In 2025, all signs point to the ongoing convergence of geopolitical tensions, economic volatility, and sustainability that defined 2024. Increasingly, organisations are placing their procurement functions at the centre of organisations’ efforts to combat these challenges. Consequently, procurement leaders are searching for new ways to mitigate headwinds while transforming the function into a lever for strategic value creation.
Industry experts widely agree that artificial intelligence (AI) is at the heart of technology-minded procurement leaders’ attempts to generate value and protect both operation and resilience. However, two years after the launch of OpenAI’s Chat-GPT, early adopters of AI tools are struggling to see a reasonable rate of return, and the blue sky promises made by AI advocates are failing to materialise.
However, there are those who believe that AI — both generative and the more traditional versions — will to play a critical role in how procurement will tackle the obstacles that await in 2025 and beyond.
While Sam Altman’s goal of an artificial general intelligence (AGI) seem like they’re hitting what people in silicon valley are increasingly calling “the wall”, some believe that 2025 will be the year that generative AI tools start to take on more autonomous, decision-making roles.
2025 will be the year we see “AI agents” take over procurement
This year, “AI agents powered by Generative AI, aka Agentic AI, will become a reality, accelerating procurement from a transactional department into a strategic force for the business,” argues Vishal Patel, VP Product at Ivalua. According to Patel, “these sophisticated systems will be embedded into comprehensive spend management platforms, designed to perform complex tasks with minimal human intervention.”
Patel believes that the next generation of Large Language Models (LLMs) — the ones that need even more copyrighted data and stolen hollywood scripts or whatever it is they’re feeding them now, than the last generation — will finally have the processing power and diminished propensity to hallucinate fictitious nonsense necessary to “access and analyse data, evaluate options, make informed decisions, and take action. Operating within a multi-layered framework, agents will integrate enterprise and/or other data sources, orchestrate workflows, interact with users and much more.”
Patel notes that, “to date, Generative AI features have operated on a ‘one-and-done’ basis—delivering outputs that serve primarily as drafts or prototypes. However, advances in LLMs now enable multi-agent AI ecosystems that perform complex reasoning, validate outputs, and provide actionable insights.”
Such a shift in generative AI’s potential would allow tools using the technology to use adaptive, real-time conversational AI capabilities to elevate procurement’s potential, Patel adds. As an example, he highlights the fact that “rather than relying on isolated AI features, smart source-to-pay platforms will leverage AI agents to manage tasks like spend analysis, RFP generation, and contract negotiation autonomously.” This, he continues, will allow procurement professionals “to focus on high-value decisions, driving productivity and strategic value. To leverage AI securely and reliably, organizations must ensure Agentic AI solutions rely on privacy-first multi-instance architecture to isolate sensitive data, utilise no-code tools to increase flexibility and rely on a unified data model to ensure outputs are accurate.”
Unstoppable evolution meets immovable regulation
The other trend intersecting with procurement and its use of increasingly autonomous AI tools relates to the ways in which governments are working to regulate the technology.
“2025 will be the year where AI regulation moves from theory to practice,” explains Patel, noting that the new EU AI Act will play a leading role in setting the regulatory pace for AI globally. “Unlike the slow rollout of sustainability regulations, the global nature of AI technology means requirements will cascade rapidly across multiple countries and regions,” he says.
However, in 2025 Patel predicts that, regardless of regulation, businesses will put into place clear documentation and transparency around AI decision-making—whether required not by governing bodies. “While some organisations may view this as a constraint, these regulations will accelerate AI adoption by providing the governance framework many organisations have been waiting for,” he explains. “The key for businesses will be transforming these compliance measures into AI solutions from the ground up, rather than treating them as an afterthought. For procurement and IT teams, this means really understanding what AI is in the tech that is being purchased, what the data privacy and security policies are, how models are being trained and much more.”
