Multi-Agent GenAI for Procurement: A CPO's Guide to Cutting Through the Hype (2024)

• Published byAmit Shah
• April 5, 2024

• Artificial Intelligence

The procurement landscape is undergoing a digital revolution, with Generative AI (GenAI) emerging as a game-changer. Software vendors have been quick to tout the power of GenAI for streamlining your Source-to-Pay (S2P) process. There is (mis)information overload, and hence Chief Procurement Officers (CPOs) will need to be discerning during evaluations. The traditional procurement landscape is changing rapidly. New technologies like multi-agent GenAI for procurement are transforming how businesses approach sourcing and purchasing.

This guide equips you, alongside your technical partner, the CIO, to dissect GenAI solutions and ensure they deliver on their promises.

This blog post is a valuable resource for CPOs at all stages of their S2P automation journey. Whether you’re just beginning to explore automation or have already implemented basic solutions, GenAI represents a transformative leap forward. For CPOs navigating the initial stages of automation, understanding GenAI’s capabilities can help you chart a clear course towards a more streamlined S2P process. For those with existing automation solutions, GenAI presents a logical next step, offering the potential to further optimize efficiency and gain a competitive edge. In either scenario, a well-informed approach to GenAI evaluation is critical for maximizing its potential within your organization.

Watch Webinar: How Generative AI Can Set Procurement Leaders up for Success

Beyond the Hype: The Advantages of Multi-Agent GenAI for Procurement

Many vendors may simply throw around the term “GenAI” without delving deeper. A crucial differentiator lies in the architecture itself. Look for solutions built on a multi-agent model.

Imagine a team of specialists working together, each with a specific expertise:

• Intake Agent: Captures user requirements and translates them into actionable information.
• Sourcing Agent: Scours supplier databases and identifies qualified vendors based on user needs.
• Policy Agent: Ensures compliance with company procurement policies and regulations.
• Negotiation Agent: Analyzes historical data and utilizes algorithms to secure optimal pricing.
• Approval Agent: Streamlines the approval process by routing requests to the appropriate personnel.
• Risk Assessment Agent: Analyzes potential purchases for associated risks based on historical data and reports.

This division of labor ensures superior performance and accuracy compared to single-agent models that attempt to handle everything at once.

Demystifying the Technical Specifications: Key Considerations

While the multi-agent model is a foundational requirement, several technical aspects influence user experience (UX) and overall effectiveness:

• Latency and Lag: Multi-agent GenAI for procurement solutions should operate in real-time. Slow response times (latency) and delays (lag) disrupt workflows and frustrate users. Ensure the software has robust infrastructure to handle data processing demands without compromising speed.
• Scalability: Your procurement needs will evolve. The GenAI solution should seamlessly scale to accommodate new functionalities and data volumes as your business grows.
• Data Security: GenAI systems rely heavily on data. Evaluate the vendor’s data security protocols to ensure your sensitive information is protected against unauthorized access and breaches.
• Explainability and Transparency: “Black box” AI solutions can be concerning. Look for GenAI platforms that offer explainability features, allowing you to understand the reasoning behind the system’s recommendations.
• Agent Capabilities: While the presence of multiple agents is a positive sign, their effectiveness hinges on their individual capabilities. Here’s where tenant-level contextual training comes into play. A robust multi-agent GenAI for procurement solution allows customization of each agent based on your specific company’s procurement processes, terminology, and data. This ensures the agents understand the nuances of your S2P workflow and can make informed decisions.

Furthermore, consider solutions that leverage techniques like Retrieval-Augmented Generation (RAG) to enhance agent performance. RAG allows agents to access and retrieve relevant information from your internal knowledge base and external sources, enabling them to provide more comprehensive and contextually-rich responses. For instance, the Negotiation Agent could not only analyze historical data but also factor in specific supplier relationship dynamics captured in your CRM system, leading to more effective negotiation strategies.

Read more: How RAG Enhances Contextual Understanding in Source-to-Pay

Beyond Technology: The Human Element Remains Crucial

While GenAI empowers automation and streamlines processes, it doesn’t eliminate the need for human expertise. Look for solutions that complement your existing team, not replace them. Procurement professionals can leverage GenAI to handle mundane tasks, freeing their time to focus on strategic initiatives and supplier relationship management.

Aligning with Your CIO: Building a Tech-Savvy Procurement Team

The CIO’s role extends beyond simply evaluating the technical aspects of GenAI solutions. They act as strategic partners throughout the process, bringing their technological expertise to the table. Early involvement from the CIO is crucial during solution selection, ensuring the GenAI platform integrates seamlessly with existing IT infrastructure and adheres to security protocols. Additionally, the CIO can collaborate with the CPO to develop a data governance strategy that facilitates efficient data collection, cleansing, and preparation for optimal GenAI performance. By fostering a collaborative environment, the CPO leverages the CIO’s tech expertise while effectively relying on their own understanding of procurement best practices, leading to a successful GenAI implementation.

Read eBook: Democratising Procurement in 2024 through GenerativeE AI

Moving Forward: A Checklist for CPOs

Here’s a checklist to guide your GenAI vendor evaluation process:

• Does the solution employ a multi-agent architecture?
• What are the response times and processing speeds?
• Is the platform scalable to accommodate future growth?
• How does the GenAI system ensure data security?
• Does the solution offer explainability features for recommendations?
• How well-trained and comprehensive are the individual agents?
• Is there a central orchestration agent to manage agent collaboration?
• How does the GenAI solution integrate with existing procurement systems?

Conclusion: Embrace GenAI, But Do It Right

GenAI holds immense potential to transform your S2P process, but a critical approach is essential. CPOs, armed with knowledge and working alongside their CIOs, can avoid falling prey to mere marketing hype. By focusing on the multi-agent model, technical specifications, and fostering human-AI collaboration, you can unlock the true power of GenAI and optimize your procurement function for the future.

Related reads:

1. Insight to Foresight: Generative AI as the Procurement Navigator of Tomorrow
2. The Role of AI in eProcurement
4. Introducing the Generative AI Way to Manage Your Source to Pay
5. Introducing Merlin Assist: The AI-Powered Procurement Chatbot
6. Merlin Experience Center – Zycus Cognitive ProcurementMerlin Experience Center
7. AI-powered Contract Management Software- Merlin AI
8. Harnessing the Power of Multi-Agent Generative AI in Procurement Software: The Future of Intake Management Solutions

Recommended Resources