2025 Forecast: Expert Predictions For The Year

US administration changes will have unpredictable and widespread consequences on many nations across a wide spectrum of operations. 

With a focus on implementing US trade tariffs, this is likely to lead to changes in supply chain for many multinational businesses. These businesses should look closely at their existing supply chain risks and act quickly to diversify and mitigate them. Information systems that can provide actionable insights will be critical, as will be an ability to make nimble business decisions based on those insights. 

Defence spending in many developed western economies is likely to feel increased pressure, as the US increases focus and budget on internal priorities. As a result, we can expect to see increased governance in procurement and supply chain operations across multiple government departments as well as increased pressure to drive internal efficiencies. The role of Artificial Intelligence will be amplified in these scenarios to provide automated and increasingly autonomous decisions in non-strategic spend categories to drive overall procurement savings with existing levels of staffing.  

Antimicrobial resistance will become elevated as a global concern, driving developments in new antibiotics, vaccines and diagnostics.  This will have implications not just for the Pharma companies, but also in agriculture and livestock production. The pace of many processes within these industries is likely to require a step-change, with significant pressure on existing systems to meet demands for capacity, speed and collaborative usability across multiple groups. Artificial Intelligence to accelerate research and optimised workflow technologies able to handle the increased complexity and pace required will be essential. 

My predictions for 2025: 

Research and innovation into foundational models will continue to increase throughout the year as it has done for the last few years. We will see bigger language models being released by the likes of OpenAI, Google, Anthropic and Meta, with more capability than ever before and with a bigger focus on multi-modal functionality (i.e. ingesting/producing images, audio, video, etc.). Albeit they’ll still mostly be being used to craft epic poems in the style of Dante or Homer about which of your friends smells the most…  

I’m also hoping there’ll be more research into the architecture of models to improve on the current transformer design. At the moment, it feels that the answer is to just throw more computing power at the problem; but given training the Llama 3.1 family of models produced about the same CO2 output as the lifetime of 475 medium size family cars, we need more innovative solutions for the sake of the planet. Having said that, I predict there’ll be more focus on smaller models too, being used for specific purposes, and being deployed on end-user devices (edge computing) enabling offline usage and quicker responses. 

At Cyferd, the focus will be on continuing to support customers to implement AI into their own solutions in an easy and intuitive way. There’s lots of great stuff coming in 2025, so it’s going to be a brilliant year. 

SaaS Applications are Evolving

The landscape of software delivery is undergoing a transformative shift. For over a decade, fragmented SaaS (Software as a Service) applications have dominated the tech ecosystem, offering businesses streamlined tools to manage everything from CRM to ERP, HR and project management. However, the rise of interconnected services and agents; AI-driven, modular systems that communicate seamlessly with one another, are beginning to reshape how we think about software functionality and user experience.

As AI becomes more commonplace, business logic will migrate to an AI layer, which is predicted to disrupt the role of these traditional SaaS apps. Instead of relying on a monolithic platform for project management, for example, companies can deploy agents that independently manage scheduling, task allocation, and progress tracking while dynamically sharing data across systems. This modular approach offers unparalleled flexibility, allowing organizations to customize workflows and scale interconnected functionalities.

The potential to disrupt traditional SaaS models becomes increasingly clear. These agents leverage advanced AI and machine learning to anticipate needs, automate processes, and improve decision-making. For businesses, this means fewer manual integrations, faster adaptation to changing needs, and cost savings. For developers, it offers an opportunity to innovate by focusing on highly specialized tools rather than competing in an oversaturated SaaS market. The shift from standalone SaaS platforms to interconnected services represents not just an evolution in technology but a reimagining of how software can drive business success.