Procurement Intelligence: How AI Is Helping Contractors Make Smarter Commercial Decisions

The UK construction sector is under sustained cost and schedule pressure. While the out-of-control materials cost volatility of the pandemic era is, thankfully, behind us, the industry is still experiencing significant issues around rising labour costs and supply-chain volatility. According to an April 2025 TMHCC report, the number of open vacancies in construction remains stagnant, meaning companies cannot fill vacancies. At the same time, “wages continue to increase, thereby reducing companies’ profitability.”

The trend is set to continue, with the BSCIS construction industry forecast for 2Q 2025 to 2Q 2030 that tender prices will rise 15% during this period. In short, contractor labour costs will continue to be one of the key factors in squeezed margins, which leaves even less room for error at the procurement stage of the building process. And yet, the way procurement is structured has not kept pace with an increasingly digitised and data-driven industry.  

The stark fact is, procurement remains one of the least digitised workflows despite its commercially critical status. With 80% of project budget committed within the first few months, the lack of structured data, real-time insights, and automation means commercial teams are making high-stakes decisions without full visibility.

The solution? Leading contractors are now turning to AI-powered procurement intelligence to replace admin-heavy processes with faster, smarter, and more strategic decision-making.

Alastair Blenkin, Founder & CEO of ProcurePro, has firsthand experience implementing AI in construction procurement across tier-one contractors in ANZ and the UKI. We asked Alastair to help elucidate the key concepts around AI in procurement and detailed a vision for how technology is transforming the procurement process. 

What Does “AI in Procurement” Actually Mean?

Everyone understands the importance of digitisation and there is a vague awareness of AI’s growing capabilities to collect and analyse data, but how does AI help in the context of construction procurement specifically? Alastair summarises the way AI impacts the procurement process: as “AI improves access to comparable pricing information.” 

The Role of Structured, Historical Request for Quotation Data (RFQ) Data

One crucial thing AI allows in procurement is historical price comparison. In a typical scenario, a main contractor will source multiple quotes and then make a comparison between the quotes that come back. But, as Alastair points out, “Currently, the price comparison process is limited to a current price comparison where the contractor goes, “which one's the best value, which one's the lowest cost? But prices A and B could still be 15% higher than received on another project at a similar time.”

By making historical price data readily available, “AI makes it easier to connect the dots, across projects, across different subbies, across time”, encouraging a commercial director or lead contractor to think relatively rather than focusing on a single project at a time. 

Predictive Analytics Means Better Pricing Decisions

Looking to past pricing to negotiate better is one important aspect of AI’s reshaping of procurement behaviour. The other is looking to the future. With every new project being estimated, contractors should be looking at their rates/benchmarks around what prices they're procuring for. AI’s predictive analytics capabilities are especially pertinent where it comes to gauging pricing that’s relevant to specific projects. 

While AI does not have the capacity to predict exact pricing fluctuations, it can give construction professionals valuable access into longitudinal cost data. Tracking cost trends over time ultimately will help firms achieve greater insights into where the market might be headed.

Costs can become a problem, too, simply when the procurement process is too slow and the material pricing changes before buy-out. AI speeds up the process overall, which helps hedge against sudden spikes in materials pricing. For example, on average ProcurePro customers save 50% time in reducing procurement admin. 

Price Benchmarking: Improving Existing Processes

Finally, price benchmarking traditionally has involved huge volumes of data ‘buried in spreadsheets’. This isn’t something AI can - or should - do away with, because spreadsheets are used for these purposes with good reason: there is a lot of nuance in variation in procurement depending on the type and scope of the project: “When you're buying a subcontract and your comparison of different subcontractors could be 20 lines long, could be 2,000 lines long”, Alastair points out. 

AI can assist by adding more structure and accessibility to how data is organised and stored, but the point of integrating AI tools is certainly not to get rid of existing processes.

The idea is to streamline complex workflows, not try to simplify aspects of procurement that can’t and shouldn’t be simplified. Streamlining is about accelerating - speeding up the access to and organisation of comparable data. If the typical benchmarking and outlier search looks somewhat like “trying to line up the prices, then going where there's gaps, plugging numbers in, and then comparing them against other vendors”, AI plugs the numbers for the professional who then makes an informed decision. 

What this process does not involve is reducing price benchmarking to a ‘computer says yes’ scenario; instead, “AI gives procurement professionals time to apply their construction intelligence and their experience, rather than them spending time punching away in spreadsheets.” 

The best way to think of AI’s role in streamlining benchmarking is as an advanced time and data management tool, not as a replacement for judgment calls that can only be made by a commercial director or QS. 

Implementation Roadmap: Understanding the Data Maturity Curve

AI can assist firms in achieving data maturity - which is to say, in using data in faster, more integrated, and more productive ways. Alastair says that “A typical data maturity curve is a gradual transition: from “people working entirely manually, to fragmented systems connected, and then once you get connectivity, that's where you're creating the unique datasets that allow you better access to information and to unlock the power of AI.”

The fragmentation of data is the most common hurdle to overcome in creating connected workflows, not least because the assumption often is that connected systems have to all be connected up and synced perfectly. That’s never the case in reality, given that the way construction firms accumulate and process procurement data is historically disparate.

There’s often a major roadblock to implementing substantial workflow changes: people naturally resist too much change too fast, especially where the existing system is complex enough. Fortunately, achieving data maturity is not about ripping up old ‘disconnected’ systems and replacing them with brand-new, ‘connected’ ones. 

Instead, achieving maturity can be done more holistically, by looking at current ways of working and identifying areas that can be better moved from manual to fragmented to connected. Alastair describes the typical process as examining the ways people sending tenders, doing comparisons, approvals, contracts, signatures, managing their supply chain, and the level of connectivity between them.

Pitfalls & Limitations: Concerns About Data Quality and Bias in AI Models

AI use is not without its sceptics in any industry, and construction is no exception. No AI model will give 100% accuracy, and some professionals may be reluctant to adopt the new technology for fear of introducing even more erroneous judgment into the process. Alastair believes that what can, ultimately, convince construction professionals that AI is worth the hype, is one simple thing - proof: “The AI does the first-pass analysis, and if the commercial director looks at that analysis, the devil's in the detail and the Quantity Surveyor proves it for themselves.” 

The emphasis is very much on assisting human expertise - which is why, Alastair explains, you don’t want an overconfident or an ‘unbiased’ AI: “We're sort of working to a 90% AI accuracy benchmark. Now, do we want 90% to become 99%? No. We actually want indicators that the AI is not confident in the answer because that directs human attention to go, hey, thumbs up, thumbs down, edit it.”

“You're constantly getting humans through their workflow to correct, shape, edit, train to improve the AI as opposed to going, ‘AI is going to completely do our job for us”, he adds.

One specific way to continue correcting and training AI to be really helpful to construction professionals is expanding and refining its knowledge of different types of projects, not just feeding it more and more of the same type of data:

The Future Landscape: Changing the Bidding Model from Price Taking to Price Setting

The really big idea in AI-boosted procurement is that it will shift the entire bidding model from price taking to price setting. Alastair explains what that would look like as a move toward a predictive bid model, similar to how an on-demand ride-hailing service dynamically suggests fares based on real-time supply and demand. 

The implications for the industry could be revolutionary. For one, procurement could become less reliant on the quote-collection stage of the process and more on the design stage, with high-quality BIM modelling at the forefront of the transformation. Alastair sees the potential for ‘good’ BIM design to become integral to accurately pricing a project at its earliest possible stages.   

The result of this design-based approach? A massive reduction of risk, which so far has been the accepted pitfall of construction procurement.

Practical Checklist: Questions to Ask Vendors

AI is still a little bit of a Wild West territory. It can be difficult to separate true software gold from marketing hype. What should construction professionals ask potential vendors before they adopt a new AI-enabled procurement management system?

Ask Yourself the Right Questions First

So, rather than being overwhelmed by promises and features, every procurement director should start with a moment of introspection and clear goal settings.

Try Before You Buy

Next, Alastair recommends trialing any new software before buying if at all possible.

Read Contracts Carefully

Any new system implementation will come with cost, training, the workload and timeframes involved. All of that should be in the contract. However, “If a contract is used to create stickiness and mitigate risk, then I'd be wary”, warns Alastair. 

Ask for Real-Life Use Cases

While AI technology is very new, and companies may not be able to provide longitudinal data on how well the new systems are improving margins, even a short-term trial use should take the effectiveness of the product out of the world of marketing and into a real-world context. "I would want to trial it and actually use it properly, not hypothetical trials”, Alastair stresses. 

Trust your experience, proof not hype.

Written by