Too Much Content with No Impact

Writing copy, adding images, creating different versions, optimizing for each channel: What used to take years—and drain time, budget, and nerves—can now be done quickly thanks to generative AI. It’s not always perfect, but it’s reliable. But what happens when everyone can produce more at the same time?

Illustration: KI generiert
Illustration: Generated by AI

Last year, AI-generated content on the web surpassed human-generated content for the first time. Many companies are eagerly joining in, but at the same time, they’re suffering as a result. Their own messages are getting lost in the noise. The challenge is thus shifting from creation to relevance and impact. Anyone who still wants to be heard needs more original ideas, more context, and less mere execution.

A lack of resources, the desire for greater reach, and the challenge of producing enough content for various channels have long been the central concerns of many communications and marketing departments. Those who published frequently had invested in content marketing and were serious about it. Those days are over. What used to take a great deal of time for years can now be done quickly. Content is created in no time via prompts, and generative AI appears to have solved the problem of time-consuming content production.

Yet the effects of increased efficiency are almost always the same: more of the same. When it takes only minutes instead of two days to produce a text, we produce and publish more, fill the editorial calendar, and increase the frequency of publication. This phenomenon is known as the «Jevons Paradox»: increases in efficiency in the use of a resource lead to an increase in the total consumption of that resource rather than a decrease. In the specific case of AI-generated content, this means that the marginal costs of production are falling toward zero, and everyone is publishing more. The consequences? Vast amounts of mediocre AI-generated content are flooding the channels (University of Florida, 2026). More than half of all content on the internet is now created using generative AI (see Paredes et al. 2025). It is often of mediocre quality and goes fact-checked. We’re generating what’s known as «AI slop» because production happens at the push of a button (Merriam-Webster, Word of the Year 2025). Questions about who we’re producing for, with what intent, what the consequences are, and whether the content makes any strategic sense at all often remain unanswered.

AI slop isn’t a machine error. The machine usually writes reliably and, at first glance, appropriately. However, the machine continues to produce content reliably even when there’s nothing to say. It doesn’t ask whether a piece of content should be produced at all. This is precisely where a person with the ability to think strategically and critically is needed—a «human-in-the-loop» who makes clear decisions from the start and doesn’t just act as a proofreader at the end.

What AI Doesn't Take Over in Communications Work

Generative AI isn’t just changing how we produce content. It’s transforming the processes and requirements for marketing and communications professionals. This goes beyond mere prompting; rather, it’s about what the machine can’t do and shouldn’t do—that is, where and in what ways the machine cannot replace us. And shouldn’t replace us. As a result, the focus of our tasks is shifting: away from mere formulation, and toward human judgment, contextual knowledge, and subject-matter expertise. The focus is no longer on the question, «How can I use this tool more effectively?» but rather, «What can I contribute that the machine cannot do?»

Judgment: A model, for example, cannot know whether a statement is appropriate for the company in a given situation: It is unaware of both the ongoing conflict with a stakeholder and the informal agreement reached last Friday. The decision on whether to publish something therefore remains in human hands and requires the ability to select the right option from several plausible ones and to justify that choice. The bold decision not to produce or publish certain content requires strategic thinking and confident action.

Contextual Knowledge: Generative systems produce plausible content. Determining whether this content is also meaningful requires a “human-in-the-loop” approach. Knowledge of the organization, the market, current developments, and the needs of stakeholders and target audiences is necessary to produce meaningful results at the human-machine interface. Whether we use the machine as a content creator, sparring partner, or challenger, contextual knowledge makes all the difference.

Technical Expertise: Generative AI models calculate probabilities—regardless of whether the result is correct. We humans are responsible for this verification. Is that what we want? Should it be that way? Can it be that way? When we use tools for tasks in areas where we lack expertise and specialized knowledge, we cannot perform this verification. That is why generative AI supports us in areas where we can contribute our own knowledge and our own thinking.

These skills are not new. What is new is that they are no longer developed and honed to the same extent. Someone who used to write a hundred press releases developed a knack for relevance and concise phrasing in the process. Those who have a hundred press releases generated for them do not develop this sense. Skills that were previously learned on the job must be developed in a targeted manner going forward. The shift in responsibilities—from execution to accountability, from carrying out tasks to making decisions, and from knowledge to contextual understanding—takes time and commitment.

These skills are essential for deciding what and how to produce content—and whether a message will even get through. That’s because, alongside the flood of content, the ways in which content is discovered and whether it is trusted are also changing. Anyone who wants to be heard must therefore answer two questions at once: What is worth saying, and where does the trust come from that gives the message any impact at all?

Trust and credibility require a clear stance

More and more people are using AI-powered response systems instead of traditional search results (see DigitalLeverage 2025). They want quick answers and fewer clicks to get them. Google shows different results today than it did a year ago. At the same time, people are using ChatGPT and similar systems intensively every day to find answers to their questions. Traffic to websites and click-through rates are declining. This is changing how and for what purpose content is optimized. It’s no longer primarily about clicks, but rather about ensuring that a model can find, understand, and accurately reproduce a company’s own content. This requires a clear structure, explicit statements, verifiable facts, and accurate context. Generative Engine Optimization (GEO) is also changing the rules of distribution. A company’s own media, such as its website, are no longer the frontrunners when it comes to discoverability and visibility. AI systems systematically favor third-party sources over content that a company publishes about itself (Chen et al., 2025). So, if a company declares itself a market leader, it does little to convince the system. On the other hand, if a company is designated as such by a trade magazine, a study, or an independent review portal, it is more likely to be included in the AI’s response.

The logic of distribution is shifting from reach to credibility and trust. The old question was: «Where can we reach the most people?» The new one is: «Where can we place our messages and showcase our company without paying for it?» This requires a clear content strategy and the knowledge of where it’s worth speaking up. What are we a credible source for? What can we share that interests our target audiences? Who can tell the story credibly? A post published exclusively on the company blog is rarely cited by AI (Kumar et al. 2025). As a result, earned media and public relations are regaining importance. A mention in a trade publication, a quote from an independent expert, or a robust in-house study are enormously important for building trust and improving discoverability—and cannot be generated automatically.

From Experimentation to Integration

Attitude, trust, credibility: None of these can be mandated, nor can they be shouldered by a single person. The crucial question is rarely how well an individual uses the tool, but rather whether the team and the organization create the framework for it. Who sets the rules? How do roles, workflows, and quality assurance change in a marketing or communications department when suddenly everyone is working with the same tools, but at very different skill levels? And how does a team deal with this disparity in skills without some members falling behind and others charging ahead on their own?

The majority of employees at Swiss companies report using AI on a regular basis. About half of Swiss companies have identified use cases and launched pilot projects. Very few move beyond the pilot phase, and just over a third do not even measure the value of these projects (see the Columbus Consulting study, 2026). AI usage is on the rise, yet a large proportion of the pilot projects launched are not pursued further or are even deemed failures. Usage and integration are two different things. We have plenty of the former. We still have too little of the latter. Greater AI adoption does not automatically mean greater AI expertise. This expertise paradox is currently hindering the systematic integration and scalability of AI projects.

With the rapid development of generative AI, the demands placed on one’s own skills and available resources are constantly increasing. At the same time, the transition from pilot projects to full integration is stalling. If a technology isn’t embedded in work processes, using it remains an extra burden. Under these conditions, individual skills develop more slowly, or people use the tool less often because the barriers in everyday work are too high. Even «staying up to date» becomes an additional burden if a company does not offer training, has not defined use cases, and does not pursue an integration strategy. Individual employees are expected to make up for what the organization has failed to prepare. Yet technology is rarely the limiting factor. Rather, it is people and processes. There are plenty of tools available. What’s missing is clarity about what they’re used for (and what they’re not), who’s responsible for quality, and how the whole thing is embedded into existing workflows. It’s therefore crucial whether an organization creates the conditions under which experimentation can turn into integration: Who sets the guidelines? Who ensures that a successful pilot project becomes an established process, rather than just a memory of a good attempt? Who decides how quality is ensured when a machine is recording the data? Here, marketing and communications professionals are in the pole position. Their responsibilities include understanding target audiences, assessing relevance, and evaluating messages for impact and credibility. It’s not just about mastering a new technology, but about consciously applying existing expertise to the design of marketing and communication processes.

Shaping how we interact with AI, rather than just using it

Narratives about the future usually take one of two directions. One: AI will make us obsolete, so there’s no need for us to make any decisions at all—everything is predetermined anyway. The other: In the end, everything will stay the same, so we don’t need to make any decisions either—nothing will happen anyway. This “either-or” approach makes neither sense nor is it realistic for marketing and communications professionals. The relevant question has long been how AI is used. Three starting points can help us move from reacting to shaping the future.

#1 Strategically integrate AI rather than just using it on an ad hoc basis

Clear goals, responsibilities, and processes are needed for the use of generative AI. Not as a tool experiment, but as an integral part of communication—and always with an eye toward creating value. Effective use of AI requires the right framework, above all a clear stance that provides direction and instills confidence.

#2 Actively Managing Quality and Trust

Marketing and communications departments should define and establish clear quality standards for AI-generated content. Relevance, context, and credibility play a central role in this process. In an information-rich environment with a growing flood of content, curated quality supports the process of building trust.

#3 Developing AI Literacy as a Core Competency

Skills such as analyzing, questioning, and taking responsibility no longer develop on their own in everyday work. They must be developed in a targeted manner. This is achieved in a culture that allows for experimentation, uses mistakes as a source of learning, and encourages reflection. Competence is developed through conscious application, not through mere use. This takes time and commitment.

AI is effective when it is strategically integrated, competently applied, and implemented to a high standard. It is not the tool that determines the added value, but rather structure, accountability, and thoughtful action. This is a process of shaping the future, and it does not begin with the next tool, but with us.

So the question isn't what the machine can do next, but how we, as #TeamHuman, shape the interface between humans and machines.

Dr. Nadine Stutz ist Beraterin, Dozentin und Keynote-Speakerin an der Schnittstelle von Mensch, Technologie und Kommunikation. Sie befähigt Menschen und Organisationen in Transformationsprozessen, mit Fokus auf dem #TeamHuman, generativer KI und dem Aufbau digitaler Kompetenzen. Sie ist Dozentin am Institut für Kommunikation und Marketing (IKM) an der Hochschule Luzern Wirtschaft und Inhaberin von empower digital. | Bild: zVg von der HSLU
Dr. Nadine Stutz is a consultant, lecturer, and keynote speaker at the intersection of people, technology, and communication. She empowers people and organizations during transformation processes, with a focus on #TeamHuman, generative AI, and building digital skills. She is a lecturer at the Institute for Communication and Marketing (IKM) at Lucerne University of Applied Sciences and Arts (HSLU) and the founder of empower digital. | Image: Courtesy of HSLU

Thematically Relevant Continuing Education

CAS in AI for Marketing and Communication

This modular CAS program equips participants with the skills needed to use AI in marketing and communications effectively, responsibly, and in compliance with the law.

All SAS programs can also be booked individually.

SAS Generative AI for Marketing and Communications

SAS AI Transformation in Marketing and Communications

SAS Responsible AI in Marketing and Communications

SAS Applied AI Marketing Project

More Information About the CAS in AI in Marketing and Communication

> hslu.ch/ai

More articles on the topic