SAP: The Five AI Themes For Businesses to Watch in 2025

As technology providers shift focus from developing general-purpose tools to implementing specific business applications, companies seek return on their AI investments after years of experimental deployments.

Across the world, providers such as SAP, Oracle and Microsoft are integrating AI capabilities into their core products - and as a result, IDC predicts worldwide spending on technology to support AI strategies will reach US$337bn in 2025, with most investment directed towards practical applications rather than research and development.

Now, due to the increased pressure on technology companies to demonstrate concrete business value from AI implementations, major enterprise software providers are responding by developing AI systems that address a maturing market where practical implementation is replacing technological experimentation.

SAP believes the changes happening in 2025 are particularly evident in five key areas: autonomous AI agents, specialised AI models, enterprise adoption patterns, user interface design and regulatory compliance. 

Autonomous AI

The development of AI agents is progressing from basic document search capabilities to systems that can plan and execute complex business processes.

These multi-agent systems (MAS) are designed to work together, each handling specific parts of larger tasks.

SAP reports that AI agents will handle customer service exceptions and administrative tasks that have traditionally resisted automation efforts.

The systems will respond to business events such as supply chain disruptions or demand surges without human intervention.

The technology represents an advancement over robotic process automation (RPA), offering more flexibility for complex tasks that cannot be addressed through traditional programming methods.

Specialised AI models

The market for LLMs is becoming standardised for basic text generation tasks and this is pushing companies to develop specialised implementations.

In this context, Knowledge graphs, a technology for representing relationships between data points, are experiencing renewed interest as a method to improve AI system accuracy and this technique provides context for AI models, reducing errors in output.

Meanwhile, Physics-informed neural networks (PINNs), which incorporate scientific principles into their predictions, are emerging as crucial for robotics applications in industrial settings.

SAP Foundation Model targets enterprise data

SAP's Foundation Model, designed for processing structured business data, also represents a move away from general-purpose LMs towards systems optimised for specific business applications.

Additionally, enterprise software providers are developing systems capable of processing multiple types of data simultaneously, including text, voice, image, video and sensor information.

Enterprise adoption patterns

As the implementation of AI in enterprise settings is shifting from experimental projects to revenue-generating applications, companies are having to address legal and data privacy requirements specific to AI implementations.

"While 2024 was all about introducing AI use cases and their value for organisations and individuals alike, 2025 will see the industry’s unprecedented adoption of AI specifically for businesses.

More people will understand when and how to use AI, and the technology will mature to the point where it can deal with critical business issues such as managing multi-national complexities" the report states.

Jesper Schleimann, AI Officer EMEA at SAP says in a LinkedIn post: “2025 will be a transition year for AI as we move from AI pilots to wider scale adoption.

This will also require a greater focus on how we adjust our organizations to this new way of value creation.”

Enterprise software vendors develop model marketplaces

To meet the financial needs of businesses, software providers are creating platforms for deploying multiple AI models, allowing businesses to switch between different systems based on their requirements

These marketplaces aim to reduce the cost and complexity of implementing AI systems.

Therefore, the industry is moving towards an outcome-based service model, focusing on achieving specific business objectives rather than providing static software features.

User experience 

The integration of AI into business software is also changing how users interact with systems - leading to copilots replacing traditional user interfaces.

"AI won't be limited to one app; it might even replace them one day. With AI, frontend, backend, browser and apps are blurring," according to the report.

As a result, organisations are beginning to view AI as part of a collaborative intelligence network, combining human expertise with machine learning capabilities, which requires new metrics and training methods for effective human-AI collaboration.

Regulatory compliance

The Organisation for Economic Co-operation and Development's AI Policy Observatory has documented hundreds of proposed AI regulations worldwide, which indicates a fragmented regulatory landscape.

This means that companies developing AI systems are implementing their own safety and ethical use guidelines while formal regulations are being developed, as the focus is shifting from technical constraints to broader questions about the role of AI in society.

"The discussion will shift from what we try to regulate from a technical standpoint to how we innovate and what we deem fundamentally human," the report concludes.

“I believe adoption will be the hardest to tackle”, Jesper adds.

“My advice to get started: begin the education early to prepare the grounds, establish a safe AI space for your organisation to try it out and scale adoption by integrating AI in business processes to ensure user acceptance at scale.”

Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.

Discover all our upcoming events and secure your tickets today.

Technology Magazine is a BizClik brand