When AI Ate Their Writing Jobs, These Two Did Opposite Things

In early 2022, James Presbitero Jr. refreshed his inbox for the third time that morning, already knowing what he'd find. A client he'd written for over several months sent a brief message: they were "exploring AI solutions for content." Another project that was "definitely happening next quarter" had gone quiet. Within months, roughly 90% of his freelance writing income had disappeared.

That same period, a social media copywriter — she asked to remain anonymous — lost six years of steady work when the company she contracted through was sold. The new owners let all the freelancers go without notice. AI templates replaced them. She searched for comparable work for months and found nothing. Her business eventually closed. To pay her bills, she turned to online sex work.

Same disruption. Same timeline. Different outcomes.

What made the difference wasn't talent, work ethic, or timing. It came down to three specific decisions — and the first one most people get backwards.

The Market Is Growing and Collapsing at the Same Time

Before getting to those decisions, it helps to understand the market these writers were navigating — because it looks completely different depending on which layer you're measuring.

The global content writing services market is projected to reach $38.6 billion by 2033, growing at nearly 7% annually. That's genuine growth, not industry wishful thinking. But here's what that number doesn't tell you: 78% of marketing teams already use AI tools internally. Which means "I'll offer AI writing" is not a pitch. It's a description of what your potential client does at 9 a.m. every morning with their existing ChatGPT subscription.

The money is moving, but it's moving upward — to the strategy and accountability layer above the model, not to faster word counts. And the buyer-side context makes sales harder than it looks. A 2025 MIT study found that 95% of enterprise generative AI pilots return zero measurable value. Every sales call you make lands in front of someone who has already heard AI promises and watched them fail. That's the headwind every new agency will face.

This isn't a copywriter-specific problem. If you write anything professionally — internal communications, product descriptions, customer-facing content — your employer or your clients are making the same calculation Becky's contracting company made: is a human doing this worth the cost? The agencies that survive are the ones that changed what the human is doing, not how fast they're doing it.

What the Transition Actually Requires

Presbitero didn't survive by writing faster than the AI. He survived by stopping selling sentences and starting to sell systems. After his inbox went quiet, he rebuilt his offer around AI-powered content systems for founders — not "here are your blog posts" but "here is a repeatable workflow that turns your expertise into audience-ready content at scale." His positioning shifted from freelance writer to AI marketing specialist. The output changed. So did who was paying, and what they were paying for.

That distinction matters because it names the skill that actually transferred — and the skills that didn't. Konrad Sanders, CEO of The Creative Copywriter, whose clients include Adidas, TikTok, and Mercedes-Benz, put it plainly: "The best way to use AI for copywriting is by being a really good copywriter. Otherwise, garbage in, garbage out." The model amplifies the operator's judgment. It doesn't replace the need for it.

The best way to use AI for copywriting is by being a really good copywriter. Otherwise garbage in garbage out.
— Konrad Sanders, CEO of The Creative Copywriter

Run through these two lists honestly before you read the next section.

What transfers directly: editorial judgment (knowing when the AI draft is wrong, vague, or off-brand); brand voice encoding (capturing and systematizing a client's distinct tone); domain expertise that the model can't reliably verify; customer empathy; and the ability to extract insight from a 30-minute client interview.

What doesn't transfer — and will surprise you: client acquisition, which is the hardest new skill for anyone coming out of employment; scope definition, because agencies live or die on what's included and employees rarely set scope; managing AI prompts consistently across five or more client voices simultaneously; pricing conversations that don't automatically undercharge; and knowing your actual cost per deliverable.

The transfer list applies regardless of your writing background. A former HR communications manager has brand-voice and empathy skills. A customer-success writer has interview skills and domain knowledge. A technical writer has the editorial discipline to catch model errors. What nearly everyone lacks at transition is the bottom list — and that's the list most people don't discover until month three.

The Honest Arithmetic

The transition is financially survivable. But the math is specific, and the timeline is longer than most YouTube content suggests.

The single most predictive variable is how much cash you have before the first client appears. Rachel Leslie, a copywriter who left corporate employment to run her own practice, describes a six-month personal-expense safety net as the documented minimum — not a luxury, the floor. Counter-cases are instructive: one copywriter who quit with roughly $3,000 in savings describes being broke within 60 days. A marketer who quit on one month of runway reported severe financial fragility throughout year one.

Year-one income for most transitions falls in the $30,000 to $90,000 range. The $100,000 to $300,000-plus figures that circulate online are real, but they're documented primarily for the top quartile — and only when the founder carried retainer clients from a pre-existing side hustle or hit recurring revenue by month four. The 34% hourly rate premium that Upwork documents for AI-skilled freelancers is also real. It just arrives in year two, not year one.

Time to first paying client runs 30 to 90 days for disciplined operators with a clear niche and direct outreach. The first month is almost always zero dollars collected, not zero dollars billed. Build your runway calculation around personal expenses times six. That number — not your website, not your tool stack, not your niche — is the first problem to solve.

Three Structural Traps That Kill the Transition

Here's where Becky's story becomes more than a cautionary tale about hard luck. Her work didn't disappear because she was bad at copywriting — clients praised her consistently. It disappeared because an intermediary owned the client relationships. When that intermediary changed its model to AI-generated templates and self-serve tools, Becky had no direct line to any buyer. Six years of excellent work produced zero transferable client equity. Owning the delivery is not the same as owning the relationship.

Paige Schwartz, who founded Copytree in 2022, took a different path. She spent five years as a product manager at Google before leaving to build a freelance technical-writing practice — direct clients, direct relationships, a validated niche — before ever launching an agency. She didn't quit employment and immediately start scaling. The bridge stage was the protection.

Writing was my favorite part of the job, and eventually I left Google to do more of it.
— Paige Schwartz, CEO and Editor-in-Chief of Copytree

Three failure modes account for most of the transitions that quietly collapse before anyone notices. First, platform or intermediary dependency: you own the delivery, but a marketplace or agency owns the client. When they pivot to AI, you have nothing. Second, selling commodity output: "I'll write blog posts with AI" is indistinguishable from what the buyer's internal team already does, so you compete on price and lose. Third, skipping the bridge: quitting employment before validating even one direct retainer client means the sales learning curve and the financial pressure arrive simultaneously.

The diagnostic question is simple. If the platform, agency, or employer you currently work through changed its model tomorrow, how many individual buyers would call you directly? That number is your actual client equity — not your portfolio quality, not your years of experience.

This exposure isn't limited to copywriters. Customer-success writers embedded inside SaaS tools, content managers whose only distribution runs through a company CMS, HR communicators who have never managed an external vendor relationship — all share the same structural risk. Direct client relationships are the moat. Everything else is infrastructure you don't own.

What Presbitero Actually Built — and Three Things to Do This Week

Presbitero's inbox is no longer quiet. But what he rebuilt doesn't look like a traditional copywriting agency. He builds content systems — intake processes, voice documentation, editorial workflows, prompt libraries — that let founders produce consistent, recognizable content without starting from a blank page every time. The writing skill is in there. It's just no longer the product. The product is the system the writing runs on.

That distinction is the whole transition, compressed into one sentence.

The move from employed writer to AI-powered agency owner isn't a leap of faith. It's a series of small, reversible experiments — each one giving you better information than the one before. The founder who succeeds is almost always the one who ran the experiments while still employed, not after quitting.

Three exercises to run this week, before you update your LinkedIn, pick a tool, or set a quit date.

The relationship audit: List every client, employer, or platform that currently sends you writing work. For each one, ask whether an individual buyer would call you directly if that entity changed its model tomorrow. Count the yeses. That number is your actual client equity.

The runway calculation: Multiply your monthly personal expenses — housing, food, insurance, debt minimums — by six. Write down that number. If it doesn't exist in liquid savings, that's the first problem to solve, not your niche and not your tools.

The niche specificity test: Complete this sentence in one clause — "I help [specific type of organization] produce [specific type of content] that [specific measurable outcome]." If you need more than one clause, the niche isn't specific enough yet. Vague positioning is what makes buyers say they'll just use ChatGPT.

The model took the sentences. It didn't take the judgment, the relationship, or the system. Those are still yours to build — if you start before the inbox goes quiet.

Recommended Tools & Resources