There are two kinds of people using AI right now.
The first group opens a chatbot, types one question, gets an answer that’s just okay, and closes the tab. Tomorrow, they’ll do the exact same thing. They’ve been doing it for months and have almost nothing to show for it.
The second group is using the same technology to run something closer to a small, invisible company — one that works around the clock, handles research, writing, monitoring, and even negotiation, without a single additional employee on payroll.
The gap between these two groups isn’t about who has the better AI subscription. Everyone has access to roughly the same tools. The gap is about mindset — and according to Howie Liu, co-founder and CEO of Airtable, closing that gap can happen faster than most people expect.
Who is Howie Liu, and why should a freelancer in Bangkok, Jakarta, or Lisbon care what he thinks? Airtable is a no-code software platform — it lets non-programmers build custom databases and apps the way you’d build a spreadsheet — and it’s grown into a company with more than 700 employees serving over 500,000 organizations, including 80% of the Fortune 100. The company is now valued at around $12 billion. In other words, Liu isn’t a social media guru selling a course. He runs one of the more consequential software businesses of the last decade, and he’s now personally rebuilding his own workflow — and his company — around autonomous AI “agents.” He’s building a new platform called HyperAgent, designed to let anyone deploy AI agents into their team chats, email, and calendar without writing a single line of code.
In a recent interview on the podcast Silicon Valley Girl: AI, Tech and Career Growth, Liu described the exact shift that took him from casual chatbot user to someone running a fleet of digital employees — and he framed it as something almost anyone can replicate in under two weeks. For freelancers, digital nomads, and small business owners outside of Silicon Valley, that claim is worth taking seriously, because the economics of “hiring” an AI agent instead of a human are becoming impossible to ignore.
Stop Treating AI Like a Search Engine
The first shift Liu describes sounds almost too simple: most people are still using AI like Google with a chat window. Type a question, get an answer, feel mildly disappointed, repeat. That loop can go on for months without ever producing real results — because it’s built for looking things up, not for getting things done.
The technology itself has already moved past that stage. What’s now called an “AI agent” doesn’t just answer a single question and wait for your next instruction — it can take a goal, break it into steps, and carry out an entire task over minutes or hours with minimal supervision. Instead of re-explaining your preferences every single time you open a new chat, you teach an agent your standards once — your tone, your quality bar, your way of doing things — and it keeps applying that standard on its own.
This distinction matters enormously for anyone trying to earn money with AI rather than just be entertained by it. A chatbot can help you write one email faster. An agent can monitor your inbox, draft replies in your voice, flag the three messages that actually need your personal attention, and leave the rest handled — while you sleep.
From “Doing the Work” to “Managing a Team of Agents”
The second shift is a change in identity, not just tools. The most advanced AI users today aren’t running one assistant — they’re running several agents in parallel, the way a manager runs a small team. Liu has described setting up tasks for his agents before going to bed, so that by morning, work is already finished and waiting for review.
This is a genuinely new skill, and it maps almost exactly onto the leap someone makes when they go from being an individual contributor to managing a team of people. Some instincts carry over — clear instructions still matter, feedback still matters — but plenty is new, because what you’re now managing isn’t a group of employees with salaries and schedules. It’s a set of software agents that never sleep, never ask for a raise, and can be spun up or shut down instantly depending on the workload.
แนะนำ
For a solo entrepreneur, a freelance content creator, or a one-person e-commerce business, this reframing is the whole game. You stop asking “how do I do this task faster” and start asking “which of my tasks should I hand to an agent, and which decisions do I need to keep for myself.”
Fixing the “Bottleneck” Problem That Kills One-Person Businesses
Anyone who has run a content business, an online shop, or a freelance agency alone knows the specific pain of being the bottleneck. Every piece of content, every client message, every decision has to pass through you — and the moment you’re unavailable, everything downstream stalls.
Liu’s proposed fix is to build an agent that has absorbed your judgment: your feedback style, your standards, the kinds of edits you’d make. Once that agent understands how you think, it can give a “first opinion” on your behalf on the days you’re too busy to review everything personally.
This is precisely how one person today can realistically run SEO work, a weekly newsletter, and posting across multiple social platforms simultaneously — without adding headcount. The workload that used to require a small team gets absorbed by agents handling the repetitive 80%, while the human focuses on the judgment calls that still need a human.
It’s worth pausing on how directly this connects to real income data. A global survey of more than 4,360 freelancers by Freelancer.com found that 73% now use generative AI in their work, and Fiverr’s 2026 Freelancing Statistics report puts the figure even higher, at 76%, with 64% of AI users reporting a measurable productivity increase. Among the top tier of freelance talent, adoption runs above 80%, with most reporting that AI improved both their earning potential and productivity. The freelancers pulling ahead aren’t the ones who use AI the most casually — they’re the ones treating it like staff, not a search bar.
The “24-Hour Employee” You’d Never Actually Hire
Some jobs have simply never existed, because no business could justify paying a human salary for them. Liu’s example: someone who watches social media all day, every day, and only flags the handful of mentions that genuinely matter. Or someone whose entire job is continuously scanning used-car listings for one specific, oddly precise set of specifications.
These roles were never worth hiring for. They are now trivially cheap to run as agents operating 24 hours a day. Liu calls this the ability to “hire” a role that used to be a luxury nobody could afford — and it’s opening up entirely new categories of income-generating work that didn’t exist a few years ago.
This is where the numbers get genuinely interesting for anyone thinking about AI as an income source rather than a productivity toy. Upwork’s January 2026 freelancer earnings report found that AI-related work on the platform grew 60% year over year, with freelancers doing AI projects earning 44% more than the platform average — charging $75 to $200 an hour for work that simply didn’t exist a few years earlier. Broader AI agent development rates in 2026 typically range from $50 to $250-plus per hour, with full project pricing running from $5,000 to well over $150,000. Zoom out further, and the global gig economy hit $674 billion in 2026, with roughly 36% of Americans actively running some kind of side hustle. The market for “AI-agent-adjacent” income is not hypothetical — it already has real, measurable size.
แนะนำ
For readers based in or near Thailand, there’s a useful local reference point: Fastwork, a Bangkok-founded freelance marketplace and one of the largest of its kind in Southeast Asia. Founded in 2015, it’s now used by more than 700,000 businesses across the region for around 90 service categories, from graphic design to web development. It functions much like Upwork or Fiverr but with local payment rails and a Thai-first user base — freelancers list skills, clients hire directly through the platform, and payment is held in escrow until work is delivered. It isn’t perfect: users commonly cite its roughly 17% commission as steep compared with competitors. But it illustrates a broader point — the freelance infrastructure Southeast Asian workers already use is the same infrastructure now being reshaped by agent-based AI work.
There’s also a pure cost-arbitrage angle that matters if you’re a digital nomad or expat based in the region. AI and machine-learning freelancers based in South and Southeast Asia commonly charge $20 to $70 an hour, compared with $130 to $300-plus for senior specialists in North America or Western Europe — a gap that cuts both ways. It means global clients increasingly look to the region for AI-adjacent talent, and it means someone based in Chiang Mai or Ho Chi Minh City building agent-based services can undercut Silicon Valley rates while still earning income that goes much further locally.
Zoom out to markets like Vietnam and Indonesia, and the pattern holds. Both countries have seen the same explosion of AI-fluent freelancers competing on global platforms, drawn by the same math: the global agentic AI market was valued at $7.29 billion in 2025 and is projected to reach $139.19 billion by 2034, a 40.5% annual growth rate. That is not a niche trend. It’s one of the fastest-growing corners of the entire technology economy, and Southeast Asia’s large, young, English-capable freelance workforce is positioned to capture a meaningful share of it.
The Two Skills That Actually Separate People, Now That Everyone Has the Same Tools
If the tools are the same for everyone, what determines who wins? Liu points to two specific human skills.
The first is spotting the right problem to solve in the first place. He estimates this accounts for roughly 80% of success, because building the actual solution is no longer the hard part — AI has made execution far cheaper. The scarce skill isn’t “can you build it,” it’s “did you know it was worth building.”
The second is good judgment, applied at scale. The most valuable thing a person can do in an agent-filled world is exercise sound judgment across many decisions at once — essentially functioning like a CEO overseeing a dozen employees, except the employees are software. People who master this become the ones running the show, rather than the ones still typing single questions into a chat window.
A Simple Three-Step Way to Actually Start
None of this requires a technical background, and Liu is explicit that the biggest thing holding people back isn’t skill — it’s the pressure to get it right immediately.
Step one: sign up for an agent-capable AI tool with a low barrier to entry — something that doesn’t assume you can code.
Step two: find a small, low-stakes task in your own life or work to test it on, rather than jumping straight into your biggest business problem.
Step three, and most important: make it fun. The people Liu has seen get the most value out of agents are the ones genuinely enjoying the experimentation, not the ones demanding a payout from day one. It mirrors the early internet — people had to explore and play before they understood what it was actually good for.
One necessary caution belongs here, especially for anyone searching for “make money with AI” advice: the FTC has already filed multiple enforcement actions against AI “passive income” schemes, including one $25 million fraud case, and realistic expectations matter. Beginners typically earn $500 to $1,000 a month in their first six months — solid supplementary income, not the overnight riches promised in viral posts. The agent-based approach described here is a way to work more efficiently and take on more, not a shortcut to guaranteed income.
What This Means for You
Whether you’re freelancing from a co-working space in Bangkok, running a Shopify store from Bali, or juggling client work between time zones, the underlying lesson is the same: the return on AI doesn’t come from asking it better questions. It comes from handing it standing responsibilities and building a system you don’t have to babysit. Start with one task you personally dread — inbox triage, social monitoring, first-draft writing — and instead of doing it yourself one more time, spend an afternoon teaching an agent to do it your way. That single afternoon, repeated across a few tasks, is the actual difference between the two groups of AI users described at the start of this article. One of them is still stuck refreshing a chat window. The other is quietly building a business that runs while they sleep.
Key Takeaways
- Most people still use AI like a search engine — asking one question at a time — while top performers deploy autonomous “agents” that complete multi-step tasks independently.
- The real shift is psychological: moving from doing every task yourself to managing a small “team” of AI agents, similar to becoming a manager for the first time.
- AI-related freelance work already pays measurably more — Upwork reports AI freelancers earning 44% above the platform average, with specialist rates from $75 to $250+ per hour.
- Southeast Asia’s freelance economy, including platforms like Thailand’s Fastwork, sits at the center of a fast-growing global agentic AI market projected to hit $139 billion by 2034.
- The two skills that matter most now are spotting the right problem to solve and exercising good judgment across many parallel decisions — not technical ability.
Frequently Asked Questions
Q: Can you really make money with AI agents in 2026, or is this overhyped?
A: Yes, but realistically — most beginners earn $500 to $1,000 a month in their first six months, not the overnight windfalls promoted online. The gains come from efficiency and higher rates, not passive income.
Q: What is an “AI agent,” and how is it different from a chatbot like ChatGPT?
A: A chatbot answers one question at a time and waits for your next instruction. An agent can take a goal, break it into steps, and carry out a task over an extended period with little supervision, such as monitoring social media or managing an inbox.
Q: Who is Howie Liu, and why does his advice carry weight?
A: He’s the co-founder and CEO of Airtable, a no-code software company used by more than 80% of Fortune 100 companies, and he’s now building a platform specifically for deploying AI agents.
Q: Is this approach only useful for tech workers or programmers?
A: No. Liu specifically frames it as accessible to non-technical users, and the clearest gains often show up in ordinary freelance work like writing, design, virtual assistance, and social media management.
Q: What is Fastwork, and how does it relate to AI monetization in Thailand?
A: Fastwork is Thailand’s largest freelance marketplace, connecting freelancers with clients across categories like design, writing, and marketing. It’s a practical local starting point for Thailand-based freelancers looking to offer AI-enhanced services.
Q: How much can freelancers charge for AI-related work compared to standard freelancing?
A: AI-specialized freelance work commonly pays $75 to $250 or more per hour, well above typical freelance rates, according to recent industry earnings reports.
Q: Why do freelancers in Southeast Asia have an advantage right now?
A: Regional freelancers often charge $20 to $70 an hour for AI-related work, well below North American and European rates, making them attractive to global clients while still going further against local costs of living.
Q: What are the two skills that matter most in an AI-agent economy?
A: Identifying which problems are actually worth solving, and exercising sound judgment across many decisions at once — since execution itself has become cheap and widely accessible.
Q: How do I actually start using AI agents if I have no technical background?
A: Sign up for a beginner-friendly agent tool, test it on a small low-risk task first, and treat the process as experimentation rather than expecting immediate business results.
Q: Are there scams to watch out for in the “make money with AI” space?
A: Yes. Regulators have already taken action against fraudulent AI “passive income” schemes, including a $25 million case, so any offer promising guaranteed or effortless income deserves skepticism.
Q: Do I need to fire freelancers or employees to use AI agents effectively?
A: No — the goal described here is absorbing repetitive, low-judgment tasks into agents so a single person can handle more without adding headcount, not necessarily replacing existing staff.
Q: How big is the AI agent market expected to become?
A: The global agentic AI market was valued at roughly $7.29 billion in 2025 and is projected to grow to about $139 billion by 2034.
