AI to Transform Procurement via Data and Security Leads

AI-driven changes to procurement and supply chain management will sharpen in 2026, with data governance, security and so-called agentic AI emerging as central concerns for chief information officers and procurement leaders, according to senior executives at Ivalua.

The company's experts expect organisations to reassess how they deploy AI across buying processes and supplier relationships. They predict a renewed role for IT departments in procurement technology choices, a stronger focus on trusted data foundations, and closer human-machine collaboration between buyers and suppliers.

Security and governance sit at the heart of many of the forecasts, as organisations expand AI pilots into core commercial workflows. The executives also see skills, not just tools, as a differentiator, with critical thinking and process design growing in importance alongside technical expertise.

Security as filter

Alex Saric, Smart Procurement Expert at Ivalua, said security concerns will define which AI tools gain traction in business software.

"AI in enterprise software will hit a hard ceiling in 2026 unless security is a core part of the agenda. Accidental exposure of sensitive information within and beyond businesses will increase as AI usage expands, with unproven vendors and bolt on solutions proliferating. CIOs will respond by tightening standards, scrutinising the access controls, data models and security certifications of proposed technologies, thereby sidelining point tools that cannot demonstrate end-to-end governance.

"Fragmented intake apps and stitched-together AI pilots are introducing data blind spots, workflow breaks, and new attack surfaces. The winners will unify three key components: the data model that underpins AI, the workflow engine that executes decisions and the governance layer that enforces policy by role and region. With these in place, IT can scale automation safely across procurement and supply chain," said Saric.

Saric expects the profile of AI practitioners inside procurement teams to change. He said success will depend less on specialist prompt-writing skills and more on process understanding and judgement.

"In 2026, organizations will realise that successful AI - particularly Agentic AI - projects in fields like procurement and supplier management aren't driven by prompt-engineering wizards or data science experts. Instead, they will be driven by the critical thinkers and workflow orchestrators who can break down processes and decide where AI can add the most value.

"Beyond workflows, the 'AI skills' that most business functions will need are sound judgment and a strategic ability to balance cost, risk, sustainability, performance, and people skills to build strong relationships with suppliers and other business functions - keeping a human-in-the-loop for the moments that matter," said Saric.

He also expects IT departments to regain influence over procurement buying decisions as AI spreads across operations.

"As AI integrates into every corner of the organization, IT leaders will renew their involvement in procuretech decisions to bolster efficiency and act as the architects of safe and scalable AI. CIOs will focus on unifying data strategy, governance and risk across finance, procurement and supply chains so AI can operate on solid ground. The organizations that move fastest will work backwards from clean, well-governed data, choosing source-to-pay platforms with a single data model instead of disconnected tools. That foundation will let them deploy AI for complex workflows, not just simple chat queries.

"For procurement leaders, this will feel less like inviting IT to block a decision and more like gaining a trusted advisor. Teams that bring IT in from the start will set clear guardrails and prove the value of automation to risk and compliance stakeholders. Those that buy in silos will stay stuck in AI pilots that fail to deliver expected results, while competitors scale the technology into everyday operations and reap the rewards," said Saric.

Data foundations

Vishal Patel, Senior Vice President, Product and Customer Marketing at Ivalua, expects a divide between organisations that invest in what he calls "agent-ready" data and those that persist with fragmented information.

"AI won't fail because of bad prompts, it will fail because the data behind the agents can't be trusted. The real divide we'll see in 2026 is between organizations that build AI on top of agent-ready data foundations and those that don't. Most procurement teams technically "have" the data: spend tables, contract repositories, supplier records, POs, invoices. But agents can't work with raw, inconsistent inputs. They need a layer that explains what fields mean, what's included or excluded, how it should be accessed, and when it can be trusted. Without that semantic and governance layer, agents hallucinate, misinterpret, or act on partial information.

"The leaders will rely on systems of record that are actually reliable, supported by strong governance and clear data definitions, and they'll enable agents to operate within those trusted frameworks. Because if you can't verify the integrity of the information your AI is acting on, you won't just make decisions faster, you'll make the wrong decisions faster. The companies that win will treat data as infrastructure: unified, governed, contextualized, and explained in a way agents can truly understand. That's the competitive gap forming right now in procuretech," said Patel.

Context over prompts

Pascal Bensoussan, Chief Product Officer at Ivalua, said the design of AI systems in business will revolve around context rather than prompts.

"The hype around prompt engineering misses the real point: Enterprise AI doesn't live or die by clever wording. It thrives or fails based on the context you feed it. The next phase of intelligent systems requires advanced "context engineering:" instantly retrieving the right data from multiple sources, mapping relationships through knowledge graphs and giving agents the information they need to reason correctly without rigid, pre-configured workflows.

"Procurement is a perfect example. An agent assessing a supplier, a contract clause or a sourcing event can't rely on a static prompt. It needs to pull relevant internal data, external signals, policy details, historical patterns - all in real time - and connect them logically. Organisations with unified data models and strong integration backbones will unlock AI's full value; everyone else will hit a ceiling," said Bensoussan.

He expects underlying hardware performance to influence the next phase of AI adoption in procurement and supply chains.

"People underestimate how much of AI's intelligence happens during inference, not during training. The next breakthrough won't come from bigger models - it will come from faster GPUs. When reasoning models can operate in real time, agents stop being fancy chatbots and start becoming true decision partners. Suddenly they can fetch context, reconcile data and reason through edge cases with the speed of a machine and the nuance of a human. That's the moment AI stops being a tool and becomes an operational engine,"

said Bensoussan.

He also anticipates wider use of tools that let business users configure AI agents in procurement environments.

"In 2026, organizations will finally bridge the AI vision gap by giving business users tools that allow them to embed their expertise directly into automated workflows. Currently, most business teams aren't struggling with AI due to a lack of ambition, but rather because the mechanics of building and orchestrating agents still feel inaccessible. That's why intuitive agent-building capabilities will matter so much. As interfaces shift from complex configuration to natural-language design, teams will begin assembling and managing agents in the same way they build a workforce: by defining roles, delegating tasks, setting guardrails, ensuring human oversight and escalation paths for edge cases."

Bensoussan links this shift with a change in how procurement teams measure return on investment.

"The new procurement ROI in 2026 will be driven by intelligent collaboration between humans and AI. The compounding commercial gains will come from teams finally having the bandwidth to build long-term relationships. Early agent deployments show modest time savings, but the real value - often 7x or more - comes from what teams can finally do once supported by AI that handles the transactional work. As agents take over intake, AP validation, supplier onboarding and other repetitive flows, procurement gains the capacity to engage suppliers more strategically and consistently. This human-machine hybrid operating model will drive deeper collaboration, resulting in better proposals, lower total cost, stronger SLAs, and clearer visibility into risk and sustainability."

Bensoussan said simpler development tools would also reshape how procurement professionals interact with technology.

"No-code AI will let procurement teams build digital workflows and automate routine tasks without writing a line of code. Professionals who once relied on IT will be able to design their own processes, connect stakeholders, and stand up cross-functional collaboration in minutes. AI will accelerate the work, but human judgment will decide how teams align, share information and act on insights. Leaders who lean into this shift will turn procurement into a digital collaboration engine - faster, smarter and far more adaptive."

He expects the relationship between buyers and suppliers to reflect these trends, with machine-to-machine interactions handling more routine work under human oversight.

"By 2026, under human-defined rules, buyer and supplier AI agents will start negotiating, exchanging data and resolving routine decisions autonomously. Instead of weeks of back-and-forth emails, agents will sync requirements, validate compliance and surface only the exceptions that need human oversight. This process will dramatically speed up the commercial cycle.

"But, this advancement doesn't replace relationships; it elevates them. Humans will still shape strategy, nuance and trust, while agents handle the precision work beneath the surface. Organisations that prepare their data and governance now will be the first to benefit from this new era of machine-to-machine collaboration," said Bensoussan.