AI Jobs Forecast: Why Freelance and Agency Contracts Could Shrink

I expect one of the next year's employment pressures to arrive through assignments that clients stop commissioning. A business could keep its own workforce unchanged while buying less writing, design or other project work. The damage would land at a freelancer or supplier, potentially without a layoff announcement from the company changing how the work gets done.

My forecast for January, April and October 2027 is that AI will encourage more of these decisions, with uneven consequences across services. I am more confident about experimentation than widespread displacement. The evidence supports a mechanism worth watching; it cannot tell us how many contracts will disappear.

Three months: small assignments come under pressure

In its September 16 research summary, OpenAI reported next-month reuse rates of 44% for advertising or promotional writing and 37% for creating marketing materials when those activities fell outside users' usual occupations. These are company findings from a ChatGPT sample. [1]

The study tracks tool use. It does not establish whether those tasks were previously outsourced, whether the output was commercially adequate, or whether suppliers lost business. [1] My inference is that repeated use makes it plausible for some buyers to reconsider small external assignments.

By January 2027, I expect more clients to test producing routine material themselves before commissioning it. Imagine a business trying its own promotional drafts for a modest campaign. That hypothetical trial could remove one paid brief without requiring the client to hire a writer or decide that professional writing has become unnecessary.

For freelancers, I would distinguish a brief lost to another supplier from a brief no longer being commissioned. Those explanations point toward different responses. Improving a pitch has limited value when the buyer has stopped buying that particular service.

Six months: experiments reach contract renewals

By April 2027, my prediction is that some successful experiments will start changing the scope of recurring contracts. A client might still want external campaign planning or research while requesting fewer finished assets. The relationship survives, but the amount of paid work contracts.

Consider a hypothetical agency whose customers each reduce a small recurring assignment. No single customer has eliminated a position. Together, those decisions could leave the agency with insufficient work for its existing team. A recruitment freeze or job cuts would appear one business further along the chain.

That possibility is why I would take shrinking repeat commissions seriously as a career signal. Waiting for a dramatic announcement could mean overlooking a deterioration already visible in the work arriving each month.

I would also resist assuming that moving into a client's company solves the problem. If an existing employee absorbs the task, there may be no vacancy to pursue. For someone searching, the relevant opportunity would be a role with continuing responsibility for the activity, where the employer is actually recruiting.

Twelve months: reasons to keep buying expertise

By October 2027, I expect the division between internal and purchased work to depend increasingly on what suppliers possess that clients cannot readily assemble. Access to useful information, established systems and specialized knowledge could matter alongside the ability to produce a deliverable.

Loop CEO Hannah Bravo's recent announcement offers one possible example. She describes models drawing on patterns across its merchant network and emphasizes partnerships and integrations. These are management's claims about the company's offering; the announcement does not independently establish its performance or customers' results. [2]

A client gaining a drafting tool does not automatically gain a supplier's pooled data or connections to other systems. I therefore expect some outside services to retain a strong reason to exist. For career planning, I would examine what a prospective employer brings that its customers struggle to obtain alone.

I regard those advantages as conditional. Information may become easier to obtain, and integrations easier to build. When choosing a specialization, I would favor learning why customers depend on a service over assuming today's technical difficulty will endure.

The supplier's own jobs remain exposed

Employment at a company selling technology carries its own uncertainty. Vermont Public reported on October 21 that Vermont Information Processing was cutting 66 jobs, just over a month after acquiring Encompass Technologies. The software business serving beverage companies was promoting AI applications, while its spokesperson cited an organizational review and changing business priorities. [3]

That reporting does not establish that customers brought work inside or that AI replaced the departing employees. [3] Its relevance is narrower: a supplier's ambitious product direction can coexist with job reductions. I would not present joining an AI vendor as an automatic escape from pressure elsewhere.

Nor would I dismiss every supplier vacancy. A business can cut some activities and need people for others. I would judge the particular role by the work customers continue to need, without assuming the sector's label provides security.

What would change my forecast

The positive alternative is substantial: cheaper production could encourage clients to attempt projects they previously abandoned, creating additional assignments for suppliers. Providers could also use AI to offer more useful work themselves. If repeat commissions and contract scope expand alongside client adoption, my prediction of a squeeze would weaken.

Conversely, evidence from actual buyers that they are absorbing previously purchased work would strengthen it. Today's usage study and company announcements do not establish that aggregate outcome. I would revise this forecast against purchasing behavior, rather than treating each new model release as confirmation.

My concern is that a company can describe its AI adoption as helping employees while the employment costs arrive outside its payroll. For the person whose regular assignments disappear, that boundary offers little comfort. The work can continue while their place in it disappears.