Is AI content working? What data says about marketing, search and audiences
Almost every marketing team uses AI to produce content, but only a minority sees better results. The deeper change is in how people discover brands.
In 2026 almost every marketing team uses AI to produce content, but only a minority sees better results. The deeper change is outside production: in how people discover brands and in what they find credible.
Everyone is using AI, but few are taking value from it
AI is now a standard marketing tool, but few organizations are seeing real return. A large part of marketing departments and organizations are already using some form of AI, but, of those, 68% are still at the exploratory or developing stage.
Only 30% of CMOs believe their processes are mature enough to scale AI.
Sources: Salesforce, State of Marketing, 10th ed., Gartner, CMI/MarketingProfs
We’re typing faster, not necessarily better
AI speeds up writing, but the gain shrinks as you move from productivity to quality and results. Text creation is the dominant use: 89% of B2B marketers with AI use it to generate or optimize copy, and 53% for images and video.
Among those using AI to create content, productivity and operational efficiency are the dimensions that improved most with AI usage, while 12% say content quality decreased.
Academic research points the same way, with one important nuance: stories written with AI ideas were rated as more creative, but ended up more similar to each other (Doshi and Hauser, Science Advances, 2024).
If everyone uses the same tools the same way, content converges and differentiation becomes the scarce resource.
Half of new articles are already AI-generated
About half of the articles published on the English-language web are now mostly AI-generated, but that content barely reaches people.
Twelve months after ChatGPT was launched, AI articles were 39% of the total. By November 2024 they had overtaken human-written articles, and the share has plateaued at about 50% since. However, studies show these articles rarely appear in Google or ChatGPT.
Google’s position explains part of this. Since 2023 the company says it rewards quality, not the production method. In March 2024 it began treating content produced at scale to manipulate rankings as spam. Earlier this year, the company reinforced the importance of creating non-commodity content that is helpful and people-first in order to perform better.
The result for brands is a paradox. Producing became almost free, so volume stopped being an advantage and became noise.
Fewer clicks, more answers
The biggest change for brands is in distribution: AI answers keep the user in place and cut traffic to websites. Content is still read, but increasingly by machines that summarize it.
People are getting the answers they need without ever leaving the search page. The irony is that the websites powering those AI summaries are getting less of the traffic. When Google shows an AI summary, only 8% of visits result in a click on a search result, and just 1% result in a click on a cited source.
Sources: Pew Research Center
Generative AI is already the number one source B2B buyers run to when researching suppliers, according to Forrester. More than 90% use or plan to use generative AI in purchase decisions.
However, the interesting thing is that these same buyers say they do not trust AI on its own. They claim that before making a purchase, they validate information with peers, experts, and the suppliers themselves.
The challenge is that these audiences are becoming harder to reach outside AI-generated summaries. Forrester estimates that B2B companies lost between 10% and 40% of their website traffic over the past year. Buyers who search with AI are one-tenth as likely to click through to a website.
The practical consequence is a new discipline, generative engine optimization (GEO or AEO), with the goal shifting from winning the click to being the source the answer cites.
Audiences penalize what feels artificial
Most consumers want to know when content is made by AI, and a meaningful share trusts the brand less on finding out. The reaction varies with age and with the brand’s category.
Consumer neuroscience reinforces the pattern. In a NielsenIQ study with more than 2,000 participants, AI ads were identified intuitively and seen as more boring and confusing. Even the high-quality ones triggered weaker memory responses than traditional ads.
There is a generational effect, though. In the Australian YouGov survey, Gen Z and millennials are more open, and Gen X and baby boomers more penalizing.
The Coca-Cola case
Coca-Cola released AI-generated holiday ads in 2024 and did so again in 2025, despite the criticism. The social media reaction centred on the loss of human warmth and the impact on creative work.
Coca-Cola’s head of generative AI, Pratik Thakar, said human storytellers had guided the direction to keep the story authentic and emotionally resonant, but the comments said otherwise. Viewers called the films forgettable and soulless, and even those who conceded the visuals had improved said they missed ads that made them tear up. Artists and animators argued the campaign devalued creative labour and the tradition of holiday storytelling.
How brands are responding
Brands are investing in AI and, at the same time, in the channels and formats AI cannot replicate:
- Physical presence and owned channels. Events rose to 23.1% of CMOs’ offline budget, from 19.4% a year earlier (Gartner, via Chief Marketer).
- Opinion with a face. 96% of B2B brands produce thought leadership, but in 37% fewer than 5% of in-house experts contribute. LinkedIn, newsletters and events are the most effective channels.
- Being cited, not just found. News publishers plan to put less effort into traditional Google search and more into YouTube and original reporting.
- Fundamentals before tools. The most effective teams credit content relevance and quality (65%) and team skills (53%), ahead of technology (43%) (CMI).
Work is changing more than disappearing
AI has not yet reduced the weight of people in marketing budgets, but it has already lowered demand for task-based writing and design work. The effect is uneven between in-house teams and freelancers.
Only 14% of organizations using AI say it reduced headcount in the past year. A year earlier, 32% expected that reduction (McKinsey). Among content and marketing professionals, layoffs rose 30% and workloads grew for those who stayed (CMI, Career and Salary Outlook 2026).
Job posts for writing and coding on a major platform fell 21% in the eight months after ChatGPT. Image-creation posts fell 17% (Demirci, Hannane and Zhu, Management Science, 2025). The jobs that remained are more complex and better paid, according to the same study.
Get The Story Sells in your inbox every week.
Subscribe on Substack