AI in Procurement: Start With the Problem, Not the Technology
AI can create significant value in procurement. But adopting AI isn't a strategy.
The rapid development of generative AI and intelligent automation is creating new opportunities across procurement — from analysis and market research to contract review, sourcing support and routine workflow automation.
The temptation is to start with the technology: identify an AI tool and then look for places to use it.
A better starting point is the opposite.
What procurement problem are we trying to solve — and is AI the right intervention?
Organisations that answer that question first are more likely to identify practical use cases, manage risk appropriately and invest where AI can genuinely improve procurement performance.
Start with the problem.
Procurement functions rarely suffer from a shortage of technology. More commonly, the underlying challenges involve fragmented processes, inconsistent data, manual effort, unclear accountabilities or capability constraints.
AI may help address some of these issues. It may also simply automate a process that should first be simplified, redesigned or stopped.
Before selecting a technology, procurement leaders should be able to clearly define:
• the problem or opportunity
• who is affected
• the current process and its limitations
• the value of improving it
• the risks involved
• what a better outcome would look like
Only then does the technology question become useful.
Where AI can genuinely help.
The strongest AI opportunities in procurement tend to be activities involving large amounts of information, repetitive analysis, document-intensive work or tasks where people spend significant time finding, organising and interpreting information.
Practical applications can include:
Market & supplier intelligence
Accelerating market research, comparing supplier information and helping procurement teams identify relevant market developments.
Contract review & management
Extracting obligations, identifying key clauses, comparing contractual positions and helping teams monitor important dates, commitments and risks.
Routine procurement work
Drafting first versions of documents, summarising information, preparing meeting material and reducing repetitive administrative effort.
Spend & category analysis
Analysing expenditure, identifying patterns, highlighting anomalies and supporting category planning.
Sourcing support
Assisting with requirements, market documentation, evaluation frameworks, supplier questions and analysis of responses.
Commercial analysis
Supporting analysis of pricing, proposals, variations, supplier performance and other commercial information.
The objective isn't to remove procurement judgement. It is to use technology to reduce low-value effort and give experienced people better information, more quickly.
AI doesn't replace commercial judgement.
Procurement decisions often involve ambiguity, competing objectives and consequences that cannot be reduced to data alone.
AI can analyse information, identify patterns and generate options. It cannot take accountability for the decision.
Experienced human judgement remains particularly important when:
Commercial trade-offs are required
The lowest price, highest score or apparently optimal answer isn't necessarily the best commercial outcome.
Negotiation matters
Successful negotiation requires judgement about relationships, leverage, timing, behaviours and the interests of the parties.
Risk is material
High-value contracts, sensitive information, critical suppliers and significant business risks require appropriate human oversight.
Probity and fairness matter
Procurement decisions must remain explainable, defensible and consistent with applicable governance and evaluation requirements.
Context changes the answer
AI can work from the information available to it. Experienced practitioners understand organisational context, stakeholder dynamics and consequences that may not appear in the data.
The goal should therefore be AI-enabled procurement, not AI-autonomous procurement — combining technology with appropriate human judgement, accountability and control.
Readiness comes before implementation.
Identifying a promising AI use case is only part of the equation. An organisation also needs to be ready to implement it.
Before progressing an AI initiative, procurement leaders should consider five areas:
Process
Is the underlying process clear, consistent and worth improving?
Data
Is the information required by the use case available, reliable and appropriately managed?
Technology
Can existing systems support the use case, and how will AI interact with the broader procurement technology environment?
Governance & risk
Are there appropriate controls for confidentiality, privacy, security, accuracy, bias, decision-making and accountability?
People & capability
Do users understand how to work effectively with AI, challenge its outputs and recognise when human intervention is required?
An attractive use case with poor underlying readiness can create more cost, complexity and risk than value.
A practical approach to AI in procurement.
Procurement leaders don't need an enterprise-wide AI strategy before they can begin exploring the opportunity.
A more practical approach is to start small and build from evidence:
1. Identify the problem
Start with a genuine procurement or commercial problem — not a technology looking for a use.
2. Assess the value
Determine whether solving the problem could materially improve performance, decision-making, risk or productivity.
3. Test whether AI is appropriate
Consider whether AI offers a genuine advantage over improving the process, using existing technology or changing the way the work is performed.
4. Assess readiness and risk
Understand the process, data, technology, governance and capability requirements before committing significant investment.
5. Pilot, measure and learn
Test promising applications in a controlled environment, define the outcome you expect and measure whether the technology actually delivers it.
Then scale what works — rather than committing to technology and hoping the value follows.
Start with the problem.
AI will undoubtedly become an increasingly important part of procurement. But successful adoption won't be determined by how quickly organisations deploy new technology.
It will be determined by whether they apply it to the right problems, with the right foundations, controls and human judgement.
For procurement leaders, the question isn't simply “How can we use AI?”
A better question is:
“Where can AI genuinely improve the way we work — and are we ready to use it well?”
LocSam's AI Procurement Readiness review provides an independent assessment of potential AI use cases, organisational readiness, risks and practical next steps.