A 3D bar chart intersected by a dotted trend line, symbolizing the interpretation of complex data over time. The vertical bars rise and fall in varied patterns, marked with numerical values to indicate fluctuations.

Agentic should-cost modelling is changing procurement and product design

Procurement teams can now negotiate, uncover value and inform product development by delivering cost intelligence at unprecedented speed and scale

Linas Polteraitis, chief supply officer at Terex Corporation, a US-based specialised equipment manufacturer, shares how, early in his career in procurement, he struggled to choose between two competing suppliers because he did not have a should-cost tool to help him understand what the product should actually cost. “I was contending with people who had 30 years’ experience in the industry, trying to pick the brains of one to negotiate with the other.” 

He contrasts this with the situation today, with AI levelling the playing field by providing procurement professionals with the knowledge required to support negotiations. Although should-cost modelling has long helped procurement teams understand the underlying economics of what they buy and what it should cost, AI is transforming the process into an even more effective tool.

Should-cost modelling is ripe for modernisation as the traditional process is slow, according to Gauranga De, vice president, consulting, at GEP. “You need a subject-matter expert and a process expert to build a should-cost, and much more beside. It takes time,” he explains. What is more, all the inputs are constantly in flux, meaning they need to be constantly updated – which is not possible in a heavily manual process. So, more time, more cost, and the potential for more errors when things fall through the cracks. “Even relevant dataset for the back cover of a phone can take three to four days to source and verify,” Mr De says, by way of example. 

Futuristic abstract background featuring flowing digital data waves in blue and green colors combined with financial charts, graphs, and HUD interface elements

Because it is so labour- and cost-intensive, traditional should-cost covers only the most expensive items, leaving mid-spend savings unexplored because the return on investment does not justify the effort. “If you have, say, 40,000 line items in a complex production process, you can only do should-cost for, at most, less than 1% of that,” Mr De explains. “That leaves an unexplored substantial opportunity ‘hump’ outside the items should-costed. And the omission of these can lead to significant slippage between estimates and actual costs.”

Gauranga De headshot

"The costing of a complex part could have taken ten days. Now it takes ten minutes."

Gauranga De, vice president, consulting, GEP

The advent of agentic should-cost means that the items in this additional opportunity ‘hump’ can be incorporated into the model, which itself is constantly rather than periodically updated. Human subject-matter experts input their knowledge and AI agents constantly bring that knowledge to bear on the task at hand. “The costing of a complex part could have taken ten days,” says Mr De. “Now it takes ten minutes.”

This means that should-cost is not just used more thoroughly on a specific product – it can be used on many more products than was previously practical.  The horizons of its use are being hugely expanded. “I don’t know that there’s a procurement application that’s not worth building an AI should-cost for,” says Mr Polteraitis. “There are very few cases where that wouldn’t be appropriate.”

Improving supplier relations

AI makes should-cost faster and more scalable, which is driving a more fundamental change: from periodic analysis to continuous intelligence. Rather than simply asking what a product or component should cost, organisations can use the insights these models provide to get a better understanding of the factors driving cost – from direct-material prices and manufacturing economics to labour and other inputs – how those factors are changing, and what that really means for business decisions.

First, it makes dealing with suppliers more transparent, says Mr De. “Once you go back with component cost detail, there is very little room for objection because the model showcases everything.”

Linas Polteraitis headshot

“In negotiations with suppliers, we can now ask ‘tell me what’s wrong?’ and we can correct it in real time.”

Linas Polteraitis, chief supply officer, Terex Corporation

Paradoxically, this leads to less antagonistic supplier relations. “In negotiations with suppliers, we can now ask ‘tell me what’s wrong?’ and we can correct it in real time,” says Mr Polteraitis, regarding assumptions about costs. “It leads to a very healthy situation, where we go from negotiating across the table to getting on the same side of the table and saying, ‘Okay, here’s what we see. Tell us where we’re wrong.’” 

GEP’s Mr De is even more optimistic about how agentic should-cost modelling is not only improving transparency with suppliers, but also helping companies to anticipate future costs. AI’s ability to model vast amounts of data makes should-cost predictive, he points out, and therefore potentially able to estimate costs months into the future. 

Two women have a professional discussion across a desk in a bright office.

Supporting better pricing in R&D

The impact of agentic should-cost resonates beyond procurement across the entire life cycle of a product. “The biggest challenge between R&D and procurement is that tensions arise between projected versus real cost,” says Mr De. With agentic AI models, the ability to constantly visualise the present and future costs of various alternative designs and material choices feeds into the R&D phase from the outset, rather than it being a case of first design, then cost. 

Mr Polteraitis confirms that Terex applies this pricing intelligence in new products to re-evaluate design assumptions against cost targets. Such a comprehensive view also leads to greater transparency when a product is in its ‘sunset’ phase, where agentic should-costing can highlight operating efficiency gaps.

A woman with curly hair and a man wearing glasses lean over to look closely at a white object on the print bed of a 3D printer. They are in a workshop setting, focused on the printing process.

The rollout of this capability may, of course, create some unease among procurement professionals. However, Mr Polteraitis is adamant that agentic should-cost does not negate the need for market expertise, but instead helps people throughout the business to understand the financial implications of design choices, making cost management a shared responsibility across the organisation, rather than just a procurement function.

Adoption is generally straightforward when tools empower users, and any friction typically signals underlying issues that precede and are unrelated to AI, believes Mr Polteraitis. Indeed, he says, “everyone was enamoured with the tool” developed with GEP.

While it is true that this will change the nature of procurement, professionals have to be “comfortable with being uncomfortable and acknowledge that it’s okay that they don’t know,” he adds. “They have to want to have that thirst for knowledge and that appetite for curiosity.”

Terex’s chief supply officer returns to his experience straight out of university, contrasting it to those now coming into the workforce, as his children will soon do so. “They’re trying to understand how AI is a threat. But in fact, it neutralises some threats. It’s less about how much you know, because there’s an abundance of knowledge. It’s more about synthesising and applying it.” 

As such, human expertise will always be necessary to provide context and validation. “This is an interesting lens to view AI through, which most people don’t think about,” Mr Polteraitis concludes. 

More from GEP

Start of carousel.

Article

More than the ghost in the machine: human agency in agentic AI deployment

The increasing uptake of agentic AI by procurement departments is not marginalising human roles. But it is changing them

Article

Agent of agents: working towards total agentic orchestration in procurement

As multiple agents collaborate across complex workflows, they increasingly deliver not just automation of routine tasks, but strategic business goals

Article Thumbnail | How ACR is crafting a strategic procurement function (1)
Article

How ACR is crafting a strategic procurement function

Before being reeled in by automation’s allure, high-performing procurement teams are upskilling

Article Thumbnail | How agentic AI enables agile procurement in an uncertain world (1)
Article

How agentic AI enables agile procurement in an uncertain world

The technology’s autonomous capabilities give teams greater value and influence

Article Thumbnail | How agentic AI is transforming procurement’s strategic focus (1)
Article

How agentic AI is transforming procurement’s strategic focus

The AI revolution sweeping across corporate functions, promising to increase efficiency by not just delivering but acting on radically improved data insights, is changing the fundamental nature of procurement.

Article Thumbnail | How Indorama Ventures built an AI strategy on long-term thinking (1)
Article

How Indorama Ventures built an AI strategy on long-term thinking

The petrochemicals producer has embarked upon an ambitious digital transformation journey

Article Thumbnail |  The AI bridge between finance and procurement functions (1)
Article

The AI bridge between finance and procurement functions

CFOs and CPOs are building strategic alliances in the war against costs

Article Thumbnail | Inside Caterpillar’s indirect-procurement reboot (1)
Article

Inside Caterpillar’s indirect-procurement reboot

The Fortune 100 manufacturer has enhanced visibility, efficiency, compliance and savings through strategic centralisation

End of carousel.

Contact GEP to learn more about supply chain transformations

Click here

More from Economist Enterprise

EE Logo (1)

Start of carousel.

How far will AI agents go?
Supported by GEP
Report

How far will AI agents go?

Examine how autonomous AI agents are reshaping procurement and supply chain operations. Based on an Economist Enterprise survey.

The transformative role of supply-chain leaders in today’s business
Supported by GEP
Report

The transformative role of supply-chain leaders in today’s business

Explore how executive supply chain roles are evolving to manage geopolitical risk, disruption, and modern procurement strategies.

The current and future use of gen-AI in supply chains
Supported by GEP
Report

The current and future use of gen-AI in supply chains

Discover how global enterprises are adopting generative AI across procurement and supply chain workflows to boost efficiency and insights.

Reskilling procurement teams for the age of agentic AI
Supported by GEP
Report

Reskilling procurement teams for the age of agentic AI

This Economist Enterprise report which explores the reskilling challenges for procurement teams in the age of Agentic AI.

Greener supply chains need smarter AI systems
Supported by GEP
Report

Greener supply chains need smarter AI systems

Learn how artificial intelligence helps enterprises build greener supply chains and overcome environmental, social, and operational challenges.

The agentic AI implementation challenge
Supported by GEP
Report

The agentic AI implementation challenge

The Agentic AI Implementation Challenge explores the challenges of adopting artificial intelligence in procurement and supply chain.

End of carousel.