The One Question Procurement Still Can’t Answer: What Are We Actually Buying?
For all the technological progress of the last decade — automation platforms, Supplier Management suites, AI copilots, category tools, risk dashboards — Procurement remains stuck on the same foundational problem it has always had:
What are we buying? From whom? At what price? And when?
These four questions define Procurement. But they are not all equal. Not even close.
Because if you don’t know exactly what you are buying, the other three don’t matter.
And this is where Procurement’s real transformation must begin.
The uncomfortable truth: Supplier Management evolved, but item identification did not
For 20+ years, organizations have heavily optimized one side of the equation: “From whom should we buy?”
It’s why the market is full of excellent solutions for:
Vendor Management, Risk scoring, ESG tracking, Onboarding workflows, Performance reviews, Supplier discovery, Compliance monitoring….
Supplier data has become structured, enriched, and increasingly automated. And Procurement teams have adopted these tools well.
But the other side — the most essential side — has remained dramatically underdeveloped:
uniquely identifying the material, product, service, or part being purchased.
Because without knowing what an object truly is, every downstream process is compromised:
category management, sourcing, contracting, catalogs, P2P, planning, inventory, spend analysis, automation, AI agents…
It all breaks — or at best, becomes unreliable.
Why “WHAT you buy” is the only blocking problem
Let’s be clear:
Misidentifying the supplier ➡️ causes inefficiencies.
Misidentifying the price ➡️ affects savings.
Misidentifying the timing ➡️ affects planning.
But misidentifying the object itself ➡️ blocks the process entirely.
If you don’t know what you are buying, you can’t:
source it, benchmark it, standardise it, compare it, automate it, plan it, store it, reorder it, or let an AI agent take action on it!
Every single capability in Procurement and Supply Chain requires one thing above all:
a unique, unambiguous, machine-readable identification of the object.
AI agents magnify the problem — they don’t solve it
With the new wave of agentic automation, the stakes have become dramatically higher.
AI agents don’t “interpret” uncertainty the way humans do. They need:
clear object identity complete logical attributes harmonized structure enterprise-wide consistency
If you feed an AI agent a material that is unclear, mislabeled, duplicated, or poorly described, the error becomes catastrophic at scale.
One wrong assumption → thousands of wrong micro-decisions.
This is why machine-consumable data matters more than ever. And why item identification and categorization come before everything else.
The minimum attribute set: the missing blueprint in Procurement
Another blind spot in the industry is the lack of standards for:
What is the minimum set of attributes Procurement needs to buy anything without errors?
For many organizations, this remains undefined. Different plants choose different descriptions. Different teams use different taxonomies. Different systems interpret attributes differently.
The result?
Humans patch inconsistencies manually. AI agents fail silently. Automation stalls. Savings leak. And complexity grows.
Identifying the object — consistently, unambiguously, and at scale — is what unlocks every other optimization.
Why Creactives focuses on the part everyone else ignores
Most Procurement technologies assume the data is correct.
Most MDM projects try to clean it once.
Most AI vendors rely on probability instead of domain logic.
Creactives does the opposite.
We solve the problem at its source, thanks to our Supervised AI, a unique combination of:
✔ Machine Learning trained on 20+ years of Procurement & Supply Chain data
✔ Knowledge Engineering that embeds domain rules, semantics, and taxonomies
✔ Human domain experts who validate, refine, and evolve the logic as industries change
This hybrid approach ensures something general-purpose AI can’t deliver:
accuracy, explainability, reproducibility, and contextual truth.
It is why Creactives can:
categorize at granular level
identify materials and services uniquely
harmonize attributes across systems
align naming logic across the enterprise
govern new data continuously.
This creates the one thing every AI agent, P2P system, and sourcing workflow desperately needs:
a reliable, stable, machine-consumable Data Foundation that tells the organization — and its machines — exactly what each object is.
Why this changes everything
Once the “WHAT” is correct:
AI agents can execute sourcing autonomously
Forecasting becomes accurate
Catalogs become searchable and maintained
P2P no longer depends on tribal knowledge
Duplicate materials disappear
Negotiation strategies improve
Risk becomes contextual
Spend analytics finally reflect reality.
Machines can act with confidence. Humans can trust the output. The entire enterprise accelerates.
Conclusion: The real revolution is not AI — it’s the ability to know exactly what you buy
Procurement has evolved in every direction except the one that matters most. And as organizations rush into agentic automation, this blind spot becomes a structural risk.
AI will not fix it. New platforms will not fix it. More supplier data will not fix it.
Only one thing will:
the ability to uniquely identify, categorize, and harmonize every object the company buys — continuously and at scale.
This has always been Procurement’s most fundamental challenge. It is now its most strategic one. And it is the exact capability Creactives has been building for 20 years — through a unique blend of Machine Learning, Knowledge Engineering, and domain expertise.
Because before AI can transform Procurement,Procurement must finally know what it is buying.