oversight.

A freelancer could specialize in reviewing AI-generated outputs for factual errors, consistency, brand voice, and usability.

You could even focus on one industry.

For example, reviewing AI-generated product descriptions for e-commerce stores.

Or checking chatbot responses against a company’s approved documentation.

The positioning matters.

Don’t sell yourself as someone who fixes ChatGPT mistakes.

Sell a quality-control process that helps businesses publish with greater confidence.

4. AI Implementation for Traditional Industries

There’s an entire world outside the technology bubble.

Local contractors. Property managers. Cleaning companies. Small logistics businesses. Independent retailers.

Many have repetitive administrative problems but little interest in becoming AI experts.

They don’t want a complicated technology strategy.

They want fewer missed inquiries, faster estimates, and less paperwork.

A freelancer who understands one traditional industry can develop practical AI-supported workflows around its specific needs.

For example, helping property managers organize maintenance requests and prepare tenant communications for human approval.

The advantage isn’t knowing every AI platform.

It’s understanding one customer’s daily frustrations better than a generalist does.

5. AI Training for Nontechnical Teams

Installing software is easy.

Getting people to use it properly?

That’s another story.

Businesses introducing AI tools need employees who understand what those tools can do, where they fail, and how to use them responsibly.

This creates opportunities for freelancers offering practical training.

Not another generic webinar explaining what ChatGPT is.

Actual workshops built around everyday tasks.

How customer service teams can draft better responses. How administrative employees can summarize documents. How managers can create reports without exposing confidential information.

The more specific the training, the more useful it becomes.

The Real Opportunity Isn’t AI

Here’s what I think freelancers are getting wrong.

They’re choosing niches based on whichever AI tool is trending.

But tools change.

Client problems don’t disappear nearly as quickly.

Businesses will continue needing organized information, reliable processes, trained employees, and better customer experiences.

AI simply gives freelancers new ways to solve those problems.

And you don’t need to become an expert in everything.

Choose one industry.

Find one expensive, frustrating problem.

Learn how AI can help solve it.

Build a repeatable service around that outcome.

That’s a much stronger position than adding AI expert to your LinkedIn headline and hoping someone notices.

The goal isn’t to find a niche with absolutely no competition. Those rarely exist.

It’s to find a problem where your understanding makes you difficult to replace.

Because the next generation of successful AI freelancers won’t necessarily be the people who know the most prompts.

They’ll be the ones who understand what clients actually need.

Stop selling AI. Start selling the problem it solves.

AI Freelance Niches That Aren’t Already Overcrowded was originally published in Freelancer’s Hub on Medium, where people are continuing the conversation by highlighting and responding to this story.

Many people are usually afraid to be alone mainly to avoid the silent mode.

I’m not referring to the silence where you’re alone doing nothing or probably thinking about something not related to your entrepreneurship journey. The silence I’m talking about is the silence of effort, the one nobody sees, the one that kept you doubting whether it’s going to work or not, the one that feels like “Should I give up or try again?”

Nobody sees any of that. They only see the results later on and call you “lucky”.

That kind of silence is where most of the actual building happens.

I came into the world of entrepreneurship in the year 2020 with no clue of how to start and scale something. All I had was just ideas. No money, no mentor, no friend, or anything at all that tells me “Bro, you’re getting it, keep going” or maybe “Bro, this is wrong, do it this way”

I started by experimenting the internet hustles: SaaS, freelancing, affiliate marketing, dropshipping, and YouTube. All gave me signs (traction, tiny audience), and few bucks in my pocket from different platforms (my first online income was $60 in 2021) and zero income in others.

I do this silently over the years which nobody sees. Sometimes I feel like quitting, sometimes I feel like trying something else, and sometimes I feel like nothing is working.

All I wanted was to create something that pays me every day so that I can live my dreams and relax for the rest of my life.

If people knew about your personal hustles, how would you feel? I’m not going to be happy at all, because I don’t like people knowing about my craft unless they want to support it and make it grow.

So I chose to build in silence.

And what I discovered wasn’t something most people can comprehend unless you’re a solo entrepreneur. The phase that happens before the actual story, “Hey, how did you do it?”

There’s no success without a story behind it, no matter how short or long it took. The silence phase is the key elevator of many successful businesses: ideas come, you see things clearly, you create a plan, you test into the real world by trying it out, it failed, you re-strategize, and failed again, you get depressed overtime, come back again, and finally it worked.

That’s what entrepreneurship silence does.

One Important Thing About Silence Is That It Gives You a Reply

In the beginning phase, silence is manageable. You started something new. Your energy is full. You toil. Things get built. You believe strongly that it’d work out. Momentum is running. Nothing is happening publicly. Nobody around you sees it. You can feel it.

Later on, silence begins to whisper.

Then you begin to ask yourself:

• “Would this actually work as I expected it to?”
• “What of if it doesn’t work, what will I do?”
• “Who would help me make it work?”
• “Am I lacking something else?”
• “Do I need a Plan B?”

It’s painful to ask yourself these questions especially when you’re doing it alone. You want a feedback. You want success to appear before you and tell you “It’s going to work”

The silence phase feels like rejection even when it’s just common silence. It’s simply time passing, not failure reading.

I learned one important thing about the silent mode: It’s not an evidence of failure, but the normal texture of progression.

Every person who built something real on the internet or outside of it all passed through the silent phase, including Elon Musk (started with Zip2 with no revenue), Jeff Bezos (Amazon took years before making profit), and many more.

Silence is part of the process, not a proof that it’s not working. What feels like nothing today becomes the beginning story of something tomorrow.

The Wrong Part About The Story They Don’t Talk About

Most of the success stories we consume usually have one pattern of outcome: “after everything, it worked”

The listeners already know the end of the story. So every hard moments in the story carries a kind of hope, the one that says “No matter what you’re going through, it’d finally work out at the end”

This is not 100% guaranteed. The brutal truth is: You can work for many years without making it. It sounds harsh, but it’s the bitter truth. That gap between the beginning and the end of the story is where silence lived.

When you’re building all alone, you have only your personal decision, but no guarantee. You go through days where your decision feels like clarity and also days where it feels like error. You do what you do — but getting different results across several period of time for reasons that aren’t able to be seen.

All you have is hope without a success timeline. Nothing is certain. And this kind of feeling is one of the loneliest parts of success.

The story tellers usually don’t tell you this but I will: success is not only when you’re making money. A poor written article that gave a bad feedback is also success, because you’ve learnt what a poor article is and would do better next time. The product with 1-star review is also success, you got a feedback from real people, understood what killed your product, you learned on how to improve it for better. And so on.

World definition of success: once it’s making money, then you’re successful. Once you’re married, bought a house, car, clothing, or on a vacation, then you’re success.

My own definition of success: success is the sense of satisfaction or the fulfilment you get when you’ve achieved what you wanted to achieve. No matter how good or bad it went. No matter how short or long it took. If it fails, you keep trying, until it works. Success is anything backed by decision:

• If you want to stay broke for the rest of your life and you did, then you’re successful
• If you want to cough and you did, then you’re successful
• If you want to lie and you did, then you’re successful
• If you want to take your bath and you did, then you’re successful
• If you want to buy food at the supermarket opposite you house and you did, then you’re successful
• If you want a girlfriend and she accepted, then you’re successful
• and many more.

Lemme go deeper:

• If you inherited a $500M mansion but you don’t like the mansion or maybe the mansion is not the kind of mansion you’ve be dreaming of, then you’re not successful
• If you’re not married to the man or woman of your dreams, then you’re not successful
• and many more.

The world tied success only to material things. I made you understood what success actually is which nobody taught me about.

The Time the Silence Breaks

What actually breaks the silence? The moment things start working: your product made a sale, a client ordered your service, your startup got a user, your article got recommended, and so on.

What triggered it? The silent period when nothing feels like its working but actually working in the background which you cannot see.

It feels great to watch your hard work payoff. No longer the absence of result or feedback, but a different level of loneliness. The one that stack money, clients, and results silently without any public awareness.

Hear me out, if you’re currently inside the silence phase, understand that: the silent phase is just the natural part of the process. It must exist before something exists. Not a problem. Just a mode.

What I Want Anyone Within the Silence Phase to Know

Your silent phase is not a sign of failure. You’re just early. You’re just new to the system. Your content or offer is still on the testing phase which the first conversion stage of everything that works.

Don’t take it personal. Don’t assume that your idea is bad. Don’t give up too soon. Ignore the lies your brain is telling you. Overlook anyone telling you that it’s not going to work.

Your doubt is real. The silence is real. Silence is the price you pay for building something before the algorithm, market, or audience has caught up to it.

You matter. Your idea is unique, just that it hasn’t be proven in the real world. You’re the next big thing in the industry. Don’t quit. You’ll make it if you persist. Ride on.

Conclusion

I’m both in the silent phase and in the winning phase. My Medium, Pinterest, and LinkedIn profiles are still growing. The digital assets I built years ago are still gaining traction up till today (traffic, sales, contacts).

I’m still building. Nothing is permanent once you keep going. It’s just a matter of time. There’s always a next level after a level.

Don’t think you’re alone; I’ve been there and I’m still there.

Building something nobody knows today consistently, is something they’ll know tomorrow. Your silence is not the absence of progress but the presence of it.

Let’s Connect

I’m happy you made it down here. If this hit — give me your claps if the article deserves it, repost the article to your network if you believe it’d help, but most importantly, follow me right away, so that you won’t miss my future masterpieces — the one that helps you build, grow, and scale beyond.

Building Alone in Silence Is Where It All Begin — Most People Quit On This Phase was originally published in Freelancer’s Hub on Medium, where people are continuing the conversation by highlighting and responding to this story.

A while ago, every freelancer seemed to become a prompt engineer overnight.

LinkedIn bios changed.

AI expert appeared everywhere.

People started selling giant prompt libraries containing 500 ways to ask ChatGPT to write an email.

For a minute, it felt like learning a few AI tools was enough to build an entire freelance business.

That phase is changing.

Businesses are becoming less impressed by AI itself.

They care about what it can actually do for them.

Save five hours a week?

Great.

Reduce repetitive admin?

Interesting.

Turn messy company information into something employees can actually find?

Now we’re talking.

That’s where I think the better freelance opportunities are appearing.

Not in selling AI because it’s trendy.

In using AI to solve boring, expensive problems.

Here are nine niches worth watching.

1. AI Workflow Automation

This is probably the most obvious opportunity, but there’s still a huge difference between knowing automation tools and understanding business processes.

Companies have repetitive tasks everywhere.

Lead follow-ups.

Customer inquiries.

Meeting summaries.

Data entry.

Internal notifications.

Reporting.

A freelancer can map those processes and build automations using AI alongside tools like Zapier or Make.

The key is positioning.

Don’t sell a fancy automation.

Sell the ten hours it could save every month.

Businesses understand time.

2. AI Knowledge Base Setup

Companies have information scattered everywhere.

Google Drive.

Notion.

Slack.

PDFs.

Random documents someone created three years ago and apparently became the only person who knows where they live.

Then the company wants an AI assistant that can answer internal questions.

There’s a problem.

AI can’t magically organize bad information.

Freelancers who can clean documentation, structure knowledge bases, remove outdated material, and prepare information for AI-powered retrieval can become extremely useful.

It’s not glamorous.

That’s part of the opportunity.

3. AI Content Quality Control

Creating AI content is easy.

Publishing good AI-assisted content is harder.

Businesses using AI for blogs, emails, product descriptions, support responses, and social posts still need someone checking the output.

Is it accurate?

Does it sound like the company?

Is information outdated?

Does every paragraph sound suspiciously like a robot wearing a blazer wrote it?

AI quality control combines editing, fact-checking, brand consistency, and human judgment.

The stronger niche is going specific.

Become the person who reviews AI-generated content for one particular industry or type of business.

Specific expertise makes generic AI output much more useful.

4. AI Chatbot Optimization

Building a chatbot isn’t the end of the project.

It’s usually the beginning.

Businesses need someone to review conversations, identify where users get stuck, improve responses, update information, and decide when the bot should hand someone over to a human.

That’s an ongoing service.

And ongoing services are interesting for freelancers because you’re not constantly starting from zero.

Instead of selling chatbot setup, consider selling chatbot performance.

That changes the conversation completely.

5. AI Training for Small Teams

Plenty of employees now have access to AI tools.

That doesn’t mean they know how to use them well.

Some barely use them.

Others paste everything into them and hope for the best.

Businesses need practical training around real workflows.

A freelancer could build workshops for marketing teams, customer service departments, recruiters, administrative staff, or sales teams.

Not a two-hour presentation explaining what artificial intelligence means.

Nobody needs another one of those.

Show people how AI fits into the work they’re already doing.

That’s much easier to sell.

6. AI-Powered Research Services

Research takes time.

Competitor research.

Customer research.

Market research.

Content research.

AI can speed up parts of that process dramatically, but businesses still need someone who knows what to investigate, which sources deserve trust, and how to turn information into something useful.

That’s the opportunity.

You’re not selling AI research.

You’re selling a useful briefing, competitor analysis, trend report, or decision-ready summary.

AI becomes part of your workflow.

Your judgment remains the product.

7. AI Customer Support Systems

Customer support can become repetitive quickly.

Where’s my order?

How do I reset my password?

What’s your refund policy?

Can I change my appointment?

Businesses can use AI to handle parts of those interactions, but someone still needs to design the system.

That includes organizing FAQs, writing response guidelines, creating escalation rules, testing answers, and improving workflows over time.

This niche becomes particularly interesting when you specialize.

AI customer support for e-commerce stores is much clearer than AI consultant for everyone.

Clients understand specific offers faster.

8. AI SOP and Documentation Creation

Every business has processes that live exclusively inside someone’s brain.

Then that person goes on vacation and suddenly nobody knows how anything works.

AI can help turn recordings, meeting notes, existing documents, and explanations into structured standard operating procedures.

A freelancer can interview team members, document workflows, organize instructions, and build an internal operations library.

The AI speeds up drafting and organization.

The freelancer makes sure the final documentation actually reflects reality.

You’re essentially helping businesses turn tribal knowledge into reusable systems.

That’s valuable.

9. AI Implementation for Boring Industries

This might be my favorite opportunity.

Everyone wants to build AI services for startups and creators.

Meanwhile, thousands of less glamorous businesses still deal with repetitive paperwork and manual processes.

Property managers.

Accounting firms.

Recruiting agencies.

Contractors.

Logistics companies.

Local service businesses.

These companies don’t necessarily want an AI transformation strategy.

They want fewer missed leads.

Faster estimates.

Better follow-up.

Less admin.

Pick one industry and learn its problems deeply.

Then build simple AI-supported services around those problems.

You don’t need to know everything about AI.

You need to understand your client’s Tuesday afternoon.

The Opportunity Is Getting More Specific

I think the generic AI freelancer is going to have a harder time standing out.

AI automation expert tells me very little.

AI workflow specialist helping property managers automate maintenance requests tells me much more.

That’s the shift.

Tool expertise matters.

But context makes that expertise valuable.

So don’t try to learn every AI platform released this month.

You’ll lose your mind before you finish updating your bookmarks.

Pick one problem.

Pick one type of client.

Learn how they currently solve that problem.

Then figure out where AI genuinely makes the process faster, cheaper, or easier.

Build your offer around that outcome.

Because clients aren’t sitting around hoping someone sells them more AI.

They already have plenty of that.

They’re looking for someone who can turn AI into something useful.

And that might be one of the best freelance opportunities hiding in plain sight.

9 AI Freelance Niches That Are Actually Taking Off was originally published in Freelancer’s Hub on Medium, where people are continuing the conversation by highlighting and responding to this story.

The project was a four-article content package. Fixed price, three weeks, two rounds of revisions. Clean.

Then, in week two, the client wrote: “While you’re in there, could you also draft a LinkedIn post for each piece?” Fine. Twenty minutes each. A few days later: “Can you rewrite them for our SEO keyword list?” Sure, probably an hour. Then came a request to “tidy up” older blog posts, which turned out to be eight articles that needed full rewrites for tone.

No single request felt worth pushing back on. Together they added about 14 hours to a project I’d priced at 30. I only noticed at invoice time.

That’s how scope creep usually works. It rarely arrives as one big change. It shows up as small, friendly asks, each too minor to make a fuss about. By the time you notice, the work is done and the price is fixed.

Below: five early signs, a simple way to quote or close each extra request, and an email for when a project has already drifted.

Why scope creep is hard to see from the inside

Mid-project, you’re focused on doing good work, not tracking what was agreed. And saying “that’s extra” about a 20-minute task can feel petty, especially with a client you like.

The client usually isn’t scheming. Most people genuinely don’t know where “four articles” ends and “a bit more content” begins. If you don’t draw that line, nobody will. The goal isn’t rigidity. It’s noticing drift while it’s still cheap to talk about.

The 5 signs your project is growing past the brief

1. Requests start with “while you’re at it” or “quick one.” These phrases frame new work as part of the old work. “While you’re finishing the hero post, could you also write the About page copy?” isn’t a revision. It’s a new deliverable. When you hear these openers, check the brief before answering.
2. New people join the project with their own opinions. You were working with the marketing lead. Now the founder, a sales manager, and someone’s cousin who “writes great LinkedIn posts” are leaving comments. None of them were there when scope was agreed, and more voices almost always mean more work.
3. Feedback rounds stop being about fixes and start being about direction. A revision is “shorten the intro” or “swap this example.” A change of direction is “actually, we want it to feel more playful” in round three. If you’re redoing work that matched an approved angle, you’re not revising. You’re starting a new draft.
4. The deliverable list changes, but the deadline and price don’t. You quoted for four articles. Now they need social captions, email subject lines, and meta descriptions, all “by the same Friday.” When the list grows and nothing else moves, the agreement hasn’t caught up with the project.
5. You’re doing tasks that aren’t in any document. Uploading drafts into their CMS. Writing product descriptions. Joining a weekly call nobody mentioned at the start. A simple test: if you can’t point to the line in your proposal that covers a task, it’s outside the scope. An occasional favor is fine. A weekly default is not.

One sign alone isn’t trouble. Two or three at once means it’s time to talk.

Quote it, swap it, or close it

When an extra request lands, you need one of three short replies, not a big conversation.

Quote it. Treat the request as a small new piece of work with its own price. “Happy to add LinkedIn posts for each article. That’s outside the current scope, so it would be $150 and add two days. Want me to go ahead?”

Swap it. If the budget is fixed, offer a trade. “We can do the full SEO rewrite if we drop the newsletter intro. Same price, same deadline. Which would you prefer?” The project stays the same size, and the client chooses what matters most.

Close it. Some requests don’t fit now. Park them instead of saying a flat no. “Great idea. Let’s add it to a phase two list and look at it after this batch ships.” A shared “later” list shows you listened without committing you to anything.

Notice what’s missing: no apology, no long explanation. A calm reply makes extra work sound like a normal business decision, because it is.

One habit helps: answer extra requests in writing, even if they came up on a call. A two-line follow-up (“Just confirming: the extra About page copy would be $300, and I’ll start once you say go”) prevents “I thought that was included” later.

The email that puts the scope back in place

Sometimes you notice the drift late, after several extras have piled up. That’s the moment for one clear reset email. Here’s a version to adapt:

Subject: Quick check-in on the content package

Hi Sarah,

Hope your week’s going okay. The four articles are in good shape — I just wanted to pause for a second before we keep adding things.

We originally agreed on four articles, two revision rounds, and delivery by March 14. Along the way we also picked up:

• LinkedIn post for each article (done)
• SEO keyword rewrite across all four (in progress)
• Full rewrite of eight older blog posts (not started)

I was happy to throw in the LinkedIn posts. For the rest, I don’t want the deadline (or the writing) to get messy, so here are two easy options:

Option A: Keep the SEO rewrite and older-post cleanup in this round for $450, and move delivery to March 21.

Option B: Stick to March 14 with the original four articles, and treat the extras as a short follow-up after this batch.

Both work for me. Reply with A or B whenever you get a chance and I’ll lock the timeline today.

Thanks,
Nick

Why it works: it sounds like a normal check-in, not a confrontation. It lists the extras so the client sees the drift themselves, offers two real choices instead of a demand, and ends with a yes/no next step.

You don’t have to bill retroactively for extras you already did free. Mentioning them (“I was happy to throw in the LinkedIn posts”) shows goodwill and makes the boundary feel fair rather than sudden.

A boundary you can reuse on every project

The easiest fix is to make extras a normal part of the process from day one. Add one short paragraph to every proposal:

“This quote covers the deliverables listed above. Anything outside that list is welcome. I’ll send a short quote for it before starting, so there are no surprises on either side.”

It tells the client extras are possible, sets the expectation that they cost something, and means your later quote follows the agreement instead of inventing a new rule.

Pair it with a five-minute check each Friday: compare what you worked on against the brief, and raise any mismatch on Monday while it’s still small.

FAQ

Is it okay to do small extras for free?

Yes, sometimes. A quick favor can build goodwill. The key is to name it as a favor (“Happy to throw this in at no charge”) so it doesn’t quietly become the new baseline.

What if the client gets upset when I quote an extra?

Most won’t, especially if you stay calm and offer options. If one regularly refuses to pay for clearly new work, that’s useful information for next time.

How much should I charge for small additions?

Use your normal rate with a sensible minimum, such as one hour. Tiny tasks still involve context-switching and back-and-forth.

Should I stop working while we sort out the scope?

Keep going on the original scope and pause only the extras until the client confirms. The project keeps moving, and nobody feels held hostage.

Related reading

Scope creep often starts with a vague brief, and Build a Client Intake Checklist Solo Freelancers Can Reuse Forever helps you pin details down early. If your base price leaves no room for extras, Raise Your Freelance Rates in 2026 Without the Awkward Email covers how to adjust. And if a client ignores every boundary, How Solo Freelancers Say No to Bad-Fit Clients Without Guilt helps you decide whether to work with them again.

Scope creep isn’t a sign of a bad client or a bad freelancer. It’s what happens when nobody watches the edges of a project. Learn the five signs, save the reset email, and add the boundary paragraph to your next proposal.

If you want more down-to-earth help with the business side of freelancing, like pricing, clients, and protecting your time, follow along. I write for solo freelancers who’d rather fix the system than work longer hours.

What’s the smallest “quick one” that turned into the biggest job you’ve ever done? Share it in the comments.

5 Signs Your Freelance Project Is Quietly Growing Past the Brief was originally published in Freelancer’s Hub on Medium, where people are continuing the conversation by highlighting and responding to this story.

I sold my nights for more money, and you can sign the same deal without noticing.

On the side of my company, I was selling automations to businesses: small programs that do a repetitive task on their own, the way an assistant would.

The promise was to take manual work off their hands and make them more productive.

Each one was built by me, installed by me and looked after by me, and it had to run day and night, 24 hours a day and 7 days a week.

In exchange I received a little more money each month, and without noticing I had traded my sleep for it.

Here is the receipt: 6 of those programs that I kept alive myself, and the one night the machine running them stopped, 4 customers wrote to me before I had eaten breakfast.

There is a name for it

The name is sleep debt, and the idea behind it is simpler than the word suggests.

A short night does not cost you one night.

It costs you the day after, because each decision you take is a little worse than it would have been.

You cannot see the difference from the inside, since the judge of your decisions is the same tired brain.

This is why the debt shows up as a bad call in the afternoon, on the job that pays for the project.

Most of the time I ran on caffeine (coffee, energy drinks) to keep the pain away for a while, and the bill came later.

Saturday does not pay it back either, because the day it cost you has already been spent.

A little more money each month that keeps you awake is a loan taken against those days, and the repayment falls due whenever a machine decides to stop.

What it looked like

This was less than 6 months after I created my company, LogiGreen, where we help companies work in a more efficient way with data and AI.

My first project was with a large factory in Switzerland, and it took a large part of my time: long meetings, workshops to organise, and site visits, since I live in Paris.

At that time I had started a dopamine detox, cutting the things that kept my brain over-stimulated: videos, music, coffee.

To keep my brain busy I studied a new skill, automation, with a tool called n8n that lets you build those small programs without being a developer.

One example: a customer asks by email when his order arrives, the program reads the email, takes the date from a spreadsheet, then answers him.

I published what I built, and people started writing to ask for programs like that for their own business.

I took a first project and then a second one, and everything was going well: projects were piling up and money was flowing into the company.

With both, I took full ownership: I installed the program, and I was the one keeping it alive.

It ran on a machine I rented and looked after myself, and I invoiced a monthly fee for keeping it running 24 hours a day, 7 days a week.

As soon as I started to sign projects and to get that additional money, I forgot the dopamine detox.

After 2 projects, 4 more arrived, from different industries and with different levels of complexity.

At the beginning the programs ran well, and when one broke it was usually a file that did not look like the one I had been shown, or a case I had not been told about.

I fixed those from my laptop, sometimes in the middle of a meeting with my main customer.

The customers in other time zones changed the arithmetic: while I slept, they were at work.

The first night it happened, a program stopped while I was asleep and I woke up to a voice message.

I fixed it the next day, and the customer told me it may have cost him money while it waited for me.

From that night I carried one question awake and into bed: what happens if something breaks while I am not here?

One day the machine that ran all of it went down.

The next morning, 4 customers had written to me, and the first thing I did with my eyes open was check the phone.

I skipped breakfast and went straight to the computer.

Later I built a watchdog, a small program that checked on the others for me, so that at least a machine would wake before I did.

But the decision that mattered came after that, and it was not technical: I stopped selling my nights, and the three moves below are what replaced them.

The receipt

Here is what the deal looked like once I put it on paper: the numbers in the order they arrived, and what each one did to my nights.

• 2 projects, then 4 more: 6 programs that I kept alive myself
• if one stopped, the customer felt it at once, and I had to put it back up as fast as I could
• running promised 24 hours a day, 7 days a week, written into a monthly fee
• customers in different time zones, so night had stopped meaning anything
• 1 failure that waited for me to wake up, and 1 customer who may have lost money
• 1 night when the machine stopped, 4 messages, 1 morning that started at the computer

Three moves

Three things changed after that morning, and you can copy them without keeping a machine alive for anyone.

1. Get paid to build, and let the customer keep it alive

You take the problem, design the solution and install it.

The customer owns the running of it, on their own machines or with someone whose job it is to be awake.

A fee for keeping something alive at night is a fee for your nights, and your nights are not for sale at that price.

2. Put a machine on watch before you put yourself on watch

The watchdog I built afterwards, a small program whose only job is to check that the others are alive, could have existed before the first customer signed.

A program that notices it has stopped, tries again on its own and writes one line in a report turns a voice message in the night into something you read after breakfast.

3. Decide what the first hour of the day is for while you are rested

The morning I skipped breakfast to open a laptop, the anxiety had made that decision for me.

You can make it in advance and write it down, and the phone stays where it is until the hour is over.

The debt collects on judgement, and the first hour of the day sets it.

Your nights are not for sale

Money that only arrives while you are awake in the middle of the night is a loan, with you as the collateral.

Follow Built on the Side here on Medium: one story every Wednesday, with the notion it hides.

I am Samir, an engineer who started his company after ten years in supply chain across Asia and Europe, with too many things built on the side.

I am not a therapist and I am not a coach. I kept the receipts, and I tell you what it cost in the head.

The stories are here, one every Wednesday. The rest is on my blog and LinkedIn.

I Sold My Nights for More Money: What Sleep Debt Costs a Side Project was originally published in Freelancer’s Hub on Medium, where people are continuing the conversation by highlighting and responding to this story.