The Quiet Revolution at the Front Door of Enterprise Spend: How AI Agents Are Taking the Wheel

MOUNTAIN VIEW, Calif. – October 21, 2026 — For decades, the flow of capital within the world's largest enterprises has been governed by a frustrating paradox: the companies building the most advanced technologies on the planet are often bogged down by the archaic, bureaucratic software required to buy a simple server, a software license, or consulting services. The "front door" of corporate spending has long been a bottleneck of rigid enterprise resource planning systems and endless requisition queues.

Today, that front door is being fundamentally redesigned. Pactum, an enterprise AI company known for automating supplier negotiations, has launched its new Intake Agent. The tool allows employees to submit purchase requests in plain language via platforms like Microsoft Teams, Coupa, and SAP Ariba. Behind the scenes, the agent translates these unstructured requests into fully coded, policy-compliant requisitions, effectively cutting purchase request cycle times by 90%.

But to view this merely as an incremental software update is to miss the broader strategic rationale. This launch marks a definitive industry pivot from generative AI "copilots"—which merely assist human workers—to fully autonomous, multi-agent networks capable of executing complex financial workflows from start to finish.

From Copilots to Autonomy: The 'Hands-Off' Era

The enterprise technology landscape is currently obsessed with the copilot model, wherein artificial intelligence acts as a highly capable intern, drafting emails or summarizing documents while a human remains firmly in control. Pactum's approach with the Intake Agent represents a departure from this paradigm, moving toward what industry analysts call "agentic AI."

Agentic AI systems do not just generate content; they execute governed, multi-step workflows. When an employee requests new office equipment or specialized software, the Intake Agent gathers the necessary documents, identifies the correct procurement category, and assigns the proper accounting codes. It asks clarifying questions only when absolutely necessary. Once verified, the Intake Agent hands the digital paperwork off to Pactum's other autonomous workers—such as the Requisition Alignment Agent, which recently surpassed one million automated checks—to negotiate pricing and terms directly with the supplier.

"Building software that assists procurement professionals is relatively easy. Building software that runs the whole process on its own, from the first request to the final purchase, is much harder," said Kristjan Korjus, Co-Founder and Chief Scientist at Pactum. "It is the difference between driver assistance and a car with no one behind the wheel. Most procurement AI still needs a buyer at the wheel for every request. Our AI agents take the request all the way to the purchase on their own, and it is remarkable to watch them pass the work from one agent to the next."

This multi-agent orchestration is where the true leverage lies. By 2024, agentic AI had advanced to the point of handling isolated tasks. Now, as we move through 2026, these agents are communicating with one another, forming an autonomous assembly line for corporate spend.

Fixing the Bottleneck: Plain English Versus Procurement Red Tape

For the average employee, corporate purchasing is an exercise in navigating red tape. Finding an approved supplier or selecting the exact accounting code requires specialized knowledge that spans finance, IT, and operations. When an employee guesses wrong, the requisition is rejected, triggering a frustrating cycle of follow-ups that can stretch procurement timelines from days to weeks.

The strategic brilliance of the Intake Agent lies in its user interface—or rather, its lack thereof. By embedding the agent directly into everyday communication tools like Microsoft Teams, as well as established procurement platforms like Coupa and SAP Ariba, Pactum is meeting employees where they already work. The AI absorbs the complexity of corporate policy, allowing the user to simply state what they need.

The efficiency gains are not merely theoretical. Enterprise benchmark data reveals that automated alignment and intake systems have dramatic impacts on operational velocity. SUEZ, a major utility and environmental services company, utilized Pactum's AI architecture within its Coupa system to achieve a 90% reduction in requisition cycle times, processing compliance and quote checks in minutes for nearly 3,800 suppliers. Similarly, a leading aerospace and defense contractor managed over $180 million in spend by cutting procurement lead times from days down to seconds.

"Procurement teams can now run indirect procurement with full autonomy while remaining in control and managing the exceptions," said Kaspar Korjus, Co-Founder and CEO of Pactum. "Our Intake Agent helps resolve questions earlier so requests can move forward with less back-and-forth and teams can focus on the decisions that need their expertise."

The Land-and-Expand Play: Capturing the Strategic Touchpoint

From a market perspective, Pactum's move to the front of the procurement funnel is a calculated land-and-expand strategy. The company initially built its reputation on the back end of the process, using AI to autonomously negotiate contracts with long-tail suppliers. By introducing the Intake Agent, Pactum is aggressively capturing the initial employee touchpoint.

This territory is highly contested. The "intake-to-pay" market has seen explosive venture capital interest, highlighted by competitors like Zip, which recently secured a $190 million Series D round, elevating its valuation to $2.2 billion. Other players like Tonkean have also raised significant capital to automate business operations.

However, Pactum's integration strategy offers a distinct competitive moat. Rather than attempting to rip and replace incumbent enterprise resource planning giants, Pactum operates as an intelligent orchestration layer within them. With self-service integrations that read live catalogs and contracts, the AI ensures that platforms like SAP Ariba and Coupa remain the ultimate systems of record. Pactum's status as a Platinum Partner in the Coupa App Marketplace further cements its position deep within the enterprise software ecosystem.

By controlling the intake—the very genesis of a purchase—Pactum dictates the downstream flow of data. It ensures that its negotiation agents receive perfectly formatted, policy-compliant requisitions, thereby maximizing the efficiency of the entire autonomous network. It is a textbook example of vertical integration within a software ecosystem.

The Future of the Human Buyer

The rapid deployment of agentic AI inevitably raises questions about the future of human capital in procurement. With systems capable of taking a plain-English request, coding it, sourcing the supplier, and negotiating the contract autonomously, the traditional role of the tactical buyer is facing an existential shift.

Industry data suggests that the "wait and see" era for AI in procurement is definitively over. Market research indicates that 58% of the sector is already actively using or piloting AI, and analysts at Ardent Partners project that by the end of 2026, more procurement teams will utilize AI than traditional eProcurement systems.

Yet, enterprise technology leaders emphasize that this shift is less about workforce reduction and more about the reallocation of human intellect. The "manual middle" of procurement—the tedious data entry, the endless email chains correcting accounting codes, the routine supplier vetting—is rapidly disappearing. By automating these transactional burdens, human professionals are freed to focus on strategic supplier relationships, supply chain resilience, and managing the complex, high-value exceptions that algorithms cannot resolve.

Scaling these autonomous systems, however, requires rigorous governance. Giving an AI the authority to commit company funds necessitates bulletproof security and compliance frameworks. Pactum's adherence to SOC 2 Type 2 and ISO/IEC 27001:2022 standards is not just a regulatory checkbox; it is the foundational trust required for enterprises to take their hands off the wheel. As multi-agent networks continue to mature, the companies that thrive will be those that learn to manage digital workforces with the same strategic rigor they apply to human capital, fundamentally reshaping the velocity of global commerce in the process.