Artificial intelligence will also make great progress next year
AI agents, AI with synthetic intuition and a high demand for further training and human expertise. We present nine trends that will shape the development of artificial intelligence in the coming months.
1. The AI agents are coming
Microsoft has the copilot, SAP calls it Joule, Atlassian has Rovo and Salesforce recently launched an entire platform called Agentforce: No doubt, the AI agents are coming. They are programs that carry out tasks independently like personal assistants. You summarize emails, create presentations, answer customer inquiries and create a to-do list for the upcoming day. “2025 will undoubtedly be the year of the AI agent,” says Nina Koch, Director of Customer Success Central Europe at Slack. “Never before has there been a technology that was able to carry out complete, multi-stage workflows completely autonomously, from the decision to execution to the reaction to new information, without human intervention.” Swen Büttner, Managing Director of MGID Germany, also sees this So: “AI agents, advanced systems that work independently and complete tasks independently will prevail.” In companies, this could alleviate the shortage of skilled workers and curb cost issues; areas such as customer service, sales and marketing in particular are seen as benefiting from the development. And in the mobile version, the assistants become your constant companion. Chunyang Chen, Professor of Software Engineering and AI at TUM: “Smartphone assistants have great potential and are being increasingly integrated into work processes.”
2. AI develops synthetic intuition
To date, multimodal models have processed text, image and audio files. Now development is about to take a new leap. Multimodal systems not only process the data types, they develop an understanding of their deeper connections. “AI systems develop a kind of synthetic intuition through the simultaneous processing of different types of data,” explains Thomas Porwol, Managing Partner Innovation & Digital Experience at DDB, and also gives an example: A product post on Instagram, combined with the tonality of the comments, Sharing behavior and current sales figures suddenly lead to a new, holistic understanding of market developments – in real time. Or to put it another way: A fashion trend on TikTok, combined with climate data and economic indicators, leads to precise predictions about future consumer preferences. AI also develops an intuitive understanding of markets. “GPT-5 and Google's Gemini will probably set new standards,” says Porwol. “This next generation will not only process but truly understand different types of data – from market trends to cultural shifts.”
From short videos to AI
These social media trends will really matter in 2025
If you want to keep up with social media marketing today, you have to remain flexible and consistently adapt to trends. Stefan Pichlmeier from Metricool highlights the most important social media trends for 2025 and shows how brands and content creators get the most out of their activities. …
3. Inference speed increases
If AI systems develop into thinking units, the reasoning time of the individual models will also decrease. Until now, GPT-4 still needed some time for complex tasks, such as preparing a master's thesis. This process will accelerate significantly as the inference speed of learning models continues to increase. “As early as 2025, we will see how large language models can handle highly complex tasks in minutes instead of hours – with a quality that still requires days of 'thinking time' today,” says Thomas Porwol. “This acceleration means not only faster, but fundamentally better results.” In a fraction of today’s time, for example, contracts can be legally analyzed, technical concepts or strategic analyzes can be thought through with all possible scenarios.
4. In-house teams for AI
There is probably hardly any agency or service provider left that works without the support of Generator AI. AI is also becoming more and more popular in companies. In November, the Federal Statistical Office published a study according to which one in five companies uses AI, and a Bitkom survey from October came to similar results. That doesn't sound like much, but it is a significant increase compared to the previous year and will continue. As a result, strengthened in-house AI teams will be set up in the company in 2025, especially in the marketing departments. This, believes Siggi Rakovich, CEO and founder of Hunch, will reduce dependence on traditional agencies. The marketing experts, including those in the agencies, would develop into “master marketers” who manage entire campaign life cycles, while the AI took care of the routine tasks in the background. An example is the creation of websites. Using intelligent tools, it is now relatively easy to create landing pages. Frank Froux, CEO and founder of Matelso: “With these tools, companies can make their online presence more efficient, react more quickly to market changes and become more independent of the shortage of skilled workers.”
5. Increased upskilling
All companies that use AI must train their employees accordingly. The EU AI law already stipulates this. However, this is not the only reason why further training measures will increase significantly nationwide. AI is considered a key technology for efficiency and competitiveness. At the same time, all applied knowledge of which standards and ethical principles must be adhered to must be adhered to. “In-depth knowledge of how and where you can use AI in everyday work will be one of the most important skills in 2025,” says Sven Wiesner, Managing Director of Neon Gold Innovations. “Companies must therefore invest heavily in suitable further training courses in order to create a uniform knowledge base in the teams.” This can also prevent employees in companies from being divided into two camps: those who already use AI on a daily basis. And those who are still afraid of contact and may say goodbye to an important development.
6. Governance models are in demand
The targeted development of skills among users, but also among developers, will also be a focus in 2025. In addition, the EU AI Act requires companies to develop new regulatory models that go beyond the usual approaches. Companies must establish clear policies, control mechanisms and processes to meet the law's requirements – such as transparency, fairness, data protection and risk minimization. “The introduction of the AI Act presents companies with the challenge of implementing governance models that significantly exceed currently prevailing approaches,” emphasizes Andreas Liebl, CEO of the AppliedAI Initiative. The EU AI Act will become effective after a somewhat difficult “getting to know you phase” and the gradual identification of points of conflict with existing regulations, confirms Tom Peruzzi, spokesman for the management and CTO of Virtual Minds. This has serious implications for data management. In the future, users will need to know exactly whether the training data on which the models are based is trustworthy, relevant and reliable in the long term, emphasizes Anila Shah, Project Lead Digital Strategy & Consulting Esome Advertising Technologies. “This means that data transparency and provenance are crucial for investment decisions.”
7. Human expertise becomes important
It may sound like a contradiction. However, due to rapid technological development, the importance of human skills and assessments is becoming more important again. The individual skills at the human-AI interface would come into greater focus again, observes Simon Graff, AI expert and tech specialist. Low-threshold AI tools would allow ideas to be implemented quickly and would lead to a “generic background noise”. Outstanding projects that use AI creatively and across borders will stand out. Graff: “Excellent human ideas remain irreplaceable. And the ability to use AI in a controlled and consciously curated manner is becoming more important than ever.” This desire for human expertise is also noticeable elsewhere. With the expected flood of AI bots, human advice can become a luxury and a differentiating feature in e-commerce, for example. Sören Stamer, CEO and co-founder of CoreMedia: “In situations where automation cannot be achieved, this creates empathetic and personal customer experiences.”
Artificial intelligence
How AI enables new business models in the creative industry
The creative industry has always been characterized by its innovative strength and changeability. But never before have we faced such a profound change as the introduction of artificial intelligence. …
8. Debate about the environmental impact of AI
Even if the topic of sustainability has moved back in the list of priorities for many companies due to increasing economic pressure, it will be discussed intensively again in 2025 in connection with artificial intelligence. On the one hand, this is due to the immense energy requirements of the AI models, which is why Microsoft, Google and Amazon are actually discussing using nuclear energy. To other regulatory requirements such as EU law, which requires a transparent assessment of the ecological impacts of AI systems. “It is noticeable that both the supply and demand sides are paying more and more attention to solutions and tools that are more predictable in terms of energy technology and sustainability and that cause less consumption,” says Tom Peruzzi. Providers like vshosting, which operate their data centers. “In addition to the alarming rise in energy costs, companies are also increasingly focusing on the ecological footprint they leave behind,” says CEO Damir Spoljaric.
9. Quantum computing pushes AI
Energy demand will also increase because companies like Google and Microsoft are investing heavily in the future of quantum computing. These require extremely low temperatures, a special infrastructure, but also a lot of energy. The fascination is great because quantum computers can process many states at the same time and thus solve extremely complex problems. The potential value is immeasurable and rich, from accelerating drug development to genetic reprogramming in healthcare to moving closer to nuclear fusion, writes Benjamin Bohne, group vice president at Cloudera, in his tech trends for 2025. “Next year will be “The significant influence of quantum technology can lead to a race in which companies want to use the technology to improve AI capabilities and achieve a competitive advantage.” The large tech companies have therefore been investing a lot of money in research for years Quantum computing, also in 2025. They also expect the technology to significantly increase computing power in the area of artificial intelligence.
