I recently had the opportunity to participate in a morning discussion session organized by Numeum on the impact of artificial intelligence in B2B organizations. Among the many perspectives shared that morning, a statement by Emmanuel Obadia summed up what we observe every day working alongside the procurement teams we support at Oalia:
“AI isn’t the transformation. It’s the stress test for your transformation.”
Behind the excitement surrounding new tools, one thing is clear: AI alone does not transform companies. It reveals which ones have already built robust processes, reliable data, and differentiating business expertise. It acts as an accelerator but also as a revealer.
Buyers have changed faster than organizations
The figures shared this morning illustrate the scale of the phenomenon (source: Forrester Research, 2025):
- 94% of B2B buyers already use AI in their buying journey
- 83% of the buying journey takes place without any contact with a supplier
- Twice as many buyers consider generative AI to be a more important source of information than publisher websites, sales representatives, or even product experts.
An increasing proportion of purchasing decisions are now made even before a company engages with a prospect. The traditional journey : Search → Browsing → Comparison → Initial Contact has given way to a direct response model: the user queries an AI engine, which provides a summary. As it was put that morning:
“Yesterday, we were chasing clicks. Now, we’re chasing citations.”
For marketing teams, this shift means new priorities: building their authority, producing truly useful content, and becoming credible sources in the eyes of search engines. But this transformation goes far beyond the marketing function. It affects every level of the organization: sales, finance, operations, human resources, and, of course, procurement.
Widespread Adoption, Value Still Fragile
A second observation shared at this event particularly struck me: while 78% of companies report having widely adopted AI, only 6% say they achieve a significant return on investment in less than a year. At the same time, one-third of pilot projects never move beyond the experimental stage due to a lack of measurable results (source: McKinsey).This gap between adoption and performance comes as no surprise to anyone who works closely with procurement departments on a daily basis. AI is not a solution in and of itself, it amplifies what already exists. An organization with reliable data, well-managed processes, and recognized expertise will accelerate its progress further and faster. Conversely, existing weaknesses, such as siloed data, incomplete databases, and poorly formalized processes, quickly become apparent. This is precisely what we’re seeing at Oalia: the procurement departments that get the most out of AI are those that had already invested in the quality of their supplier data, the structuring of their categories, and the formalization of their decision-making processes.
The challenge isn’t AI. The challenge is decision-making.
This is probably the most formative lesson of the morning. Technology has value only when it tangibly improves decision-making. In procurement departments, the reality is particularly complex. A growing number of suppliers, increasing regulatory constraints, CSR requirements, risk management, and cost pressures: teams must balance sometimes conflicting objectives on a daily basis. In this context, the promise of AI is not to replace human expertise, but rather to help buyers access relevant information more quickly, identify weak signals, and inform their decisions. But as systems become capable of making recommendations, anticipating outcomes, or even acting autonomously, a fundamental question arises:
- To what extent can we delegate analysis without delegating decision-making responsibility?
- How can we ensure the quality of decision-making?
- How can we prevent algorithmic bias or incomplete data from influencing a strategic decision?
This is precisely one of the key questions at the heart of the emergence of agent-based AI and a topic we analyze in detail in a dedicated article. “Agent-Based AI: What Are the Risks for Procurement Decision-Making?” The more tools advance, the more human judgment becomes strategic. In fact, the word that resonated most during this morning’s session was not “automation,” nor even “performance.” It was “humanization.” The more technology advances, the more the value of human expertise, discernment, and judgment becomes decisive. For procurement departments, this means one thing: AI can process information, but it is the buyer who makes the decision and assumes responsibility for it. At Oalia, we design our solutions based on this conviction: AI is a powerful tool when it is integrated into controlled processes and supports teams that retain full decision-making authority. Rosalie Narcy, Marketing Director Oalia 