The Invisible Handshake: AI to AI procurement negotiations

From the November 2025 edition of Supply Chain Management Review.

November 2025

The November 2025 issue of Supply Chain Management Review explores the topics of global supply chain resilience, innovation leadership, and data-driven transformation. Highlights include strategies for building…

Browse this issue archive.

Access your online digital edition.

Download a PDF file of the November 2025 issue.

Supply chain managers and procurement professionals are watching a new kind of negotiation take shape, one that doesn’t involve people sitting across a table or trading emails. Instead,  artificial intelligence (AI) systems are quietly working out the details of low-level contracts behind the scenes. The goal? Free up procurement teams from repetitive haggling so that they can focus on higher-value work.

Walmart helped put this concept on the map a few years ago when it engaged startup Pactum to pilot AI-driven negotiations with part of its global supplier network. The project targeted Walmart’s long-tail vendors, or those smaller contracts that large organizations often don’t have the time or resources to revisit. Pactum’s system automates those negotiations, standardizing routine terms and moving agreements forward without human involvement.

The result is what some are calling the “invisible handshake.” Instead of procurement teams trading calls and spreadsheets, AI agents reach agreements on price, payment terms and delivery details in minutes. For supply chain managers, this signals a future where technology handles the baseline work of negotiation while people focus on strategy, relationships and exceptions that still require human intervention and expertise.

The rise of AI-to-AI negotiations also reflects a much longer journey in supply management technology. Jim Fleming, CPSM, CPSD, manager of product development and innovation at the Institute for Supply Management (ISM), notes that early systems like enterprise resource planning (ERP) laid the groundwork by digitizing and tracking transactions.

“Those systems were capable, but they still required a lot of human interaction,” Fleming says. Traditional buyers still had to sit on both ends of the table, work through terms, and then record the results, for example. Artificial intelligence can effectively remove that manual layer and free up procurement and supply chain professionals to focus on more value-added work.

Fleming sees these shifts as part of the broader pattern of industrial revolutions, where each wave arrived faster than its predecessor. The fifth industrial revolution, he explains, is compressing cycles of innovation into decades rather than centuries. “We’re seeing transitions now that are in the 10-year to 20-year time frame,” says Fleming, who likens the pace to what consumers have experienced with smartphones: a constant layering of new features, new functions and entirely new behaviors that quickly become part of their daily lives.

Balancing efficiency with ethical boundaries

Like many AI-related innovations right now, the “invisible handshake” is still a work in progress. Dawn Tiura, president and CEO of Sourcing Industry Group (SIG), agrees AI has vast potential in procurement but warns that its role in negotiations comes with new risks. “It may be proven in procurement for handling routine tasks and repetitive tasks,” she says. “However, it’s a bit less proven in the negotiation space, where the models haven’t matured enough to show they can stay within an organization’s ethical boundaries.”

Tiura is particularly concerned about what happens as the systems keep learning. Just as an entry-level employee gains experience and starts making their own choices, for example, an advanced AI model could begin to move beyond its original parameters. “That’s the biggest worry I have regarding negotiations,” she says. “As an AI model continues to get smarter and smarter, it also can choose to make its own decisions.”

For procurement leaders, this raises questions about oversight, accountability and when a model’s judgment might drift too far from what a company intends. On a positive note, Tiura says AI is already excelling in lower-risk applications. She points to uses like invoice matching, supplier onboarding, purchase order creation and spend analytics, as well as scanning contracts for risky clauses and helping forecast demand. These applications reduce manual workload and improve supply chain stability. And in areas like inventory optimization, the technology is already delivering measurable gains by minimizing stockouts and cutting carrying costs. Looking ahead, Tiura says the focus should be on learning to manage AI responsibly. Procurement professionals need to become hands-on with the tools—even if it’s just through no-code platforms—to better understand how they’re built and how to guide them. “The goal is to capture the efficiencies while staying vigilant about ethics, human oversight and the limits of machine decision-making.”

Local governments test the waters

The push toward AI in procurement isn’t just happening inside corporations. Local governments are testing it too, often with mixed reactions. Brooke Smith, NIGP-CPP, MIS, MMC, UCC, city recorder in Murray, Utah, has been speaking on AI in procurement forums for several years and says the audience has changed dramatically. “The first year I had like 10 people in the audience and they were all people in AI and in tech,” she says. By the next year, she was facing a 50/50 split between enthusiastic adopters and those who were wary.

That shift, Smith notes, is now tilting toward more openness. In the last year she has seen more people who were once skeptical to begin experimenting with AI tools. Many are starting small (e.g., using AI to draft emails or rewrite letters) but are now asking how it can help directly in procurement. The questions are moving from curiosity to application: how does it benefit my agency, how can it stretch staff capacity when budgets are tight, and what role could AI play in covering workforce gaps?

Smith says the biggest challenge now is education. Some procurement professionals have never tried AI, for example, while others are already experimenting with AI agents and advanced tools. This creates a wide learning gap she works to bridge in training. “I’m being asked to present at conferences, but how do you talk to all people across the board?” says Smith, who tailors sessions with both simple prompts for beginners and demonstrations of how AI agents are being built to automate more complex processes.

Smith stressed that agencies and other organizations should approach AI with what she calls the three Cs: be creative, be cautious and be critical. Creativity opens the door to finding new uses, caution ensures sensitive information stays protected, and critical thinking keeps professionals from relying too heavily on the technology. “You have to experiment, but you can’t lose those critical thinking skills,” she says. “That balance is what will keep procurement professionals relevant as AI becomes part of daily work.”

The human element in a digital shift

For Fleming, AI-to-AI negotiations fit into a bigger story of digital transformation. He says supply management has moved from a transactional role to a more strategic one, where professionals use data and logic to guide decisions. That evolution means automation will take on more of the routine tasks, but people still must shape the strategy behind it.

Governance may be one of the biggest hurdles. Companies want to try new tools, but IT often sets the limits. “You have to set up some safe boundary conditions inside of the company that IT will allow individuals to explore,” Fleming says. Clear rules help organizations balance innovation with security and compliance. Procurement and IT need to work together to make sure experimentation happens in a safe, managed way.

Oversight is also key inside the negotiation process itself. Fleming notes that AI-to-AI negotiations are happening today—mainly in low-value, low-risk categories. In those cases, automation can take care of the basics while humans check the results at the end. The rise of agentic AI, or small bots designed for specific tasks, will push this trend further. Still, he cautions, “You don’t want to just turn it on, walk away and figure it out later.”

For now, Fleming says AI-to-AI negotiations still require human interaction and process checks. He believes the pace of adoption will depend less on the technology itself and more on how companies handle governance, oversight and the cultural side of change management.

Where it goes next

Tiura says the next phase will move beyond human-to-AI negotiations into true AI-to-AI conversations, where supplier bots negotiate directly with buyer bots. That shift could bring speed and efficiency, but it also raises questions about transparency and control. For example, she says companies are already seeing value in AI for tasks like supplier onboarding, invoice matching and contract review.

Pilots with Walmart and others show that suppliers have found AI tools easy to use and often prefer them, she adds, but full bot-to-bot negotiations are still rare. “That’s going to change and we’re going to start seeing more and more of those bot-to-bot interactions,” she says.

For procurement leaders, the focus now is on setting clear expectations. Tiura says every organization needs to decide how it wants AI used, define what ethical use looks like and make sure employees understand the boundaries. That clarity will help build the trust needed for AI-driven negotiations to scale. “We’re already seeing AI-to-AI negotiations emerge in low-level contracts,” she says, noting that when the technology hits its stride, “it has the potential to be a true game changer.”

SC
MR