While generative AI (GenAI) has taken the world by storm – with many noting the release of ChatGPT in 2022 as a pivotal moment in which it became clear what GenAI could do for people – many companies are experiencing with McKinsey & Company has termed the ‘GenAI paradox’: whereby the scalability of chatbots and copilots has meant eight in 10 businesses have deployed GenAI, but the same number report no material impact on their earnings.
Fortunately, there is a solution to the paradox, McKinsey believes: agentic AI.
Agentic AI refers to AI systems that can operate autonomously and make decisions independently of people. These capabilities mean these agents could, in theory, be deployed to autonomously screen CVs, book flights, or, in Pactum’s case, negotiate contract agreements for the procurement industry.
Pactum was founded in 2019 by Kristjan Korjus, Martin Rand, and Kaspar Korjus. They have created an agentic AI platform that automates contract negotiations and finds agreements for parties within as little as 15 minutes, according to the company website.
Contract negotiations in the procurement industry can be tricky. Fundamentally, both the suppliers and buyers need to negotiate favourable terms for each other and settle on a fair price, delivery and production time, standards of quality, and so on.
Pactum believes that AI agents will speed up this process and find the best deal for both parties, independently of human oversight, thereby freeing up human employees to get thinking strategically. There seems to be a consensus as to the benefits this will yield: in June 2025, the company announced its Series C fund raise, reaching $54 million, its most successful yet, and bringing its total funding to more than $100 million.
It also reported that its platform had seen a 489% increase in spend handled by AI agents, a 2.5x boost in annual recurring revenue, and the addition of 25 new Global 2000 customers.
The hope, from Pactum, is that while enthusiasm for AI was sparked with large language models (LLMs) and chat bots, the future will be agentic.
“Unlike generative AI, which generates new content and phrasing, agentic AI in procurement has the governance of pre-defined boundaries,” explained Kaspar Korjus, Co-Founder and CEO, in a conversation with Procurement Pro. “Which it uses to reason about what’s needed to achieve a goal, plan a series of steps, execute those steps, and learn and adapt from the outcomes.”
Deploying AI agents in procurement
Pactum’s platform draws on data sources such as existing contracts, invoices, and demand forecasts, to identify the best deals.
In terms of where AI agents can be applied to the procurement industry, contract negotiation is one facet. Autonomous sourcing, supplier discovery, and continuous risk monitoring are also areas where agents could be deployed to help enterprises.
“Agents in procurement allow a shift from reactive, transactional activities to more strategic, proactive functions,” said Korjus. “In addition to cost savings, agents enable more resilient and agile supply chains.”
Pactum works closely with Chief Procurement Officers (CPOs) and wider procurement teams, who are responsible for reducing costs, navigating risks, and ensuring they drive value. “They face challenges such as generating more savings with fewer resources, unmanaged spend, market cost dynamics, the need to integrate M&A, high cost of labour, seasonality, and geopolitical events such as tariffs,” said Korjus.
Some of the advantages Korjus mentioned that AI agents bring to the table – savings, allowing CPOs to focus on strategy, facilitate procurement agility, and create competitive advantage – reflect a belief in the capabilities that these agents have. As Korjus said: “Pactum’s agents can drive savings at a scale and speed that couldn’t be achieved by humans alone. The agents are proven to autonomously generate 2-10% savings on unmanaged and under-managed spend, which for a global enterprise is significant.
“While CPOs are asked every year to do more with less, Pactum agents allow execution of the procurement operating model at scale, with efficiency and agility.”
In its report on how enterprises can take advantage of agentic AI and navigate their way out of the GenAI paradox, McKinsey said that the broader challenge of agentic AI adoption won’t be technical, but human: earning trust, driving adoption, and establishing the right governance to manage autonomy and prevent uncontrolled sprawl.
Korjus seemed to be in agreement. “Some of the most challenging aspects of applying agentic AI to procurement are the psychological and change management aspects,” he said. “Some employees’ initial reactions can be fear of job security. The best implementations of agentic AI focus on the aspect of career growth by becoming a manager of agents. Management tools like leaderboards that track savings generated via agents help drive the positive embrace of change.”
Embrace agentic AI
According to Korjus, AI-driven procurement will oversee humans managing agents to do the work they would have done previously, which will result in greater cost savings and efficiencies, as outlined earlier.
Ultimately, the best employees will embrace this transition, Korjus said.
“Procurement strategy and operating models become the focus,” he said. “More and more manual tasks will be done efficiently via agents, with the best performing employees embracing the power of agents.”
