$1,000+ in free AI tools. Access Offers →

If I had to rebuild my income from zero, I'd start with something I could sell this week.

Rent wouldn't wait for me to build a personal brand. And spending a month learning a new AI tool wouldn't help much if I still had nobody to sell to.

AI would be useful for the research and first drafts, or for turning messy notes into work I could send a client. But I'd still need to find someone with a problem I knew how to fix.

Here's how I'd approach it.

When you're starting over, it's tempting to pick a skill you'd like to learn. I'd look at what someone might pay you for now.

A former office manager could offer to sort out a business owner's inbox and calendar. That's a service a buyer can understand. "I'm learning AI automation" doesn't give them much to go on.

You may need a new way to package your experience rather than a new skill. AI can help you work through that, especially when stress makes it hard to see what you have to offer.

Before opening an AI tool, write down your answers to these questions:

Then paste your answers into your AI assistant:

Don't keep an idea just because it sounds impressive. Look for a buyer you could actually reach and a problem you know enough about to solve.

"I help businesses with AI" leaves the buyer doing all the work. They have to figure out what you do and whether they need it.

Compare that with:

"I turn a real estate agent's messy listing notes into ready-to-post property descriptions in 24 hours."

Now they know what they'd get and when they'd get it.

Before choosing an offer, try describing where the problem shows up in that person's workday. What does it cost them? What have they already tried? If you can't answer those questions, talk to a potential buyer before spending days building a service around a guess.

It's easier to show someone a finished sample than to explain what you might be able to do.

For listing descriptions, write samples for three fictional properties. If you're offering spreadsheet cleanup, use a sample spreadsheet to show the before and after. Keep fictional examples clearly labeled.

Once you've done the work, AI can help you explain it:

Making the sample also gives you a chance to try the work before committing to a month of it. You might discover that a service you thought would be easy takes far longer than you'd want to charge for.

Keep the first offer small enough to explain in one read:

For example: "Five ready-to-post listing descriptions from your rough notes. Delivered in 48 hours. $150. Includes one round of revisions."

I'd be strict about scope here. When you need money, it's easy to agree to every extra request. Then a small project takes three days longer than you priced it for.

Start with one package you can deliver reliably. You can offer more later.

I'd start with people who already know me rather than wait for social posts to bring in clients.

Make a list of 30 people. Include former coworkers and past clients, but don't overlook people you know outside work. Someone from your gym or your kid's school may know a business owner who needs help.

Aim to send five short, personal messages a day during the week. Something like:

AI can help if you're stuck on the wording. Still, read each message before sending it and make sure it sounds like something you'd say to that person.

Use a plain spreadsheet to track the name, date sent, reply, and follow-up date. After three days, follow up once. If they don't respond, leave it there.

Once you have a paying client, use AI where it saves you time:

You're still responsible for what you send. Check names and facts. Read for the client's voice, and cut anything you'd feel awkward saying out loud.

A wrong name or an invented statistic can undo the time you saved. Leave room for review when you promise a delivery date.

If you can deliver early without rushing the checks, do. Giving a first client a reason to recommend you is worth the effort.

Week 1: Inventory and proof. Work through your experience, choose a problem you can solve, and make a sample.

Week 2: Offer and outreach. Write the index-card offer and run your 7-day outreach sprint. Follow up once.

Week 3: Deliver and document. Do the client work. Keep a checklist of the steps, including where AI helped and where you needed to fix its output.

Week 4: Repeat and raise. Offer the same package to the next 30 names. If two clients bought it, try raising the price 20% for the next round.

I'd leave the logo and website rebuild alone for now. You don't need to finish either before offering five listing descriptions to an agent.

Trying another tool. It's easy to spend an afternoon testing an AI app because that feels less uncomfortable than asking someone to pay you. Stick with one assistant while you test the offer.

Watching tutorials without using them. Look up what you need for the work in front of you. A month of free videos won't tell you whether someone will buy your package.

Pricing out of panic. A $25 package can still turn into $250 worth of requests if you haven't set boundaries. Be clear about what the price covers, including revisions.

Quoting every project from scratch. Keep track of the work you repeat. If the package takes just as long on the fifth delivery as it did on the first, look at what's slowing you down before taking on more.

I'd spend the first 30 minutes on the skills inventory. Then I'd choose one service and make a sample before worrying about a website or a bigger plan.

The point of using AI here is to spend less time preparing and more time talking to potential buyers and doing paid work. You'll still have to ask for the sale. That's the part no prompt can do for you.