How AI agents are transforming procurement: What CPOs should know

Where—and how—agentic AI delivers procurement value quickly

Procurement is an almost-ideal spot for agentic AI to create value fast, since you typically have both the “fuel” that AI needs, and the tasks where it excels:

• Unstructured data that AI can turn into high-value, continually updated analysis
• Repeatable knowledge work that AI can execute, while humans review the outputs
• Ongoing internal and external collaboration, which AI can make seamless
• A digitized source-to-pay (S2P) foundation, so AI can get moving fast

Based on our work with clients and our rigorous modeling of procurement processes, we expect agentic AI to transform, very soon, at least 75% of procurement activities. Many may become “agent driven”: Agents can take much of the work, while people transition to decision making, relationship management, oversight, and strategy. In others, people and agents can together create a hybrid workforce that manages new workflows. Only a few tasks will remain entirely in unassisted human hands.

That could mean a productivity jump of at least 30%, rising to as much as 70% in agent-driven tasks. That could not only lower your costs. Agentic AI can also drive faster insights and create new strategic capacity, so you can cut your company’s cost base, boost resilience, and grow enterprise innovation. It can enable you to move beyond once-a-year category strategy and build a nimble, continually refreshed strategy, based on a unified data foundation and largely automated, “always-on” research and insights. But first, you’ll need to reimagine how key workflows operate.

Reimagine procurement: Intake and orchestration (I&O)

It happens all too often in I&O: Processes and data are scattered in different systems with different rules. It may not even be clear what the rules are. And too much work is manual and repetitive. But agents can partly automate about 80% of that work, while driving consistency across systems. For example, agents can

• Translate a business user (BU)’s request into a structured intake and assess it against procurement policy and approval thresholds
• Intelligently route request to the appropriate buying channel, assigning a procurement owner when required based on category, value, urgency, and complexity
• Check the preferred supplier list and supplier master to reuse approved suppliers where possible
• Check existing contracts and service catalogs to reuse existing agreements or trigger a new contract when required
• Check inventory and service scope coverage to help avoid duplicative purchases
• Gather required data across systems to pre-populate requisitions and contract requests, track supplier status, run risk checks and flag missing or relevant information to stakeholders
• Track and report progress across the process and notify owners and approvers if work stalls

In this workflow, agents follow your procurement policy to guide people through a compliant process, with many “nuts and bolts” invisible to users. Yet people are still in charge: They review and approve agents’ outputs and make any high-risk or high-value decisions.
But with as many as four fifths of their previous tasks automated, their work can be completed more quickly while improving quality, and you have new capacity to reinvest.

Reimagine procurement: Strategic sourcing

Too many mundane repetitive tasks, too little high-quality data, and too little time for strategy? If your sourcing faces these challenges, AI agents can help. An intake tool can automatically process a BU’s request, then puts sourcing agents to work.

• Gather data on past spend, pricing benchmarks, and supplier performance
• Identify cost and lead-time drivers in the specification
• Assess concentration risk
• Scan markets to identify pricing trends and new potential suppliers
• Generate research and scenario analysis on individual suppliers
• Draft and distribute the RFx and collect responses
• Offer data-driven guidance for negotiation strategies
• Notify stakeholders of decisions and store data for future analysis

All in all, they can take the lead on up to 25% of current activities and assist humans on another 70%, often significantly:  This kind of assistance can slash sourcing cycle time by 50% or more, giving people the time (and data) to build value creation roadmaps and make better strategic decisions: on priorities and trade-offs, or how many sources you need, or whether a purchase should be a one off or part of a long-term relationship. The result can be lower costs for you and lower risk, lower cost contracts for your company.

Reimagine procurement: Contract lifecycle management (renewals)

CLM is full of automation and strategic opportunities for agentic AI, offering new ways for your people to create value. Enterprise agentic solutions are increasingly mature, offering quick CLM value. Right now, contracting agents can, for example:

• Notify the Contract Manager (CM) and BU when a contract is expiring
• Assess supplier criticality, prior negotiations, scope and term variations, and data from multiple systems to prioritize contracting engagements
• Flag outdated terms for the CM’s review
• Enable rapid, trackable interactions with suppliers
• Analyze data sets that cover supplier criticality / performance, category strategy, and geopolitical forecasts to help the CM draft and negotiate new contract with more balanced risk language
• Analyze supplier redlines and suggest alternative language and negotiation strategies through data-driven insights
• Store new contracts and extract key details for data analysis
• Continuously monitor supplier performance, contract compliance, spend trends, and risk indicators, then recommend corrective actions

This mix of agent-automated and agent-enhanced activities can drive faster cycle times, fewer non-compliant contracts, more robust risk language, and stronger tracking and enforcement of terms. And with the new “free time” and data that agentic AI gives them, CMs can better partner with suppliers to foster strategic relationships that add value to both parties. That can directly help grow your company’s top and bottom lines.

Reinvest your new capacity: Procurement that drives strategy and innovation

As agentic AI largely automates execution work across procurement, your people could have the insights and time to create new value. Four areas are particularly rich in potential:

• Third-party strategy. With the time, data, and tools that AI agents give them, your people can create models or digital twins to simulate your supplier network. That can help drive enterprise-level strategies to reduce costs, manage risks, boost resilience, and better select suppliers with whom to deepen relationships.
• Category strategy. To drive a continually updated, “living” strategy, agents can synthesize vast amounts of internal and external data, surface the insights that matter, and generate first-cut, data-driven opportunities and levers for your category managers to consider as they finalize actions and strategy.
• Upstream decision making. With their new ability to forecast demand and other supply chain dynamics—including ones that impact what your company sells—your people can shape not just what the business buys, but what it does.
• Enterprise innovation. Once agentic AI has helped make more procurement data available for analysis, you can help the business create new products and services: by forecasting demand and costs, or helping to select and work with suppliers to co-create products or delivery models.

How to get started: 6 moves to make today

Transformative value depends on a new operating model for procurement. You can get there in stages, while creating value every step of the way.

1. Identify where you can kickstart value creation. To get that quick, significant win that pays for next steps, pick your first workflows based both on AI’s capacity to close high-value gaps that other technologies can’t and on data quality. The path to value is shorter if data is already reliable and compliant and comes with a knowledge base, even though AI can help you fix data too.
2. Upgrade your tech, strategically. Even if you already have digitized S2P systems, for AI you’ll likely need an updated digital architecture: harmonized master data models, scalable data pipelines, and interoperability across systems. With a strategic approach, you can move step by step, with AI itself helping to accelerate your modernization and keep your costs down.
3. Elevate engagement—and create excitement. Like other transformative tech, works best when people are excited about it. Fortunately, AI can offer your people so much—eliminating mundane work and enabling more creative, higher-value work, as well as improving the work / life balance—that if you offer incentives and psychological safety, you can get them on board with your AI transformation.
4. Roll out a new operating model. With early projects delivering big value and your people enthusiastic, it’s time to resdesign roles, responsibilities, workflows, and upskilling paths. People should rely less on trial and error, more on simulations and data. With less repetitive work to do, they can focus on creative, strategic thinking.
5. Let trust be a differentiator. Besides the human-at-the-helm oversight that you should embed in every agentic workflow, you’ll need a robust Responsible AI framework that can both manage risks and—by encouraging stakeholder trust and reducing the need for costly pauses and do-overs—accelerate value creation.
6. Keep improving—and scaling. With AI-specific proof points benchmarks and the tools that agentic AI itself can provide, track the value you’re creating—and identify ways to do better. Since the architecture that powers agents can be reusable, you can scale up quickly—and affordably.

Agentic AI can help you meet your CEO’s top priorities

As agentic AI helps you cut costs, better support the business, and drive enterprise innovation, it can also help you meet what’s likely one of your CEO’s top priorities: reinvention. In PwC’s Global CEO Survey, 42% of CEOs report entering new sectors in the last five years. For the last twelve months, 30% report new revenue from AI and 66% report stakeholder trust concerns in at least one area of business operations.

New sectors, new technologies, and new risks demand a new role for procurement: one with nimbler sourcing and risk management, greater transparency, and more insights to guide the business. Agentic AI can provide that and more. It’s time to get started—today.