It is difficult to talk about artificial intelligence without asking ourselves whether AI will change the way we work. When it comes to procurement, the answer is already quite clear: it certainly will.
AI can already analyze large amounts of data, compare documents, identify patterns, suggest suppliers, and flag discrepancies. According to a 2025 McKinsey study, 40% of procurement functions covered by the research have already implemented or piloted generative artificial intelligence.
But there is one thing that gets somewhat less attention: AI can be very smart, but if you give it messy data and disorganized processes, it will not suddenly perform miracles. This is exactly where we come back to something that may not be as attractive as artificial intelligence, but is still crucial for procurement – good procurement management software.
AI can find the answer. But first, it needs somewhere to look.
Imagine asking AI: “Which of our suppliers have contracts expiring in the next three months, and where do we have the greatest risk?”
We might say that this is a very simple question, but what if the contracts are stored in several different places? What if some of the data is in Excel spreadsheets, some in the ERP system, some in emails, and some in folders that only a few people can access? What if we do not have a reliable record of all our contracts?
The problem AI faces then is not that it is not smart enough. The problem is that it does not have a reliable source of data. This is one of the important lessons that comes with bringing AI into procurement: AI does not solve data disorder; it can only analyze it faster.
Among the key challenges of applying AI in public procurement, the OECD specifically highlights data quality, system interoperability, lack of skills, as well as issues of transparency and accountability. In other words, before we give procurement AI, we need to give it a good data foundation.
This is where procurement software comes in
Good procurement software is not a replacement for AI, nor is it its competitor. It is the foundation on which AI can become truly useful.
When requisitions, procurement processes, bids, suppliers, contracts, purchase orders, and invoices are connected within a single system, an organization gains something extremely valuable for AI: structured and connected data.
When data is connected, we can start asking much more interesting questions, not just “How much did we spend?” but “Why did we spend that much?” Instead of “Which supplier is the cheapest?”, the question becomes “Which supplier brings us the greatest long-term value?” And instead of “How many contracts do we have?”, the question becomes “Which contracts have the largest discrepancies between contracted and realized values?”
This is where AI can make a major difference.
We do not need AI to make decisions instead of procurement
AI is sometimes discussed as if it will soon independently run the entire procurement process, but that is not going to happen. It can find a supplier, send an inquiry, analyze bids, select the best one, place an order, check an invoice, and so on, which sounds impressive.
But in real business, things are not that simple because the best offer is not always the one with the lowest price.
The supplier with the lowest price may have problems meeting delivery deadlines. Another may be more expensive but never be late. A third may offer better quality but be financially unstable. A fourth may be reliable, but the company may have too little information about them. These are decisions that simply require context. And that is exactly why the goal should not be for AI to replace the procurement professional. It should free them from some of the tasks that should not be taking up their time.
What if AI could do what currently takes a procurement professional hours?
Let us say a procurement professional wants to analyze spending in a particular category. Today, they may have to:
extract data from the ERP system
transfer it to Excel
combine several spreadsheets
check suppliers
compare prices
find old contracts
check purchase orders
search for invoices
manually connect the data
only then start analyzing what is actually happening.
AI can significantly speed up part of this work, but only if the data is available and connected. If it is not, AI is not a magic wand – the procurement professional still has to find the data first.
And that is precisely one of the reasons why quality procurement software will remain important even in a world where AI is becoming increasingly powerful.
Software solves one problem. AI opens up another level.
Procurement software should primarily help an organization organize its process.
So that it is clear:
who requested the procurement
what is being purchased
who approves the request
which procurement procedure was conducted
which suppliers participated
which offers were received
why a particular offer was selected
which contract was concluded
how much of the contract has been realized
which purchase orders were issued
which invoices are linked to the procurement.
This may not be the most exciting part of digitalization, but it is exactly what creates order.
And when there is order, AI gets room to do what it does best – analyze large amounts of information, identify patterns, and provide recommendations.
Example: AI alerts, software connects
Imagine that AI notices something unusual, such as the price of a particular item from one supplier has increased by 18% compared to the previous period. This is useful information, but what happens next?
If we have a good procurement system, we can immediately see:
previous purchase orders
the contract with the supplier
the contracted price
actual spending
previous offers
related invoices
other purchases in the same category.
AI is then no longer just a tool that says, “Something has changed.” It can help us understand what has changed, how much it is costing us, and where we need to take action.
The decision still remains with procurement. This is a much more realistic and, in our opinion, much more useful model for applying AI.
The same applies to contracts
One of the specific examples cited by McKinsey shows how useful AI can be when working with business data.
In one case, a large pharmaceutical company used AI to compare invoices and contracts and identified more than $10 million in potentially lost value. This is an excellent example. But imagine how difficult it would be to perform such an analysis if the company did not have a centralized record of its contracts and invoices. AI cannot compare documents it does not know exist. It cannot link an invoice to a contract if that connection has not been recorded. It cannot flag a discrepancy if it does not have information about what was agreed.
That is why process digitalization is not the opposite of AI. It is its prerequisite.
And what about the human factor?
This is where we come to the part that is often lost in conversations about technology. Procurement is not just a collection of data. It also involves supplier relationships, negotiations, risk assessment, business understanding, and experience.
A procurement professional may know that a particular supplier saved the company two years ago when competitors were experiencing delivery problems. They may know that a supplier looks excellent on paper but is not reliable in practice. They may know that a particular price cannot realistically be compared with another because different conditions are included behind it. These are insights that cannot always be reduced to a single spreadsheet. That is why the best scenario is not AI instead of the procurement professional.
The best scenario is: AI + quality data + a digitalized process + procurement expertise.
Will procurement software then become less important?
Probably quite the opposite. The more capable AI becomes, the more important a high-quality data foundation will be. We can compare it to a car. AI is like a highly advanced driver assistance system. It can recognize a situation, warn about a problem, and help with decision-making. But if the car does not have good sensors, the systems are not connected, or the data is inaccurate, even the best algorithm will not be able to do much. The same applies to procurement. AI is a powerful tool. Software is the infrastructure. The procurement professional is the one who makes the decision.
That is why the future of procurement is not “AI or software”
This is not a choice between two technologies. The real question is: How do we connect them?
Software should ensure that the procurement process is structured and that key information is available in one place. AI can then help extract insights from that information that a person might not identify quickly enough. And the procurement professional decides what that information means for the business.
This is a model in which technology does not replace people, but gives them a better foundation for making decisions.
And perhaps that is the most important change ahead of us
The procurement professional of the future will not necessarily be the person who knows the most Excel formulas. Nor will it be the person who can manually compare the largest number of offers. And they will not necessarily be the person who knows how to use every new AI tool best. They will be the person who knows how to ask the right question, understand the data, and make a good decision.
And to do that, they need quality data, an organized process, and a system that allows them to see the full picture. That is why the question of the future is not: “Do we need AI if we already have procurement software?” Nor is it: “Do we need software if we have AI?” The real question is: “Is our procurement function organized well enough for AI to do what we expect it to do?” If the answer is yes, AI can become one of the most valuable tools procurement has. If it is not, perhaps the first step is a little less spectacular. Organize the data. Connect the processes. Digitalize procurement.And then let AI do what it does best.
Because the goal is not for artificial intelligence to make more decisions instead of procurement. The goal is for procurement to make better decisions – with the help of artificial intelligence.
