By Leon Gauhman
2024 reaffirmed Generative AI’s trajectory of innovation, with groundbreaking announcements that didn’t just push boundaries — they redefined them. The breaking of benchmarks became routine, highlighting the field’s relentless pace. At the forefront of this progress are multimodal models, which enable AI systems to not only process text but also interpret images, audio, and visual expressions. This leap toward multimodality equips models with a deeper and more sophisticated understanding of the world, setting the stage for more intuitive and versatile human-AI interactions, and paving the way for robots that can interact with and understand their environments.
Beyond the headline-grabbing announcements from OpenAI and Google, a quieter, more telling story is emerging: the AI landscape is maturing. Key players have now launched their flagship models, and the field is beginning to converge. OpenAI’s forthcoming model, Orion, for example, hints at a slower pace of development, suggesting a shift from raw innovation to refinement and application.
So what does this mean for tech trends in 2025? Here are six insights we have uncovered:
1. Same, same: GenAI becomes commoditised
Two years on from the launch of ChatGPT, GenAI is rapidly becoming commoditised. Thanks to the speed at which they can be trained, open source GenAI models such as Mistral, Stable Diffusion and Meta Llama 3 are fast closing the gap with OpenAI, Google, Microsoft, Anthropic — today’s market leaders. Not only are open source models being trained quicker, their use of synthetic data means the model can be trained more efficiently. Increased competition between the open source and foundational models is good news for businesses as GenAI and LLMs will become cheaper and more available.
2. Bye Bye SaaS — DIY AI will become the new normal
The next generation of software solutions will have to feature a significant element of GenAI which is increasingly commoditised. What this means is that building AI solutions on top of an expensive SaaS platform makes no sense because enterprises can take GenAI and build it internally. With spending on SaaS tools expected to reach $197 billion, business leaders will come under growing pressure to axe their expensive SaaS platforms.
Equally, with GenAI more commoditised, creating AI solutions on top of a proprietary foundational model will also make less business sense. Instead, forward thinking business leaders will exploit the competition between the open source and proprietary foundational models — which continue to converge in capabilities — to seek solutions which allow them to build custom built AI solutions internally from scratch. Some of these models will be traditional LLMs, whilst others will be Small Language Models (SLMs), which are optimised for specific tasks and consume fewer resources.
3. Generative UI will be a competitive differentiator
Current AI interfaces are designed for general-purpose interactions, which limits their effectiveness in specialised domains or complex workflows. Users often find themselves spending considerable time crafting precise prompts to get the desired results.
As we move beyond the chat interface to more sophisticated Generative AI (GenAI) systems, the future will be defined by Generative UI (GenUI): interfaces that dynamically adapt to user needs, hiding the complexity of AI behind seamless experiences. We saw this happen with computers and GUIs and with mobiles and the iPhone.
GenUI is a fundamental shift in how we create and interact with user interfaces. At its core, it leverages AI models’ ability to generate code in real-time, enabling new and bespoke UI elements and interactions to be created on demand. This shift will allow users to focus on the task at hand, rather than on how they communicate with the AI. By using GenUI to solve the right problems in the right way, companies can create delightful tools and employee experiences that help to differentiate and deliver competitive advantage.
4. 2025 will see the rise of autonomous AI agents
In 2025 we can expect technology to move towards autonomous agents such as Baby AGI and Auto GPT which have greater agency around allocated tasks. In practice they can carry out different stages of a task and act on results without needing human input.
With the arrival of multi-agent frameworks like OpenAI’s Swarm, LangGraph, and React, businesses can now bridge disconnected agents through intelligent, collaborative systems, creating a cohesive network that transforms decision-making and operational efficiency.
These frameworks are driving multi-agent systems to become increasingly agentic, meaning they are powered by autonomous agents capable of reasoning, planning, learning from past interactions, and communicating with each other. This shift allows multi-agent systems to go beyond simple task coordination, fundamentally reshaping how software integrates within enterprise environments with the potential to supercharge productivity.
5. Big incumbent enterprise names will announce large AI transformations
The arrival of multimodal AIs that can understand visual and verbal contexts and combine them to create a new, more sophisticated sense of the world, represents a step change in productivity. Businesses can now use a range of inputs including voice, video and code to aggregate a much broader range of information and context for their AI to reason with. Ultimately, multimodality represents a move towards capturing and querying an organisation’s complete collection of expertise and know-how. All of this is leading to a new era of AI enhanced employees, who are supported and empowered by AI tools, personally tailored to their needs. These bespoke tools will boost employee productivity and wellbeing at work, leading to new levels of growth and efficiency.
In 2025 expect to see big enterprise names make announcements around AI and productivity which will put pressure on the market for other companies to do the same. JP Morgan Chase’s AI assistant LLM Suite is an early example of this trend. Due to be rolled out to 140,000 employees, the GenAI tool doubles as a “research analyst” that can create and refine written documents including summarising lengthy tomes and offer creative solutions using Retrieval Augmented Generation (RAG). RAG blends proprietary data with the underlying language models in order to reduce hallucination, increase accuracy and turbocharge insight gathering. The bank estimates its AI applications could deliver up to $2billion in value.
6. Expect large-scale deployment of robotics
2025 is set to be transformative for robotics, with large-scale deployments anticipated in manufacturing and logistics, led by companies like Tesla. This growth will be fueled by GenAI, multimodality and computer vision, enabling collaborative robots (cobots) — such as Agility’s Digit — to adapt to complex tasks, interpret diverse inputs, and work safely alongside humans. Cobots are expected to support various industries, from automotive assembly to food and beverage, where their precision and adaptability can manage high-mix, low-volume production.
On the consumer side, we’re likely to see household robots capable of handling a range of everyday tasks—cleaning, unpacking groceries, and waste disposal — bringing once-novel conveniences into the mainstream.
![iStock-1438575049 (1) [Converted]](https://www.europeanbusinessreview.com/wp-content/uploads/2024/11/iStock-1438575049-1-Converted-100x70.jpg "“Does Everyone Hear Me OK?”: How to Lead Virtual Teams Effectively")
