A year ago, most of us feared that AI would shrink our copywriting team. In September 2026, in the most surprising twist of my year, we met to plan how our copywriters will handle all the new work coming in, and the conversation made it clear that we need to hire more writers. The work comes from a new service line that exists because of AI, and from the way we chose to do our AI transformation.
The fear that AI would take our jobs
We have been very forward-thinking about AI as a company. We have had a CEO-driven mandate to make AI how we work and to make sure everyone understands that our future depends on using it right. Every employee has had an AI license since 2024, and getting one requires passing a training and an assessment, so that everyone knows how to use it well. We have a dedicated AI department, a robust AI policy and more. But one subject stayed delicate and mostly unspoken across the whole workforce: is AI going to take our jobs away? In this article, I will focus on the copywriters.
As soon as LLMs became accessible, content factories started to appear, turning a style guide and a direction into as much copy as a company wanted: on topic, but with little substance and nothing truly differentiating. At the same time, AI was beginning to let marketing teams write much of their own copy in-house. In one example from May 2024, Klarna said that it was using AI for 80% of all its copywriting and had cut its spending on outside marketing suppliers by 25%. Sam Altman went further, telling the authors of Our AI Journey that "95% of what marketers use agencies, strategists, and creative professionals for today will easily, nearly instantly and at almost no cost be handled by the AI." He was describing where AI is heading, but for people who write for a living it sounded like a clear warning.
Our own copywriters had not lost any work. Using AI skillfully, they were producing more, and they could put far more research behind everything they wrote. But I expected the pressure to keep growing. What I did not foresee was a conversation, a year later, that went 180 degrees from where I thought we would be.
The change that moved the work back to the writers
Buyers now start their research by asking AI, and that shift created a new kind of work. Generative Engine Optimization (GEO), also called AI Engine Optimization or Answer Engine Optimization (AEO), is the work of getting your company cited and recommended when people ask AI tools questions. Forrester's 2025 Buyers' Journey Survey found that 94% of business buyers use AI in their buying process, and twice as many as the year before named generative AI as their most meaningful source of information.
There are many GEO products on the market, and most of them do the same thing. They track a list of prompts, tell you how often you are cited and compare that with your competitors. We measure those things too. But a tracker tells you where you stand, not what to change.
B2B, and especially technical and regulated industries where we specialize, needs something else. The buyer is a small group of specialists asking very specific questions: a lab director comparing platforms, a compliance lead checking a claim, an analyst choosing a measurement method. You don’t win them by scanning the whole internet. You win them by answering their exact questions better than anyone else, with evidence they can check. That is the gap we built our ROAI GEO Framework to fill, and in 2026 it became the fastest-growing service line at RAB2B.
Why we need more copywriters now
Getting recommended by AI takes work in three places. The first is technical: clean code, structured data and pages that AI crawlers can read and understand. Our development teams have moved fully to agentic development and keep very tidy coding practices, so optimizing a website's code has become a clear set of steps they follow, and they do it very well.
The second is outside the website: other trusted sources citing and recommending you, which takes outreach, earned media and partnerships. AI tools weigh what others say about a company heavily, so this is a large part of the work and we run it as its own program.
The third is content, and it is the part most GEO tools leave out. In our work, it is about 75% of what moves a client's visibility. The research points the same way. In the 2023 paper that introduced the term GEO, researchers from Princeton and other institutions found that adding evidence to a page to support your claims raised its visibility in AI answers by up to 40%.
The page has to work for the human buyer and for the AI at the same time. Winning takes writing with specific qualities, and there is no shortcut to any of them. AI tools recommend the page that best answers what the buyer is asking, so the page has to be that answer. That is where a skilled writer makes a difference.
Our ROAI GEO Framework uses AI to measure how often each client is cited compared with its competitors, question by question, and to show which pages are missing, which are too thin and which need proof. Each of those findings becomes a piece of writing, and a single website can have hundreds of pages that need it. That volume of skilled writing is why we will need to hire more writers.
AI is creating work we did not have before
In this case, it is copy planning, a new responsibility in our copywriting team.
Our GEO system runs a thorough assessment of the market, the client’s gaps and the opportunities to win, and gives precise directions on what to do about each one. It then turns every improvement into a brief and a task in Jira, so the work goes straight from the assessment to the team that does it.
In copy planning, a copywriter reviews those briefs, checks the research behind them and makes the critical calls on what to apply, what to refine and what to leave out before sending them to the writers. I picture it like scoring in basketball. The system makes the assist, setting the writer up with the decisions to make and the information to collect. But it is the writer who has to take the shot and score.
Most recommendations need some changes before they work, because the writer knows the client, the audience and the brand in ways the system does not.
What the results look like
One of our clients this year was a B2B analytics company that we tracked against its three direct competitors. When we focus on the questions buyers ask without naming any brand, as they research at the top, middle and bottom of the funnel, we are looking at the hardest territory to win. When we started, our client was in second place there with 28% of the citations among the four companies. Four months later it is first with 57.5%, and the other three are at 22.7%, 14.9% and 4.9%. Most of that work was writing.
Why the work keeps coming
GEO does not end. AI answers change, competitors publish, industries move and articles that were right last year go out of date. To keep being cited and recommended, a company has to keep its content current, refresh what has aged and keep sharpening what makes it different. Many of our clients have entire websites that need this work, and each page has a measurable job and a number that shows whether it is doing it. That is why our copywriting team has more work now than it had a year ago.
What this says about AI and jobs
I am still worried about AI taking jobs, and I think many leaders underestimate that risk. Paul Roetzer of SmarterX said it plainly this year: "I don't understand the premise that a responsible leader of any organization wouldn't plan for the alternative: That maybe there is this chance that we do go through a one to three to five year period where we're just going to need fewer people." Most people see the risk, just not for themselves. In SmarterX's 2026 State of AI for Business report, a survey of more than 2,100 professionals, 71% said AI will eliminate more jobs than it creates over the next three years, but only 20% were worried about their own role.
What this year showed me is that the outcome depends on how the transformation is done. Earlier this year I wrote that every professional now has to keep asking what AI does better than they do and where their value moves next. Our copywriters are one answer to that question. A content factory uses AI to produce the same content for less money. We used AI to find where the work that matters exists, and that work still needed our writers, with more skill than before. We are seeing the same pattern in other departments, and I will write about those in the coming months.
A year ago, the question nobody dared to ask out loud was whether we would still need our writers. Now the question is whether we have enough of them. What made the difference was not the technology itself but what we decided to use it for.
Frequently asked questions
What is GEO?
Generative Engine Optimization (GEO) is the work of getting a company cited and recommended in answers from AI tools such as ChatGPT, Google AI Overviews, Gemini, Claude, Copilot and Bing. It is also called AEO, for AI Engine Optimization or Answer Engine Optimization, and the result it aims for is often called AI visibility.
How is GEO different from SEO?
SEO gets your pages indexed and ranked in search results. GEO gets your company cited and recommended inside AI answers. GEO builds on SEO: the technical foundation still matters, but content that answers buyers' exact questions with evidence is what earns the citation and the recommendation.
How is GEO different for B2B and B2C?
In B2C, a buyer's choice is shaped by signals from all over the internet, such as reviews, social posts, forums and retailers, so GEO has to cover all of them. In B2B, the buyer is a small group of specialists with specific questions. The work is identifying the right buyer personas, learning what they ask at each stage of their research and making sure your content gives them the best answer.
How do you measure AI visibility?
Most tools track a generic list of prompts. We build ours from the buyer: the ideal customer and their personas, the client's and competitors' market, and the words buyers actually search with. With that, we engineer a controlled study that scans the client's market and shows how the client is positioned to win the opportunities and beat its competitors. Every month we measure how often AI tools cite and recommend the client against its named competitors.
How long does GEO take to show results?
Technical fixes and ungated content can show movement in the first months. Citation share takes time to build, because AI systems re-evaluate sources gradually. In the example above, the client's share of citations among its competitors went from 28% to 57.5% in four months. How long it takes depends on the size of the website and how old its code base is, and for large or older sites it is more likely to take six to twelve months.
Will AI replace copywriters?
Not in our experience. AI has taken over a lot of high-volume copy with little substance, and it assists with much of the research behind our writing. But AI search recommends the content that best answers a specialist buyer's exact question, and writing that takes a skilled copywriter. Since we started our GEO work, we have needed more copywriters, not fewer.
What new roles and responsibilities has AI created at RAB2B?
In our copywriting team, GEO added copy planning as a new responsibility: a writer reviews the recommendations from our GEO system, checks the research behind them and decides what to apply, refine or leave out before the writing starts. Our AI department also has three roles that did not exist before:
- An AI Systems Architect decides how the automations in each solution are built and where human specialists come in, so the whole workflow runs as efficiently as possible.
- An AI Service Orchestrator runs an AI service such as GEO from start to finish and makes the judgment call at each checkpoint.
- An AI Solutions Engineer turns the department's frameworks and methods into working software by directing AI development tools and keeps it running.
— Yas Dalkilic Head of AI, RAB2B