Ready-to-use AI prompts built for tradies — quoting, scheduling, compliance and more.

Your plumber is sitting pretty. Your bookkeeper should be worried.

That’s not a hot take from some tech pessimist with a Substack. That’s the implication of a major economic modelling project published by Anthropic, the company that makes Claude. They built a full economic scenario explorer, modelled three futures for AI’s impact on jobs and wages out to 2030, and in every single scenario, knowledge workers face flat or falling wages. Tradespeople and workers who show up in person see wages rise.

The AI company, modelling the effects of their own AI, concluded that coders, accountants, bookkeepers, call centre staff and admin workers are the ones exposed. The sparky who won’t call you back for three weeks? Still in the driver’s seat.

What Anthropic Actually Found

The Economic Scenario Explorer is based on a peer-reviewed technical report, Economic Scenarios for significant AI (Korinek et al., 2026). It models three futures:

Modest scenario: GDP lifts 1.6% above baseline by 2030. Unemployment stays normal. Knowledge worker wages drift flat, other wages tick up slightly.
Substantial scenario: GDP up 8.3%. Most respondents in a 10,000+ person survey picked something close to this. Knowledge worker wages essentially flat; non-knowledge wages rise noticeably.
Extreme scenario: GDP up 32.4%, driven by recursive AI self-improvement and rapid adoption. Knowledge worker wages fall more than 10% by 2030. Unemployment spikes to historic levels.

Three scenarios, three different assumptions about AI capability and adoption speed, one consistent result: cognitive work gets cheaper and physical work gets more expensive.

The survey of 10,000+ Americans found the typical respondent expects something close to the substantial scenario, with GDP about 10% higher by 2030 and unemployment around 5%. About one in ten expect the extreme scenario. Almost nobody expects nothing to change.

The report names the jobs at risk directly: coders, call centre agents, bookkeepers, accountants, admin roles. It also names the jobs that benefit: electricians, nurses, and other roles where you need to show up in person and do something with your hands or your body.

Australia Makes This More Extreme

The model is US-based. The economic mechanics apply here, but Australia has a layer on top that makes the divergence sharper.

We already can’t get tradies. According to Jobs and Skills Australia’s 2025-26 data, Technicians and Trades Workers make up 51% of all persistent occupational shortages in the country. The government’s own 2026-27 federal budget allocated $75.1 million to a new trade skills assessment system, partly to fast-track overseas-trained tradies into Australian licences, because the shortage is bad enough to threaten the 1.2 million homes promised by 2029. Industry groups estimate we need around 116,700 additional construction workers just to hit that target.

AI does not fix a labour shortage in physical trades. You cannot remotely patch a switchboard. You cannot AI-generate a licensed plumber. Demand for tradies is driven by housing construction, infrastructure spending and an ageing stock of homes. All of those pressures are increasing. The number of qualified Australians who can do the work is not growing fast enough to meet them.

Meanwhile, Australia’s white-collar workforce is exactly the population the Anthropic model flags. Bookkeepers, BAS agents, compliance staff, call centre workers, admin coordinators. These are common roles in SMBs across the country, often filled because the alternatives were too expensive or complex. The alternatives are getting cheaper and more capable every quarter.

What This Means for Your Business

Three things to act on now:

1. Stop waiting to automate administrative tasks. If you have staff doing data entry, invoice processing, scheduling, or routine customer enquiries, the tools to automate those tasks exist today. The substantial scenario is already priced in. Waiting to see how it plays out means your competitors who moved earlier have a cost advantage you’ll struggle to close.

2. If you employ knowledge workers, have the upskilling conversation now. The risk isn’t mass layoffs tomorrow. The risk is a gradual shift in what those roles look like, over the next two to four years, with wages that don’t keep pace with inflation. Staff who learn to work alongside AI tools stay valuable. Staff doing 2023 workflows in 2028 are exposed. That conversation is your responsibility as the employer.

3. If your business depends on tradies, get your lead times sorted and lock in relationships. Electricians who trust you, builders you’ve worked with before, plumbers who pick up when you call. The shortage is structural. The government is spending money on the problem and acknowledging it won’t be solved quickly. Relationships are the only hedge you have against it.

Anthropic built a model to show where AI creates value and where it displaces it. They published it openly. Their own finding is that the tradie wins and the accountant gets squeezed. For Australian SMB owners, the structural conditions that make this true, chronic tradie shortage on one side and automatable white-collar tasks on the other, are already in place. The model just puts numbers on what most business owners can already feel.

More hot takes: Hot Takes →

📬 The SmallBizAI Brief

One practical AI tip for Australian small business. Every Tuesday. Free.

Join business owners getting smarter about AI every week.

Subscribe free →