How to Start a Career in AI

A generation ago, career advice sounded straightforward: get a good education, polish your résumé, find a reputable employer, and climb the ladder. But what happens when the ladder itself is being rebuilt while you’re climbing it?

If you’re in university, recently graduated, or considering a career change, artificial intelligence presents an unusual proposition. It may disrupt the work you were preparing to do. At the same time, it may give you the tools to create a career that didn’t exist a year ago.

The question isn’t simply, “How do I get a job in AI?” It’s also: What can I do with AI that someone else will find valuable?

First, what counts as a career in AI?

Some people build AI: researchers, machine-learning engineers, model developers, infrastructure specialists, and safety experts. Others put it to work: automation consultants, AI product designers, implementation specialists, trainers, and business owners who solve industry-specific problems with AI.

You don’t have to work at OpenAI or build a foundation model to have a career in AI. You could help a dental practice reduce missed appointments, a manufacturer improve quality-control reporting, or a construction firm turn site notes into estimates. If applying AI is central to the value you deliver, that’s a legitimate AI career.

Look at the leaders associated with today’s AI revolution. Some, like Geoffrey Hinton, devoted decades to AI research. Others entered AI through earlier careers in computing, commerce, or technology. Both routes exist: build your career in AI, or bring the career you already have into AI.

Your existing knowledge may be your competitive advantage

Consider the North American Industry Classification System, or NAICS. It divides the economy into sectors and industries. Now imagine pairing AI expertise with virtually any one of those industries.

AI plus healthcare. AI plus hospitality. AI plus agriculture. AI plus logistics. AI plus accounting. AI plus skilled trades.

Every combination creates potential problems to solve, workflows to improve, and customers to serve. Not every idea will become a viable business, but the opportunity space is enormous.

Someone who has spent ten years managing a warehouse may understand inventory bottlenecks better than a brilliant programmer who has never set foot in one. Add genuine AI competence to that domain knowledge and you have a potentially powerful combination.

This is why I wouldn’t tell everyone to quit their current profession and start over. Sometimes the smartest move is to become the person in your industry who knows how to make AI useful.

I’ve had to reinvent my own business more than once over more than 30 years, long before AI came along. Each time the tools changed, I took what I already knew and found a new way to make it useful. That habit of reinvention has been one of the keys to my success, and I believe it will matter even more for young people starting out today.

Meet our hypothetical graduate

Imagine a 24-year-old woman with a bachelor’s degree in business administration. She’s smart, ambitious, comfortable with technology, but has no programming background. She has student debt, rent to pay, and no wealthy parents funding another degree. She wants a career in AI. What should she do?

First, let’s distinguish using AI from implementing it. Most graduates have encountered AI on their phones, in search, or through chatbots. That’s not the same as knowing how to design a reliable workflow for a business, evaluate outputs, protect confidential information, or tailor a solution to a particular group of users.

AI literacy is becoming a basic professional expectation. The differentiator is the ability to put that literacy to work.

Should she return to university? Perhaps—especially if her goal is advanced research or a highly technical position requiring rigorous mathematics and engineering. A bachelor’s, master’s, or PhD can open particular doors. But another degree isn’t a universal ticket to success, and it isn’t the only route into applied AI.

If she needs income now, a day job combined with deliberate evening study and small projects may be more realistic than taking on substantial new debt. Self-directed learning is another route. The right choice depends on her finances, interests, strengths, and intended specialization.

Start with something you actually care about

Before recommending a course, I’d ask her a different question: What else are you passionate about?

Sports? Cooking? Travel? Retail? Music? Small-business operations?

You’re likely to spend forty, fifty, or even sixty hours a week working. If you can connect your AI ambitions to a subject you understand and genuinely enjoy, you have a better chance of sustaining the effort when things get difficult.

Passion alone doesn’t pay the bills, of course. It needs to meet a real need. But a person who understands both the technology and the problem has an advantage over someone who simply announces, “I do AI consulting.”

What Sun Tzu can teach us about entering AI

In The Art of War, the ancient strategist Sun Tzu emphasizes preparation, understanding the terrain, using information wisely, and adapting to change. It’s a book worth owning no matter what you do, and its lessons translate surprisingly well to building a career in AI.

For our graduate, that means knowing her own strengths and limits, studying the real problems businesses need solved, and choosing her battles: helping one local company with a concrete problem may be a better first step than trying to build the next major AI platform.

Sun Tzu didn’t anticipate AI, but the lesson holds. You don’t need to be the smartest person entering AI. You need to be smart about how you enter it.

Proof beats promises

A résumé still matters, and recognized qualifications can carry weight. But a résumé is a claim about what you might be able to do. A working demonstration is evidence.

What have you built? What have you improved? What have you published? Can you show a before-and-after result? Can someone verify your contribution?

Create a simple website. Publish a short case study. Record a demonstration on your phone. Build a small app, an automation, or a useful research tool. Explain the problem, your approach, the limitations, and the outcome. You don’t need a television studio; a smartphone and a quiet room can be enough.

And make it discoverable. LinkedIn, your own site, relevant professional communities, and other platforms can help establish a public record of your work. A large following isn’t a prerequisite for competence, and privacy is a legitimate choice, but potential clients and employers need some credible way to assess what you can do.

Here’s a line I keep coming back to: “The harder I work, the luckier I get.” Initiative and momentum increase the number of opportunities you encounter. Sitting still doesn’t.

Five ways to turn AI skills into income

There isn’t one correct pathway. Depending on your circumstances, you might:

1. Join an AI-focused employer. Build specialized skills and apply for roles in product, operations, customer success, implementation, research, or engineering.
2. Become the AI specialist in your existing field. Improve workflows where you already understand the people, processes, and pain points.
3. Offer a small proof of concept. Approach a business with a specific proposal: “Give me a problem you can safely share, and 48 hours to outline or prototype a possible solution.” Keep the scope small and protect confidential information.
4. Freelance or consult. Try professional networks, referrals, LinkedIn, Fiverr, Freelancer, or other channels. A narrowly defined service with demonstrable value is easier to sell than a vague promise to “transform your business with AI.”
5. Invent your own product or service. Find an unmet need, build something modest, test it with real users, and improve it. You don’t need to begin with a venture-capital pitch or a fantasy of a multimillion-dollar Shark Tank deal.

Your personality matters. Someone comfortable presenting may excel at pitching businesses. A quieter builder may prefer creating an impressive product and letting the demonstration do much of the talking. Geography, family obligations, money, and experience all influence the choice.

What matters is getting out into the world, testing ideas, and learning from the response. Not every company will give you a meeting. Not every prototype will work. It’s a numbers game—and persistence matters.

Which jobs will AI change most by 2030?

Let’s look ahead. Nobody knows exactly what the labour market will look like in 2030. What follows is my read on where things are heading by 2030, not a statistically validated ranking. They concern exposure to AI-driven changes in tasks, not guaranteed job losses.

Ten occupations likely to experience substantial change

1. Data-entry clerks: Routine information transfer and validation are increasingly automatable.
2. Customer-service representatives: AI agents can handle more inquiries and transactions.
3. Bookkeepers and payroll clerks: Reconciliation, processing, and routine reports can be streamlined.
4. Administrative assistants: Scheduling, correspondence, and document preparation are changing.
5. Translators: Routine translation faces growing automation, while nuanced specialist work remains important.
6. Copywriters and content producers: Commodity commercial content is increasingly easy to generate.
7. Junior software developers: AI-assisted coding changes the entry-level mix of tasks and skills.
8. Paralegals and legal assistants: Research, document review, and drafting are becoming AI-assisted.
9. Market research analysts: Information gathering and first-pass analysis can accelerate dramatically.
10. Travel agents: Planning and comparing itineraries can increasingly be performed by AI assistants.

Ten occupations relatively resistant to near-term replacement

1. Plumbers: Every old building presents physical surprises.
2. Electricians: On-site installation, troubleshooting, and safety judgment remain difficult to automate.
3. HVAC technicians: Real-world diagnosis and repair require hands-on adaptation.
4. Roofers: Physical work in changing outdoor conditions is challenging for robots.
5. Carpenters: Custom work and renovation demand dexterity and judgment.
6. Nurses: AI can assist documentation and clinical decisions, but bedside care requires human presence.
7. Physiotherapists: Treatment combines physical interaction and personal assessment.
8. Personal support workers: Hands-on assistance and human relationships are central.
9. Firefighters: Emergencies are unpredictable, dangerous, and physically complex.
10. Early-childhood educators: Supervision, trust, and emotional development are difficult to replicate.

These professions won’t be untouched. A plumber may use AI to prepare estimates; a nurse may use it to reduce paperwork. And progress in robotics could change the picture. Being affected by AI isn’t the same thing as being replaced by AI.

But there’s a remarkable reversal worth considering. For generations, families often viewed white-collar office work as a safer career bet than manual trades. By 2030, some people who spend their days processing information at computers may face more immediate automation pressure than the electrician working in their basement.

And here’s the opportunity for an aspiring AI professional: those same electricians, plumbers, and HVAC contractors still need scheduling, estimates, marketing, inventory systems, and customer communications. The jobs hardest to replace may belong to businesses most eager for useful AI services. I work with the trades more than any other sector, so if you run one of those businesses, this is exactly the work I do.

The window is open—but not forever

My expectation is that the next several years will offer substantial opportunity to people who can help organizations implement AI responsibly and effectively. Many businesses are still figuring out what to adopt, what to avoid, and how to make it all work.

But I wouldn’t promise that today’s AI consulting boom lasts indefinitely. As AI systems become more capable, some of the implementation work people charge for today may itself become automated. The skill worth cultivating isn’t mastery of one fashionable tool. It’s the ability to understand a real problem, learn quickly, exercise judgment, and deliver a result.

By 2030, knowing how to work with AI may matter more to your career prospects than knowing how to perform many of the tasks AI can do for you.

The future doesn’t come with a guarantee

Nobody can guarantee you a successful career in AI. Not a university degree, not a prestigious employer, and certainly not a certificate hanging on your wall. The world is changing too quickly for guarantees. And the pace of change is accelerating.

You can look at that uncertainty with fear, cling to the familiar, and hope someone will eventually tell you what to do. Or you can recognize this extraordinary moment in history for what it is: an opportunity to create something that didn’t exist yesterday.

Learn everything you can. Experiment. Build something. Find a problem worth solving. Show the world what you’re capable of.

And when one door closes, don’t stand there waiting for it to reopen. Go find another door—or build one yourself.

The future of AI hasn’t been written yet, and that’s precisely what makes it so exciting. Don’t wait for someone to offer you a career in AI. You may have to invent it yourself.