Autogmentation Part 5: A detailed examination of GenAI features in procurement tech

Procurement has not been immune to the latest evolution in technology. While initial automations initially increased process velocity by taking over more tactical, repetitive tasks, they gradually evolved to cover more complex ones. At first, these only required more analytical activities to augment human workers, but they eventually evolved from predictive and prescriptive analytics towards more embedded and autonomous analytics that directly help execute the workflows.

Procurement activities and tasks are operational, tactical and strategic. Therefore, depending on the technology used and the processes digitized, the impact of technology can be divided into automation and augmentation, with augmentation being more 'autogmentation' [/aw-tog-men-tay-shun/] as we call it. For example, traditional AI enables organizations to tackle more complex use cases and follow the path to 'autogmentation,' which focuses on making processes much more effective and impactful. It augments human intelligence and knowledge with AI as a 'force multiplier' and not simply on increasing efficiency through process automation.

Generative AI (GenAI) represents another step in realizing AI's power and opening the door to new or enhanced possibilities in general for driving more value into processes in terms of user experience, efficiency and effectiveness.

For example, large language models (LLMs) are a central component of GenAI. They are good at ingesting and 'understanding' large data sets by uncovering deep data patterns that help predict what the user is trying to infer. However, the real power comes not from building an LLM from a massive corpus of internet data to create a context-free general-purpose chatbot. Instead, it comes from how rich highly-abstracted data-prediction models that human intelligence has built and trained and are increasingly training themselves to squeeze more value out of existing data and systems. It can enhance domain-specific structured business data — and the apps and experiences that deterministically use that data via algorithms and business logic.

The different approaches to using GenAI in procurement technologies reflect how providers have been responding to the potential of this new technology and the growing customer demand for GenAI features. We detailed these approaches in our recent article on the "six different 'flavors' of GenAI implementations in procurement applications.

We also covered several use cases and examples in our 'autogmentation' series (part 4) to illustrate the possibilities, limitations and challenges GenAI represents for procurement.

But, things change fast in AI, and several additional vendors have implemented GenAI since our coverage at the end of 2023. Therefore, we want to provide readers with an update. Let's look at more use cases from briefs and demonstrations we received.

This piece covers the following vendors: apexanalytix, Agiloft, Archlet, ContractPodAi, Evisort, Fairmarkit, GEP, Globality, HaloAl, Icertis, Ivalua, Malbek, McKinsey (Spendscape), Mercanis, Oro Labs, Relish, Rosslyn, SAP Ariba, Scoutbee, Sievo, Sirion and Zivio.

The examples represent the following segments: AP Automation/Invoice-to-Pay, Carbon Management, Contract Lifecycle Management, E-Procurement, Procure-to-Pay (P2P), Source-to-Contract (S2C), Source-to-Pay (S2P), Sourcing, Spend and Procurement Analytics, Supplier Management (SxM) and Supplier/Third-party Risk Management.

And we detail the following GenAI use cases:

Conversational user experience or CUX (chatbots 2.0)

Research and analysis (category management, contract analysis, etc.)

Creation of ‘new’ content like messages, sourcing events, etc.

Summarization of documents and information