The rapid advancement of Artificial Intelligence and web technologies is fundamentally reshaping the way organisations manage procurement. Through this series of articles, Oalia explores these technological developments and analyses their practical impact on procurement teams.
As we move into 2026, understanding the evolution of Large Language Models (LLMs) has become a strategic priority. Their growing capabilities are transforming Digital Procurement by enabling deeper data analysis, more intelligent automation, enhanced supplier risk management and more informed decision-making.
Our white paper, “Digital Procurement – LLMs in 2026: Trends and Predictions,” provides a structured overview of these emerging technologies and explains how Oalia integrates them into its Source-to-Pay platform to create lasting value for procurement organisations.
Throughout this series, we examine the evolution of Artificial Intelligence, advances in web technologies and the convergence of these innovations. More importantly, we explain how Oalia turns these technological advances into practical capabilities that help our customers improve procurement performance and gain a competitive advantage.
As 2026 begins, I believe it is essential to understand where Large Language Models are heading—and what their evolution means for the future of Digital Procurement.
Looking Back: 2022–2024
The rapid progress of Large Language Models did not happen overnight. Years of research and continuous improvements have transformed them from simple text generators into intelligent assistants capable of supporting complex business processes. This progress extends far beyond reducing so-called hallucinations—which I personally prefer to describe as confabulations. The real transformation lies in making AI more reliable, controllable and better suited to professional use.
2022 – From Text Generation to Intelligent Assistance
2022 marked a major turning point. Large Language Models evolved from systems that simply predicted and generated text into assistants capable of interacting naturally and following behavioural guidelines. The arrival of ChatGPT brought Generative AI into mainstream business use.
Example – Teaching AI Professional Behaviour
Refusing inappropriate requests, Avoiding harmful or biased responses, Adopting a professional and context-appropriate tone
2023 – Specialisation and Enterprise Adoption
In 2023, AI became truly deployable at enterprise scale. Rather than relying on a single model expected to perform every task, organisations began creating specialised assistants trained for specific business use cases through supervised fine-tuning and domain expertise.
Example – A Legal Assistant vs. a Customer Support Assistant
Both assistants are based on the same underlying model, but each has been optimised for a different purpose:
One is trained to analyse legal clauses and produce structured legal assessments. The other specialises in customer support workflows, ticket triage and service interactions.
[…] Continue reading… Download our white paper to explore the latest trends shaping Large Language Models and discover how these technologies are transforming the future of Digital Procurement.