What the market pays — VA work versus AI-certified work

The floor — what replaceable work pays

Start with the lower of the two pay figures JobTayo publishes: task-based work that AI now absorbs pays around $4–7 per hour. It is the more useful of the two, because it describes what is happening now rather than what is possible.

Notice what it describes. It is a price for a kind of work, not a verdict on a kind of worker. Plenty of careful, fast, reliable people are paid at that level, not because they are careless but because the job they were hired to do can now be done another way. Data entry and encoding, copy-pasting between systems, formatting and sorting, basic writing and rewriting, simple research and summarising: the Academy lists these as things AI already does well. When a task can be fully described in a message and checked by glancing at the output, its price drifts towards the cheapest reliable way of producing it, and that is no longer a person working alone.

What the gap is actually measuring

The gap between the two columns is easy to misread as a promise. It is not a price list, and it is not a before-and-after of a person. It is a description of the distance between two kinds of work, which can happen inside the same business, on the same afternoon, for the same client.

At one end, someone is handed a described task and completes it. At the other, someone looks at how the business actually runs, decides what should be built, builds it, and stands behind it when it goes wrong on a Tuesday. Neither number is a rate anyone is owed. The point is the gap itself: that it exists, that it is made of specific things, and that those things can be learned rather than wished for.

Three things do most of the work in setting where a person sits in that gap. None of them is "know more tools".

Three things that actually move a rate

What you are accountable for

The plainest version of the difference: I did what you asked is a different product from this works, and I will fix it when it stops working.

Accountability is paid for because it moves risk off the client's desk. An owner who hands over a task still owns the problem. An owner who hands over an outcome does not. That transfer is the thing being bought, and it is why two people can sit in the same call, use the same software, and be priced differently.

Look at what the Academy says still needs a person: seeing what is broken inside a business, building the system that fixes it, deciding who is allowed to see what, and checking AI's work before a client sees it. Every one of those is an accountability, not a task. None of them can be written as a ticket and handed to a stranger with no context.

Made concrete: "tidy up the enquiries spreadsheet each morning" is a task. "Enquiries are captured, nothing is lost, each sales person sees only their own region, and someone tells me when the flow breaks" is an accountability. The second is harder, slower to earn, and priced on a different basis.

Whether the client can replace you cheaply

Rates track replaceability more closely than they track skill, which is why hard work alone does not shift them. The useful question is an uncomfortable one: if you disappeared on Friday, what would it cost this client to be back to normal by Monday?

For described task work, the answer is usually one job post and a short handover. For someone holding a working system, replacement means rebuilding context: which fields exist and why, who approves what, what the access rules are, what broke last time and how it was fixed. That is real knowledge about a real business, and it takes time to rebuild.

Whether you can price an outcome instead of an hour

The hour has a design flaw in AI work. If a tool halves the time a job takes, hourly billing halves the income from that job. You get faster and are paid less for the same result, and the client gains nothing by wanting you faster. Every efficiency you learn works against you.

Pricing an outcome removes that trap, but it is not a billing trick you can adopt on Monday morning. It needs a defined result, a shared definition of "done", and someone willing to carry the risk when an estimate turns out to be wrong. In other words it needs the first two items on this list already in place. Our guide on how to price AI work without guessing covers the mechanics.

What nobody can promise you

This is the most important paragraph on the page, so it is written without decoration. No course guarantees income. JobTayo's Terms say it plainly: we do not guarantee income, rate increases, employment or clients. What any individual earns depends on their own effort, their own skill and the market conditions they are working in, none of which a training programme controls. Anyone who tells you a specific figure you will be earning by a specific month is either guessing or selling.

The same honesty applies to everything around it. The subscription is ₱999 a month and payments are not refunded. Foundation quizzes are scored automatically against a fixed answer key; a specialisation assessment is a live call with a person, billed separately. That exam costs ₱1,000, and none can be booked yet, because the specialisations are still being built. Passing is never guaranteed, and the two fees buy different things: the subscription buys the teaching, the exam fee buys the attempt. Neither buys the result. Graduates who pass a specialisation assessment may enter the JobTayo talent pool, which is not an offer of work and not a promise that any employer will make contact.

Working out where you currently sit

You cannot forecast your own rate. You can work out which side of the gap your current work sits on, which is more useful anyway. Take the last month of real work and ask three questions about each recurring piece of it.

1. Could this have been fully described in a message? If a client could write the instruction, and a stranger with no knowledge of the business could follow it, it is task work. Task work is where the pressure is.
2. Would the client notice if a tool did it instead of me? Answer honestly. If the output would be roughly the same, the work is exposed regardless of how well you do it.
3. When it goes wrong, who is expected to fix it? If the answer is someone else, you are being paid for hands. If the answer is you, you are being paid for judgement, and judgement sits in a different market.

Most people find their week is a mix, which is normal and is not a failure. The direction of travel matters more than today's ratio: which parts are growing, and which parts are being quietly absorbed. Our guide on which remote work is growing and which is shrinking looks at where that movement is happening.

Why this takes longer than a course

One honest caution about timing. The three things above are earned in a specific order, and the order cannot be skipped. You take responsibility for something small. If it holds up, you may be trusted with something larger. Some people build a record of systems that did not fall over, and for them the pricing conversation eventually changes — not because a new rate was announced, but because what is being sold has stopped being the same product. Nobody can tell you whether or when that happens for you.

Training can compress the technical half of that. It cannot compress the trust half, and nobody should tell you otherwise. What a course can reasonably do is teach the technical half, so that if a chance to own something comes up, you have the skills to build it, secure it and stand behind it. The abilities that carry that weight are set out in the skills that outlast the tools.

Where to go next

• How to price AI work without guessing — quoting for an outcome rather than an hour, without inventing a number.
• Which remote work is growing and which is shrinking — where demand is actually moving.
• The skills that outlast the tools — what stays valuable when the software changes again.
• The JobTayo Academy curriculum — what the three foundation courses actually cover.