Agentic AI is reshaping corporate procurement by moving beyond decision support to autonomous execution. Pactum is using it to automate tasks such as requisition handling, supplier communication, and compliance checks, helping enterprises manage procurement more efficiently across large supplier networks.
Pactum positions itself as an execution layer for enterprise procurement, using agentic AI to carry out work at scale rather than to plan or analyze. The company says its platform helps large companies handle routine procurement tasks within existing policies and systems, without replacing procurement teams or core software.
The firm targets a common problem in global enterprises: procurement scope has grown faster than staffing and operating capacity. As a result, execution can be uneven, especially for long-tail suppliers and routine buying processes. Pactum says its AI agents can review requisitions, check policy compliance, engage suppliers, and support negotiations under set guardrails.
The platform is designed to work alongside tools such as SAP Ariba, Coupa, and Oracle Procurement Cloud, which remain the main systems of record. Pactum integrates through standard APIs so procurement data and supplier information can feed into agent-driven workflows without major system changes.
The company draws a line between automation and autonomous execution. It says its agents can complete defined procurement actions independently within governance limits, with a particular focus on tail spend management.
Pactum argues that enterprise procurement will increasingly rely on a layered model: ERP systems will provide structure and control, while AI agents handle day-to-day execution.
In an interview with DIGITIMES Asia, Kaspar Korjus, CEO and co-founder of Pactum, shared the firm's view and strategy in the AI agent era.
Q: Could you briefly explain what Pactum does today in simple terms for enterprise clients, and what specific procurement problems you are solving that traditional procurement software has not been able to address?
A: Pactum is an agentic AI workforce for procurement that helps companies execute sourcing and supplier management continuously across their supplier base. Its AI agents handle intake, buyer operations, requisition review, policy checks, supplier engagement, and negotiations within existing enterprise guardrails.
It addresses a common problem for global enterprises: procurement teams often have broader responsibilities but little additional capacity. Many organizations have strong strategies and processes, yet thousands of suppliers and routine purchasing decisions still go under-managed because execution is too manual and inconsistent.
Pactum's AI agents close that gap by automating approved procurement workflows and working directly with suppliers. This lets teams focus more on strategy, supplier relationships, and oversight while routine execution is handled continuously.
Q: How does Pactum position itself within the broader enterprise procurement and spend management ecosystem, and would you describe it primarily as a procurement tool, an AI decision layer, or something fundamentally different?
A: Pactum says it is not a procurement tool or AI decision layer, but a procurement execution workforce that lets AI agents carry out workflows within enterprise guardrails.
As SAP and Coupa add more AI orchestration and governance features, Pactum aims to act as the execution layer inside those ecosystems. Its native integrations with SAP Ariba and Coupa allow procurement work to continue within existing procure-to-pay processes without disruption.
Pactum also says its AI agents can help manage tail spend and routine procurement tasks, including supplier requests, policy checks, and buyer decisions, so even smaller suppliers can be handled more consistently.
Q: How do you see Pactum's role evolving alongside traditional procurement platforms like SAP Ariba, Coupa, or Oracle Procurement Cloud, and what determines success in this space, data access, workflow integration, or AI capability?
A: What defines success in the procurement industry is the combination of data access, workflow integration, and AI capability working together seamlessly. Data and workflow integration provide the operational foundation, but the application of agentic AI enables procurement organizations to scale beyond human capacity constraints within enterprise environments.
Pactum's role is evolving alongside traditional procurement platforms as part of a more integrated, system-driven model. SAP Ariba, Coupa, and Oracle Procurement Cloud will continue to serve as systems of record, while execution becomes increasingly embedded and automated within those environments, shifting procurement from human-only administration toward AI-assisted procurement execution operating under governance frameworks. Pactum complements this platform shift by operationalizing procurement strategy through embedded execution workflows, continuous supplier engagement, and scalable procurement operations across the supplier base.
This also creates a clear distinction between Pactum and the broader wave of AI products that function primarily as LLM wrappers, dashboards, or copilot-style tools that still rely heavily on human intervention to complete work. Pactum's AI agents autonomously execute procurement tasks while remaining under the control of procurement policies, enabling organizations to move from AI-assisted to AI-driven procurement operations at high capacity.
Q: What does the integration with SAP Ariba mean strategically for Pactum's role in enterprise procurement systems, and how do you balance AI autonomy with the governance and control requirements of large ERP environments?
A: Integration with platforms such as SAP Ariba helps make AI agents accessible to procurement teams within their existing workflows. Rather than replacing enterprise resource planning systems, Pactum extends them by letting agents operate within clear policy and governance guardrails.
Using standard application programming interface connections, Pactum pulls supplier data, current terms, and spend history, reviews requisitions, checks policy compliance, engages suppliers, and negotiates approved terms. Each negotiation creates a full audit trail of offers, counteroffers, and decisions, giving procurement, compliance, and auditors transparency without extra manual documentation.
The approach focuses on governed execution and integration. Humans retain responsibility for strategy, policy, approvals, and oversight, while AI agents handle operational workflows within enterprise controls.
Q: How should we think about the relationship between ERP systems and AI agents in the future enterprise stack? Do you see AI agents as a layer on top of ERP, or eventually embedded within it as core functionality?
A: ERP systems are built to centralize data, standardize workflows, and provide control, but not to handle large volumes of daily operational work on their own. That work has traditionally depended on human teams managing sourcing, suppliers, approvals, and negotiations.
AI agents add a new layer by carrying out procurement tasks continuously using existing data, policies, and workflows. ERP platforms remain the system of record and governance, while AI agents become the system of execution.
The result is not just task automation, but the ability to apply procurement strategy across the full supplier base in real time. Over time, enterprises are likely to expect procurement systems to do more than store information and manage workflows: they will be expected to complete operational work autonomously within business controls.
Q: Beyond procurement, what other enterprise functions are most suitable for autonomous AI agents, and how do you think decision-making authority will shift between humans and AI systems over the next 5–10 years?
A: Procurement has become a key test case for autonomous AI because it combines structured data, clear rules, and measurable results. Human oversight still matters, but well-defined workflows and governance can let AI agents handle large-scale procurement tasks reliably.
The same approach can extend to finance, supply chain, and pricing, where repeatable, rules-based decisions fit within set guardrails. Over the next five to 10 years, authority will remain with enterprises, while AI systems take on more of the execution. The winners will be organizations that pair human oversight with governed AI embedded in daily workflows.
Article edited by Jerry Chen
