面向精益采购团队的代理人工智能：更好的可见性、更快的决策、全面控制

AGENTIC AIFOR LEAN PROCUREMENT TEAMSBETTERFASTER VISIBILITYDECISIONS PROCUREMENTAT THE SPEEDOF THE MARKET If that immature approach to procurement limits growth duringnormal business cycles, it is even more problematic in today’sbusiness environment, which is defined by uncertainty. For example,2025 has been a year of frequent and unpredictable changes in the id-market procurement teams face challenges inany macroeconomic environment. Some of those M mid-sized firms are both intensely focused on rapid growthand have not yet established the necessary structures and The challenge can be especially acute when mid-market companieshave various business units implementing their own procurementstrategies and policies. There are real-world implications to afragmented approach to procurement. From wasted staff time and This fluid and risky business environment exposes the deficienciesof procurement when it is managed with a patchwork of tools,spreadsheets, and manual work. Beyond the typical risk of errors “Mid-size companies need to move fast,” said Alex Zhong,Senior Director of Product Marketing for GEP, a provider ofprocurement and supply chain software and services. “Limitedresources mean they must operate with greater agility and The procurement dilemma mid-market companies face isdriven both by people and process. Like large enterprises,fast-growing mid-market firms need structures, processes, andpolicies to effectively manage their spending. But when theyattempt to implement new procurement tools and procedures, “Mid-size companies need to move fast. Limited resourcesmean they must operate with greater agility and resilience,making faster decisions and adjusting in real time rather than ALEX ZHONG,Senior Director of Product Marketing for GEP THE EMERGENCEOFAGENTIC AI Mid-market firms lacking defined procurement policiesand systems can accelerate their maturity, standardization,and flexibility by implementing agentic AI. Unlike previous To be clear, agentic AI is not the same as large languagemodels (LLMs) like ChatGPT. While LLMs can answer complexquestions and produce text and graphics to communicateanswers, they depend on people for direction. Agentic AI builds Agentic AI can make sophisticated decisions, providedagents are built and trained to achieve complex outcomes.In procurement, for example, negotiation is a complex andnuanced task that requires synthesizing and analyzing external “Agents aren’t generalists. They’respecialists, built around functions that can ALEX ZHONG,Senior Director of Product Marketing for GEP DATA IS THEFOUNDATIONOF AGENTIC AI Agentic AI’s sophistication and capabilities depend oncomprehensive and clean data from across an organization. Thiscan be particularly challenging for mid-market companies. “AIcan’t work without data,” said Raj Aggarwal, Director of Product More specifically, the types of data mid-market companiesneed to begin leveraging the procurement potential of agenticAI include historical spending patterns, metrics evaluating the Assembling, consolidating, and cleaning this data are necessaryfirst steps for agentic AI. But the investment of time and resourcesis worthwhile because it builds a data foundation for several CATEGORY MANAGEMENT Mid-market companies often lack explicit and formalized categorymanagement structures. However, AI agents can still developspending strategies for various categories, monitor market signals, SPEND VISIBILITY AND ANALYTICS POLICY ENFORCEMENT When data about company spending is stored in different databasesand spreadsheets, there is little chance that employees will find timeto rigorously assess whether that money is being spent wisely andstrategically. When an AI agent has access to that data, it can raiseconcerns about spending that exceeds contractual agreements Establishing procurement policies is straightforward; consistentlyexecuting it is not. Procurement policy agents can help by SUPPLIER ONBOARDING An AI agent can eliminate the labor-intensive work that goes intoonboarding new suppliers. From requesting information, pre-filling supplier profiles, and conducting verification checks, agents SUPPLIER RISK MANAGEMENT Supplier risk is everywhere. From geopolitical and trade disputes toextreme weather and supplier financial instability, tracking potentialsupply chain disruptions is vital for mid-market companies to meet A TOTALORCHESTRATIONAPPROACH TO GEP’s approach to solving mid-market procurement challengescenters on what the company calls Total Agentic Orchestration(TAO) for S2P (source-to-pay)—a platform that embeds anorchestration layer directly into procurement workflows. Rather THINK:Using information about impacts, agents can thenrecommend alternative suppliers by assessing whether newpartners can handle the volume of orders required, evaluating EXECUTE:Once mid-market leaders review proposedstrategies, agents can autonomously manage implementation This approach enables AI agents to app