Generative AI in procurement

By Oliver Knapp and Pieter Niehues

Recent advancements in machine autonomy have been remarkable. Their capabilities have evolved from rule-based automation, where machines follow predefined rules to perform tasks, to predictive analytics, where statistical models forecast the future, to cognitive automation, where machines use robotic process automation (RPA) and artificial intelligence (AI) to analyze data and make decisions.

In November 2022, OpenAI launched ChatGPT, a generative AI (GenAI) based on large language models (LLMs), which is now pushing this technological development to a new level. Previous technologies focused on “making data available” through digitized processes and data lakes, for instance, or “making data read and speak” via tools such as chat bots. GenAI takes this a step further by “making data create”. This means it can autonomously generate human-like creative output like new content or solutions such as strategic approaches or contracts.

GenAI is a real game-changer for procurement. It will enable technology to constantly perform human-like tasks simultaneously at much higher speeds, significantly, boosting efficiency. This will optimize procurement productivity and purchasing results, by enabling more and better-prepared sourcing events and supplier interactions, for instance.

GenAI’s implications and value for procurement

When it comes to digitalization, during the last 20 years chief procurement officers (CPOs) have typically focused on areas like source-to-contract, purchase-to-pay process, spend analysis, digital contracts and supplier web-portals. More recently, natural language processing (NLP) applications like chatbots and negotiation bots have emerged, which enable, more interaction such as asking questions or giving commands.

Now, however, GenAI applications enable creation of new content based on existing data. By automating tasks, providing valuable insights, and improving transparency, generative AI can help CPOs take their organizations to new heights with tremendous rewards expected across six core dimensions:

• Increased efficiency: includes process automation and simplified execution of procurement tasks, such as strategy development, tender management, negotiation preparation and contract management.
• Productivity improvement: includes freeing up employees from mundane tasks and allowing them to focus on value-maximizing activities, such as performance-based supplier development.
• Innovation: such as analysis of market developments, suggestion of alternative material specifications, and AI-supported ideation platforms.
• Risk mitigation: includes GenAI-powered early warning systems, evaluation of risk scenarios, and development of risk mitigation plans.
• Spend reduction: through areas like improved knowledge management, optimized procurement decisions, and accelerated procurement procedures.
• Sustainability and compliance: includes real-time CO2 footprint analysis and P2P process compliance monitoring.

But which use cases should companies start with to immediately and effectively leverage GenAI? As a starting point, GenAI should be used to quickly generate knowledge and ideas using LLMs, while at the same time increasing transparency of internal company data and knowledge. Altana, winner of the 2023 Intelligent Spend Management (ISM) award sponsored by Roland Berger and SAP, developed a knowledge graph including GenAI features that connects internal data from ERP systems, such as spend categories, suppliers, and countries, with external data like category risk, supplier risk, CO2 impact or country risk.

In a next step, a variety of LLM applications are conceivable using networked data and knowledge. These include analyzing and optimizing processes (GenAI-conducted process mining), evaluating suppliers (inventory and potential new suppliers), and risk analysis. With GenAI, procurement can quickly analyze complex business activities and predict future market developments. It will empower CPOs to revolutionize procurement strategy development and implementation by enabling a comprehensive and dynamic evaluation of a company’s current state in alignment with relevant sales and procurement market dynamics.

The procurement endgame: Procurement as we know it will cease to exist

There are almost no limits to the uses of GenAI, and procurement organizations can develop LLMs into new “omniscient” digital colleagues. This requires new ways of working and thinking, which means the operating model must be completely transformed. Repetitive text- or document-based tasks in particular are well suited to LLM agents. This raises the question as to whether procurement should be set up as an efficient machine for handling standardized tasks or fill a key strategic role as a value generator. Ultimately, these aspects will dictate the extent to which activities performed by humans are taken over by technology.

What is certain is that fewer resources will be required for category management, sourcing, negotiation, and operational process execution tasks, as well as for supporting activities such as spend analysis, risk controlling and contract management. Traditional roles and functions must be adapted to the “new normal” in an AI-driven environment. Focus and tasks will shift and new skills will be necessary. Traditional roles such as category managers, tactical buyers or procurement controllers, for example, will evolve into strategic relationship managers, LLM operators, data scientists or prompt engineers to effectively leverage GenAI and eliminate AI hallucinations.

By embracing and fully integrating GenAI, CPOs can ultimately achieve better procurement results using only 50-75% of the resources. Procurement leaders must re-answer their “right to play” and redefine their destiny in an AI-powered environment.

The road to a GenAI-fueled procurement future

The use of GenAI in procurement promises to revolutionize ways of working through end-to-end process automation, creative idea generation, and fast and reliable task execution. To give procurement an AI boost, companies need to follow five steps:

• Identify target use cases: CPOs should not start by analyzing capabilities of AI technology and aligning new ways of working to it, but rather identify key areas of improvement along the procurement process to increase value proposition and efficiency and translate these into prioritized use cases.
• Create your AI strategy and environment: Based on identified use cases, companies should derive a stringent AI strategy, and create a corresponding AI environment cross-functionally. Consequently, the IT infrastructure must be reconfigured to build the metadata model as the foundation for effective, efficient, and compliant use of GenAI.
• Reshape the operating model: It must be identified what role procurement should play on the spectrum between "efficient machine" and "value generator". The former focuses on maximizing process efficiency, with technology taking over large parts of the function. Procurement as a "value generator" leads to new objectives and shifts in human procurement, including toward relationship management, innovation, sustainability, and supply chain excellence.
• Rebuild your workforce: CPOs must transform their teams to pursue new goals and use AI in a targeted way. This will include retraining and upskilling existing employees as well as recruiting new talent with the necessary skills and knowledge in areas like LLM deployment, prompt engineering, and data science.
• Intensify and expand relationship management: Procurement should strengthen relationship management to unleash additional value-creation potential. This includes internal relationships with business partners to strengthen cross-functional collaboration, and external relationships with key suppliers to improve supplier engagement.