How Agentic AI Will Turn Automation to Autonomy
In a recent fireside chat, Jeffrey Rajamani, Senior Analyst at Forrester Research, offered valuable insights into how agentic AI will shape procurement by 2030.
Speaking at Zycus' Unlock Deep Value with Agentic AI in Procurement digital event, he offered a compelling vision of procurement's transition from simple automation to true autonomy, presenting both opportunities and challenges for procurement professionals.
The evolution of AI in procurement
Jeffrey, speaking at Zycus' event, outlined three evolutionary cycles of generative AI that were progressively transforming procurement operations. The first cycle, prompt engineering, focused on crafting input prompts to guide AI models. The second cycle introduced Retrieval-Augmented Generation (RAGs), which enhanced trustworthiness by incorporating verified data sources. The third and most transformative cycle was agentic AI.
"Agentic AI systems or agentic AI workflows transition from passive tools to autonomous agents. They're capable of perceiving the environment, they're capable of making decisions and taking actions to specific goals," explained Jeffrey.
Unlike earlier AI iterations, agentic systems could work independently or in coordination with other agents, adapting to dynamic conditions without human intervention.
To watch any of the talks from the event on demand, click here.
What makes agentic AI different?
At its core, agentic AI represented a fundamental shift in how AI systems operate. "Agent systems should be synonymous with autonomously taking decisions, autonomously taking actions and then learning on its own â continuous improvement, we call itâto achieve specific goals," said Jeffrey.
This autonomous AI operated through four key stages: perception (gathering data), reasoning (processing information), action (executing decisions) and learning (continuous improvement). The system's ability to operate independently enabled it to handle complex tasks with remarkable efficiency.
Transforming procurement functions
Agentic AI was making significant inroads across the entire source-to-pay cycle, which Forrester labelled as "supplier value management." Jeffrey identified several key areas where this technology was already proving valuable:
Spend analysis and cost optimisation
"Agentic AI can analyse spending patterns and search for cost-saving opportunities. We can implement budget controls by blocking or flagging unauthorised suspicious spending," notes Jeffrey. This capability allowed for real-time monitoring and intervention, reducing wasteful expenditure.
Dynamic sourcing and supplier selection
One of the most promising applications involved autonomous supplier scouting and selection. "AI agents can scan the market for the best suppliers or pricing for a given material," Jeffrey explains. This dramatically accelerated the sourcing process while also improving outcomes.
Contract management
Agentic AI could autonomously draft contracts after supplier selection and negotiate terms using pre-defined thresholds. This not only sped up procurement processes but also ensured consistency in contract management.
Risk mitigation
"AI continuously monitors suppliers for compliance issues, legal compliances, regulatory compliances, financial risks, geopolitical challenges and so on," said Jeffrey. This proactive approach to risk management helped organisations anticipate and mitigate potential supply chain disruptions.
Procure-to-pay automation
Beyond strategic procurement, agentic AI was also transforming operational processes. "We see there are a lot of accounts payable agents, buying agents that can automate purchase order creation, invoice matchingâbe it three-way matching, four-way matchingâprocessing payments and so on," Jeffrey highlighted.
The future of agentic AI
Looking ahead to 2030, Jeffrey offered a bold prediction: "One in every three large organisations will start using agentic AI at scale and they'll be reaping a lot of benefits by then because they would've moved from experimentation to exploration to exploding."
He envisioned three types of agentic interactions: agent-to-agent (autonomous teams negotiating with suppliers), agent-to-human (augmenting human capabilities) and agent-to-environment (autonomous interaction with internal and external systems).
For procurement professionals, the message was clear: agentic AI represents a transformative opportunity, but implementation requires strategic thinking.
"Complexity and scale become the catalyst for transformation," Jeffrey emphasised. "A successful agentic AI isn't about blindly following the robots. It's about humans and AI working together in a partnership."
Rather than viewing agentic AI as a threat, procurement leaders should embrace its potential to free up time from manual tasks, enabling more strategic focus. As Jeffrey concluded: "The key is don't sit and stare, but adopt AI at the right time and in the right place."
To watch any of the talks from the event on demand, click here.
Explore the latest edition of Procurement Magazine and be part of the conversation at our global conference series, Procurement & Supply Chain LIVE.
Discover all our upcoming events and secure your tickets today.
Procurement Magazine is a BizClik brand
