Procurement is currently in the midst of its most transformative era ever.

Today, AI is reshaping business strategy, driving significant efficiencies and creating entirely new opportunities.

One company born from this shift is Opstream, an AI-powered procurement orchestration platform that streamlines procurement processes, empowers teams with self-service workflows and eliminates inefficiencies. Opstream’s mission is to enable procurement and finance teams to optimise purchasing workflows, ensure compliance and achieve cost savings without sacrificing speed or flexibility.

But despite the promise of AI, CEO and Co-Founder Lihi Lutan recognises the danger of treating it as a silver bullet or magic wand that fixes everything. 

In fact, to think of it that way would be a costly mistake, as Lutan stresses the value of focusing on fixing the process first instead of layering AI on top of antiquated operations holds the key. “In the past couple of years, as everyone started talking about AI, most people have thought about AI primarily at the front end of GenAI – asking questions and receiving replies in natural language,” she explains. “But I want to take a step back and say that we need to embed AI from the very beginning, before a human can even see it. If I go back to how we handle integrations, we use AI within the integration layer itself to do things intelligently and to enable deployment for new customers in a matter of weeks rather than months.

“AI is embedded there, but it also extends to the front end. How? Not only through natural language interactions, but also because every question you ask can be traced back through the integration layer. We can determine whether the answer exists in an ERP, a CLM, or a compliance system, not directly upstream, but integrated with upstream systems, and then use AI to provide those answers immediately.”

Mor Cohen-Tal, Chief Technology Officer and fellow Co-Founder at Opstream, adds that people often refer to AI as a new entity and disregard the previous forms that have come before it – but reflecting on the past can actually act as an advantage. “What’s new is this is the latest generation of AI. There’s a great deal we can learn from previous iterations as we adopt these new technologies,” she explains. “One analogy I like to use is GPS, because it’s one of the most successful examples of AI adoption in our daily lives. If you think back 20 years, getting from point A to point B meant opening a map and planning your route. Today, we don’t even think about it. Whether we use Waze, Google Maps, or Apple Maps, navigation has become second nature. Why? Because we trust it and don’t even think about it anymore.”

With this in mind, Cohen-Tal believes there are several learnings that can be drawn upon from the human adoption of navigation. These are using real-time data instead of outdated generic information, keeping a human-in-the-loop and maintaining explainability. “These principles are at the core of what we do at Opstream,” she adds. “We apply these three areas across all our processes so that we can provide this to our customers. AI that keeps a human-in-the-loop, that is explainable and that earns our customers trust while leveraging data in an intelligent way.”

Indeed, the potential efficiencies that AI can unlock is seismic. To go further, agentic AI is the latest craze dominating CPO agendas with agents front and centre of many conversations. But AI is not yet at the point where it can be left to make decisions without governance which is why the human-in-the-loop is still an essential component of the process. Lutan admits striking the right balance is an ongoing challenge but believes full automation should actually not be the end goal. 

“I don’t think AI should be responsible for 100% of the process,” explains Lutan. “The question should always be about how you combine it with a human-in-the-loop? First of all, our AI engine assigns a confidence score to each output. Our engine scores itself based on how certain it is about the answer. You can then set triggers based on that score. 

“For example, if we’re talking about invoice reconciliation and the invoice is a clearly typed PDF with a straightforward match, the system will assign a very high score, and you may decide that only one person needs to review it. But if it’s a handwritten invoice and the score is lower, you might need 10 different people involved to ensure everything is correct, especially if a large amount of money is involved. We don’t assume in advance where human oversight is needed or who that person should be. Instead, those decisions are guided by the AI’s own confidence score for the output.”

Beyond governance, Lutan believes another major challenge facing procuretech companies is the industry’s tendency to treat every organisation the same way. According to Lutan, this is one of the most outdated parts of procurement’s operating model and believes that it could prove costly to organisations. “That’s a big mistake,” admits Lutan. “Every company is a snowflake. Each has its own data, and they all care about different things. Traditional procurement, intake, and orchestration platforms assume that everyone cares about renewal dates and contract end dates, but they don’t account for the fact that every company has a different way of calculating and negotiating savings.

“For me, the next level is about how we personalise the tools and processes for each company. When I look at our customers, we see an average of 17 different dates being tracked per customer, and those dates vary significantly from one company to another. Most of the tools out there can’t support that because they are very rigid in their data model.”

Moving forward, Cohen-Tal believes the future of procurement revolves around procurement’s role as a driver of proactive value creation instead of its legacy perception of being siloed out of the way of the action. “It’s about how we change and become strategic to the business when they say ‘Oh, we need a new provider for something’ and are proactively leveraging third-party vendors strategically,” she explains. “We become strategic when we help mature businesses to leverage AI and select the right vendors. That’s when AI becomes this incredible opportunity for procurement.”

And as far as Lutan is concerned, she affirms an emphasis should be placed on learning AI but in a way that is specifically unique to individual companies. “Procurement professionals now need to understand how to procure AI tokens,” explains Lutan. “Five or 10 years ago, procurement teams needed to know how to manage cloud spend. Now, the focus is shifting to AI tokens. Master that. Don’t let IT control it because you should be the expert on AI tokens.

“Knowledge is readily accessible today, so use it. Be prepared, leverage AI, bring the first AI tools into the organisation, and help lead the transformation. I think many procurement practitioners have been told by their boss, ‘Go to this conference and find out about AI.’ Some people are worried they may be laid off because they believe AI will replace them. Don’t be afraid of AI – embrace it.”

Find out more about Opstream here.

For BAT, a leading consumer goods manufacturer, recoding procurement means going back to the foundations of how the function operates. Speaking at DPW New York, Eduardo Chrispim, Procurement Excellence and Digital Transformation Leader at BAT, said the event’s theme of ‘Recode’ resonated because digital transformation must extend beyond systems alone.

“We tend to think about recode in the technical meaning of the word, but for me it is much bigger than that,” he explains. “In digital transformation, it is about recoding and reconfiguring the culture, the people, the talent and the mindset. It is also about how processes are built, going back to the drawing board and redefining the ecosystem of solutions.”

That thinking is shaping BAT’s work with ORO Labs, as the business looks to rethink its procurement technology and create a more connected buying experience. Lance Younger, EVP, EMEA GM and Global Alliances at ORO Labs, sees ‘Recode’ as an indicator of a wider trend across procurement. “It is about going back to the basics, the foundations, and then rebuilding them again in a completely different way,” he says.

Rebuilding from the foundations

With procurement teams under pressure to move faster as markets become more volatile, this rebuilding work is becoming increasingly urgent. There’s a need to manage volatility while also managing risk and looking for new ways to innovate. There’s huge potential for new tools, particularly artificial intelligence (AI), to solve some of these challenges, but technology is just one element of the wider transformation.

“Procurement processes and platforms are extremely fragmented,” Chrispim says. “That deserves special attention in redefining how the ecosystem of tools should work together. We need to rethink the data flows, rethink the process, and rethink the intent of the process, why it exists in the first place.”

There’s often the temptation to add a new tool to solve an immediate problem, but this can lead to further fragmentation if the organisation doesn’t address what’s going on behind the scenes. 

Younger has seen this pattern before. “In years gone by, people have done a technical transformation and then wondered why it has not been successful,” he says. “Now more than ever, you have to reshape the operating model. That means the skills, how you deliver the service from procurement to the business, and how you deliver it to suppliers.”

Fix the process before adding AI

AI can be powerful for procurement teams, but Chrispim urges caution. He says that without looking at the underlying processes, “you’re basically scaling chaos.”  

“It is very important to step back and understand why the process exists,” Chrispim adds. “It is a good time to think about the problem and the solution. What are we trying to solve with the platforms?”

The AI conversation has changed significantly in recent years. The emergence of large language models means organisations can now work differently with structured and unstructured data, using semantics to interpret information in more flexible ways. However, those outputs are only as useful as the data feeding them. 

The same goes for automation. “Procurement should not be completely automated,” says Chrispim. “You should automate the most transactional and repetitive activities, such as tactical sourcing, supplier onboarding, and some supplier screening activity. But there are areas where we need the human element, including strategic partnerships and building relationships with stakeholders and suppliers.”

Both Chrispim and Younger agree that AI must be built on good foundations and shaped by proper governance and controls. If procurement teams are going to rely on tools to recommend decisions, they need confidence in the processes and data behind them.

Finding the right partner

This focus on foundations shaped BAT’s decision to work with ORO Labs. The relationship began a year ago, as BAT looked to rethink its technology stack and make deliberate choices about the partners that could support its procurement transformation. The value case was practical as well as strategic, with BAT looking to reduce friction across the buying and sourcing journey. 

“We wanted to partner not only with what a solution provider has now, but also understand their vision and how we can share that to become a best-in-class procurement function,” says Chrispim. “We found ORO Labs to be a leading organisation in the intake and orchestration space, which resonated with the problem we had in bringing more efficiency to the buying process, improving the buying experience, and improving the sourcing experience for procurement users.”

The alignment was also clear from ORO’s side. “There was a nice fit from day one with where BAT wanted to go, and we saw the ability of the teams to work together,” Younger says. “That all came together at the Procurement Leaders Awards where BAT won Procurement Team of the Year. Oro Labs won Top Procurement Technology Provider the same night. It’s almost the perfect replication of where we are now, but also where we are going to be in the future.”

Procurement’s changing role

Cost remains important, but BAT is pushing beyond a purely cost-centric model. Around 20% of BAT’s procurement team do not have savings or cost targets. Instead, they are focused on innovation, growth, and value creation.

“Sustainability is another area,” Chrispim adds. “Procurement is unlocking sustainability goals and commitments that organisations are making. Procurement is creating much more value than just cost, and that will continue to be amplified in the coming years.”

These priorities might vary across organisations. “What’s important is that procurement departments make a conscious decision about what matters to them,” says Younger, “whether it’s cost, risk or sustainability. Then they follow through on it with conviction.”

Turning ambition into action

That need for conviction is also what makes forums such as DPW New York so valuable. The event brings together a community of procurement leaders, practitioners and solution providers who are not only discussing the future of the function, but also actively working through how transformation happens in practice.

“DPW is a nice community of people that actually do things,” Chrispim says. “It is very tangible. You understand what solution providers are implementing, but you also meet peer companies and understand what is working and what is not from those actually driving the projects.”

For BAT and ORO Labs, the theme of ‘Recode’ reflects where procurement is today. Transformation must start with the foundations, as this will support faster decisions, stronger supplier relationships, and a function better equipped to deliver value beyond cost.

Today, the role of procurement is increasingly challenging. 

Leaders are being asked to unlock greater value, and often with fewer resources. 

It is a key reason why SpendHound was created – to give procurement teams the data, leverage, and confidence they need to shape purchasing decisions proactively.

SpendHound is a SaaS spend management platform that helps companies understand what they should be paying for software and AI. Companies use SpendHound to gain visibility into their technology investments, track renewals, and benchmark vendor pricing against peer organisations. As a subsidiary of YipitData,

SpendHound leverages benchmarking insights across more than 10,000 software and AI vendors to help organisations optimise spend, negotiate from a position of strength, and maximize their investments.

Tom Ruello, VP, GTM at SpendHound, explains that what his organisation has found is that procurement teams don’t necessarily have the resources to optimise spend but are asked to save a lot of money on their bottom lines. “What we’ve done is essentially flipped the model on its head, unlike other companies that will charge a lot of money and then struggle with really fast growth,” he discusses. “We have created a model where we provide a product that is 100% free for companies with less than 1,000 employees. For those companies with more than 1,000 employees, we charge just a $10,000 a year flat fee with a guarantee of $150,000 of savings. 

“The nice thing about the model is that it has allowed us to grow really quickly. In two and a half years, we’ve grown to more than 1,200 customers and now we’re close to 100% growth year over year. That growth has built this unique, proprietary benchmarking data set that gives power back to procurement. With access to stronger, democratised spend data, procurement gets that strategic voice internally within their business and improves their results.”

Becky Martins, VP Marketing at SpendHound, adds that it feels like a “no-brainer” to leverage SpendHound because of the low barrier of entry. “We have an amazing data set that is generated from real contract and peer spend data, which is what people want rather than just the sticker price from a vendor,” affirms Martins. “It’s something that I believe we do better than anyone else on the market.”

Martins affirms she has lots of ‘empathy’ for buyers because of the tension and conflict they are forced to deal with. “It’s a really tough job,” she admits. “You have a strategic charter, managing supplier relationships, spend visibility and other things that the business depends on, but yet you’re often measured from a savings and budget variance perspective only, and you really don’t have all the tools that you need to be successful.”

Ruello and Martins were speaking to us at DPW New York, where this year’s conference theme was ‘Recode’ which is the idea of redesigning how procurement and supply chain operates today. With this in mind, Martins believes that, given the transformation happening within procurement, this could be the function’s defining moment. She argues that this is where tools like SpendHound are on hand to help procurement leaders navigate change and deliver more value to their organisations. 

“At SpendHound, we want to be the ones to arm procurement leaders with actionable intelligence and defensible benchmarks so they can be more strategic thought partners to the business,” adds Martins. “Being honest, you can’t just get a seat at the table – you really have to earn it. The way in which you do that is being that default person that the organisation goes to at the beginning of a negotiation cycle instead of at the end, because you have the intelligence, insights and benchmarks that no one else has so you can guide the negotiation process in as strategic a way as possible.”

Ruello agrees with Martins and asserts that Recode is about essentially redefining how procurement actually operates and eradicating legacy ways of thinking. “There’s a bit of a moment in time right now where every single company’s talking about AI and trying to implement new agents,” he explains. “The reality is that companies are hiring fewer procurement people. Recode is that opportunity to think about how we can make sure we’re doing things in a way that’s going to provide that seat at the table so that we can be strategic drivers within the business. Candidly, it’s also about ensuring we’re not going to be replaced by those AI agents that everyone’s trying to sell to us.”

The noise around AI is loud. Attend any procurement conference and the dominant topic of conversation will revolve around various forms of AI and the opportunities it can unlock.

For Ruello, he admits the buzz is real but explains that the technology is only as good as the foundation it is built on. “A lot of companies aren’t necessarily investing heavily in building those foundations,” he says. “Everyone wants the impact and the value of the platform on top, but they’re not willing to invest in what’s actually required.” 

“There are two key pieces to that. Firstly it’s about the way that you clean your data. At SpendHound, we do that. We clean up all of your software data and provide a clean view of exactly what you’re spending money on. Then the other piece of that is enriching AI with other types of data and we think of SpendHound as very much a data company. We provide data and insights to power those agents to feed any of your AI processes you have in place. We’ve recently built an MCP server so we can push data to any AI platform you’re using to provide that flexibility. The reality is you’ve got to be making moves because the space is going to change whether you do or don’t.”

Find out more about SpendHound here.

Despite years of digital transformation, some procurement functions still heavily rely on manual processes.

Indeed, many companies are still comparing spec sheets and emailing suppliers for data. 

For Valdera, the needle has not moved far enough.

Founded by brother and sister Dheev and Sruti Arulmani, Valdera is the only AI-powered procurement technology built for chemicals and raw materials, used by category and sourcing leaders across pharmaceutical, personal care, CPG, and manufacturing. The platform unifies supplier discovery, market intelligence, and category management in one system, automating sourcing and qualification, exposing pricing and supply risk in real time, and giving teams a decisive edge on cost, speed, and resilience. Valdera is redefining how the world’s most demanding industries source the molecules they depend on. 

Dheev Arulmani explains that the chemicals and raw materials industry has an infinite number of specifications which means the use of AI is not just a nice-to-have but actually becomes essential. “If you look at indirect or even finished goods on the direct side, like pipes, fittings or equipment, there’s a finite SKU associated with each one of these individual projects,” he explains. “Chemicals and raw materials are a whole different world for a number of reasons. The main thing being that there’s an infinite permutation of specifications. You can have hundreds of different properties, chemical structures, regulatory requirements and safety requirements. What we realise is that you have to build the data infrastructure from the ground up specifically for that space and then you have to build the workflow on top of that, whether it’s agentic AI models or how materials are qualified and onboarded. You have to build every one of those steps specific to this industry and I think that’s the real enabler to drive value in this space.”

Given the level of innovation at companies’ fingertips today, Chase Macaione, Global Head of Sales at Valdera, believes the time is now for procurement to rip up the script as the entire workflow is too outdated in most places. “Fortune 500 companies are all running a lot of their workflow through email and spreadsheets and based off of 15-year relationships,” he explains. “I think one of the biggest reasons for this is supplier discovery. There’s a limited bandwidth in most companies for how many materials can be actively managed so people end up relying on the same two or three relationships that they’ve had for years. That means it’s really difficult to get the level of transparency that you need from the market to go and find new suppliers and people don’t have the time to do it in a lot of cases. Being able to use AI to take the details of the specifications and do that matchmaking is really powerful.”

Dheev Arulmani explains that the chemicals and raw materials industry has an infinite number of specifications which means the use of AI is not just a nice-to-have but actually becomes essential. “If you look at indirect or even finished goods on the direct side, like pipes, fittings or equipment, there’s a finite SKU associated with each one of these individual projects,” he explains. “Chemicals and raw materials are a whole different world for a number of reasons. The main thing being that there’s an infinite permutation of specifications. You can have hundreds of different properties, chemical structures, regulatory requirements and safety requirements. What we realise is that you have to build the data infrastructure from the ground up specifically for that space and then you have to build the workflow on top of that, whether it’s agentic AI models or how materials are qualified and onboarded. You have to build every one of those steps specific to this industry and I think that’s the real enabler to drive value in this space.”

Given the way in which AI is transforming companies of all different shapes and sizes, many organisations are eager to introduce new large language models onto their old, and often broken, supply chain processes. However, Valdera advocates for purpose-built AI mapped to individual sourcing workflows in the form of discovery, qualification and negotiation. Arulmani reveals he started the business with a data science lens and sought to build the world’s largest data set of chemical specifications, suppliers, catalogs, and packaging that could bring everything under one central hub.

“What has really evolved in the last couple of years is the agentic capabilities of large language models,” explains Arulmani. “What we’ve been able to do is take that core data that we’ve always had and apply this AI layer on top, which is arguably even more commoditised now. The real power is being able to integrate that data that sits underneath with the agentic capabilities of these models to drive value across that workflow for a category manager. But I think it is very hard to do if you don’t have that data underlying it. For us, that’s a real differentiator and many of the other players are focused on their own individual verticals.”

Macaione adds that one of the major areas overlooked is the non-deterministic nature of AI which makes it challenging to have it implemented in real enterprise processes. “You have to have determinism,” he affirms. “One of the things about being specific and vertically integrated means that we can guard against some of that non-determinism and especially when you talk about chemistry and buying chemicals because there are a lot of highly deterministic things that you have to get right. By building proprietary Retrieval-Augmented Generation (RAG) models, we’re able to guide and benefit from the AI while guiding it towards the right answer and consistent answers every time. That’s something that’s going to be really important, not just in our space, but for AI adoption and enterprise overall.”

For Valdera, improving efficiency in procurement is only one challenge made possible by AI. In the chemicals industry, data is of paramount importance. Buyers are protective of their formulas and do not want their data used to train models that could help rival companies. It’s something Arulmani recognises and he stresses there is no silver bullet in chemicals and raw materials. “I can’t go in and say, ‘Hey, I’m going to switch that overnight.’ For pharma companies, that could take two or three years to get approval but for some industrial businesses that might take several months, so the devil is in the details when it comes to raw materials,” he explains. “I think the real key here is how can you enable that workflow and how do you drive human relationships? How do you drive qualifications and onboarding that are enabled by R&D while digitising all the painful steps in that process, finding the suppliers, routing to them, estimating the logistics, tariffs, and comparing the specifications. How do you take those, we call them the dirty dishes of procurement, off the plate of a category manager so that they’re focused on the most strategic things that can drive the biggest value for the business?”

Today’s world is full of challenges and complexity unlike ever before. The past five years have been filled with disruption in the form of the likes of COVID-19, trade tariffs and wars – all of which have caused major headaches for businesses across a multitude of industries. While both Arulmani and Macaione admit they can’t stop these ‘black swan’ events from happening, being proactive is a large part of the puzzle in order to overcome issues and pivot quickly. “The key is being able to proactively eliminate some of that risk,” explains Arulmani. “You’re never going to be able to eliminate 100% but in those cases, the question becomes how can you be as agile as possible?”

Macaione adds, “These black swan events are not going to stop, they’re going to continue to happen so agility really becomes the name of the game. With a lot of materials out on the market, it’s a bit of a zero-sum game. If you’re not staying ahead of your competition from a technology standpoint to be as agile as possible, then you’re going to get left in the dust and endure business continuity disruptions which are bringing the Chief Procurement Officer to the forefront within a lot of organisations.”

Find out more about Valdera here.

Many procurement teams are successfully using artificial intelligence (AI) to surface insights and support decision-making, but this is only the first step. The next phase of procurement AI focuses on execution.

This is already underway at Ahold Delhaize USA . Speaking at DPW New York, Sean Sieper, Director within the Indirect Center of Excellence at Ahold Delhaize USA, explained how his team is using agentic AI. “We have begun piloting AI-supported negotiations with select suppliers,” says Sieper. “It is an important early step in how we’re exploring agentic AI responsibly, and we are encouraged by the results so far.”

The work is being delivered through Ahold Delhaize USA’s partnership with Pactum. Jim Rahill, Global Vice President of Sales at Pactum, explains: “We are delivering autonomous negotiations to Ahold Delhaize USA’s Indirect Center of Excellence, with the goal of achieving cost savings. We’re a couple of months into the relationship, and we’ve had great results so far.”

For Ahold Delhaize USA, the relationship marks an important milestone. “For us, this is about moving to practical application in a controlled environment within our team, with strong governance built in from the start. It is an important milestone for us,” says Sieper.

From recommendations to execution

That focus on doing, rather than only advising, is central to how Rahill sees the role of agentic AI in procurement. “The technology is advanced enough where it can take a data set and provide analysis or insights,” he says. “But when you involve a third party, such as a supplier in our case, that agent has to go out and do the work within the defined guardrails. Then they must bring it back and update their source systems.”

This is ultimately the promise of AI, but it’s also the most difficult thing to get right. Pactum’s execution layer is where autonomous negotiation creates real value. The aim is to identify opportunities but also act on them at scale to deliver cost savings. 

Aligning with DPW New York’s ‘Recode’ theme, Rahill sees this as an inflection point in procurement. “The world is shifting away from traditional enterprise SaaS software where companies purchase modules,” he says. “Now, agentic AI, the speed of development, and the time to value required from companies like Ahold Delhaize USA are such that you move faster to deliver value in weeks, not months.”

For Sieper, the difference is capacity. Generative AI can help people perform tasks more efficiently, but agentic AI can take on defined work so teams can focus elsewhere. “The real opportunity is helping our team scale routine activity so our associates can stay focused on strategic supplier relationships and higher-value work,” he says.

Building confidence through governance

When using AI agents to interact with a supplier, there need to be guardrails in place. The technology must operate within the rules and commercial boundaries set by the client while maintaining the supplier relationship.

This makes governance more important than ever. Not only does this protect relationships, but it enables Ahold Delhaize USA to show stakeholders how the supported negotiations are delivering value.

“Our supplier relationships are extremely important to us, and critical to our business,” Sieper says. “Transparency into how decisions are made is essential. It helps build confidence with stakeholders by making the process visible and understandable.”

For many stakeholders, this will be the first time seeing an AI agent in action. They’re seeing the real back and forth of an interaction to understand how the agent responds.

Rahill says that is often the moment when the value becomes clear. “When you see that completion, that the AI agent actually went out and did something versus just surfacing an insight, that is where the magic happens.”

Keeping humans in the loop

Both Pactum and Ahold Delhaize USA are clear that autonomous negotiation is still a journey. The point is not to remove human oversight, but to build confidence gradually.

“It is all about working with Pactum and putting in our guardrails for what we would expect from the negotiation, and then doing that at scale,” says Sieper. “Today, we maintain full oversight of each negotiation and interaction. Over time, as confidence builds and governance matures, the level of review can evolve appropriately.”

Pactum takes that governance process seriously, particularly when an AI agent is executing on behalf of a department for the first time. “It is a comfort level the client has to achieve before they reduce the amount of human engagement,” Rahill says. “At the beginning, we recommended it because we want to prove and show that this technology can be leveraged at scale.”

From pilots to outcomes

The pressure to prove value from AI is increasing. “We have cost, capacity and external considerations,” Sieper explains. “There’s a clear expectation to demonstrate measurable value, while ensuring we implement these capabilities responsibly and protect our supplier relationships.”

Rahill says that expectation is not unique to Ahold Delhaize USA. Across the market, procurement leaders are under pressure to show results, and to show them quickly. “It is no longer acceptable for it to take six months to implement and another six months for an ROI to be achieved,” he adds. “With the pace of change and how fast these tools are innovating, we need to prove value quickly.”

Even so, both Sieper and Rahill are cautious about overhyping the maturity of agentic AI in procurement. Sieper says the function still has “a long way to go”, with investment barriers, pricing structures and governance questions still to be worked through. Rahill also pushes back on the idea that agentic AI will make traditional enterprise software obsolete in the near term.

The more immediate challenge is understanding where to apply the technology responsibly. Pactum starts with questions such as which suppliers, categories, and opportunities can deliver value most quickly. Within procurement at Ahold Delhaize USA, it means starting slowly, building confidence, and then progressing as the organisation becomes more comfortable.

A forum for change

As any industry navigates technological change, there’s a need for practical, grounded conversations. This is part of what makes DPW New York so valuable. Sieper says the event gives procurement teams a chance to spend time with vendors and practitioners working through the same problems.

“People feel more comfortable moving from booth to booth and having meaningful conversations,” Rahill adds. “Professional and deep conversations, but in a more relaxed setting.”

For Ahold Delhaize USA and Pactum, that environment reflects the wider procurement AI journey. The opportunity is not simply to experiment with new tools, but to identify where agentic AI can deliver real outcomes, build trust through governance, and help procurement teams move from insight to execution.

Tail spend has been a nuisance for procurement functions for decades.

Often fragmented, cumbersome and individually low value spend, it’s traditionally been regarded as not worth the time and effort required to manage effectively. Yet, these small transactions quietly add up in the background, accounting for approximately a fifth of total procurement expenditure.

But as procurement continues to embrace an industry-wide transformation amidst a rapid deployment in AI across the board, the time is now to deal with the lingering challenge of tail spend.

Fortunately, Axiom is actively addressing this problem. 

Axiom has built a single ecosystem that handles everything from catalog purchasing to one-off supplier payments encompassing an ‘all in one place’ approach to make life easier for procurement teams. “The problem is that there are different component elements of tail spend,” explains Alistair Cane, Co-Founder and Director at Axiom. “There are elements that require sourcing, a robust catalog strategy and then there are the long tail where you just want to pay a new supplier.” 

Axiom has built a single ecosystem that handles everything from catalog purchasing to one-off supplier payments encompassing an ‘all in one place’ approach to make life easier for procurement teams. “The problem is that there are different component elements of tail spend,” explains Alistair Cane, Co-Founder and Director at Axiom. “There are elements that require sourcing, a robust catalog strategy and then there are the long tail where you just want to pay a new supplier.” 

“The challenge you have with other systems is that they are just handling a part of that. But when you think about it from the buyer’s journey, they want a complete end-to-end capability and not to have to jump from one platform to another. Within our tail spend solution, we have all of those modules beautifully integrated – a single place for the organisation to manage tail spend.”

Axiom is firmly customer-focused. It initially built the platform alongside Roche in 2021 which represented a significant step in catalog procurement. The platform not only fully digitises Roche’s global catalog content, but also revolutionises the buying process and experience, delivering real-time savings. Developed through deep collaboration over two years, this new procurement technology solution empowers organisations to establish, and seamlessly manage, their MarketPlace for any catalog content and related tail spend. Over time, Axiom’s product has evolved directly from the problems that clients raised.

“We worked with Roche closely to build the platform in the beginning,” explains Aaron Ueckermann, fellow Co-Founder and Director at Axiom. “That was our core MarketPlace and catalog management solution and that went live and we made continuous iterations on it. What we discovered is sometimes there’s things that just aren’t available in the catalog and users need to do one-off sourcing events to find new things or large quantities of existing things. So we launched MarketSource and that solved that gap.

“But we were then approached by another customer who said they wanted users to get things from inventory rather than buying something new. That caused us to develop our MasterMaterials and inventory management solution. We’ve continued on this theme across various modules and what we’ve found is we never really build anything just for the sake of a shiny feature for a demo. We always build off the back of customer problems so that we deliver something that they don’t currently have in their existing solutions.

“This has worked very well for us, but our next challenge is about ensuring we also place some big bets such as new product lines which customers aren’t immediately asking for where we’re looking ahead for new problems to solve. Keeping that balance is fun.”

Traditionally, purchasing cards (P-Cards) were used to streamline low-value transactions and help enable faster purchasing. These allowed for controlled spending, more efficiency and improved supplier relationships. However, one of the major drawbacks of P-Cards from a compliance perspective is that authorisation only happens retrospectively after the event has taken place.

This is where Axiom’s MarketPay solution comes in. MarketPay is built to handle the full tail spend flow instead of fragments of it. It operates on a disclosed procurement agency model – which means Axiom acts on the customer’s behalf instead of as an intermediary. Powered by its procurement agent, it automates the steps that typically slows everything down from validating request data to enforcing thresholds and compliance. Every payment also runs through a global payment account held in the customer’s name which allows the team to stay in control at all approval stages.

“MarketPay is essentially a disclosed procurement agent model,” explains Ueckermann.“What this means is that if you want to buy something from a supplier who you don’t currently work with or have contracted, you just upload a quote, specify an email address for who needs to receive the purchase order and then you send that off to the supplier. They onboard into Axiom and we take care of all the compliance requirements specific to the business before the supplier invoices through Axiom. But the key part is the legal relationship is still between the customer and the supplier. We see that the weakness of standard merchant of record models is that you end up buying through a third party, which can lead to compliance and tax issues, especially when it’s cross border. Now MarketPay resolves these, but it also keeps the simplicity and benefits of uploading a quote, specifying email address and getting paid.”

Modern procurement is changing quickly as new technologies come to the fore. Over the past 12 to 18 months, the biggest craze has been the rise of agentic AI and how agents are freeing procurement teams up to take on more strategic work. Cane believes the advancement of tools like agentic AI is redefining how the function lives and breathes. “One of the beauties of the Axiom platform is that there are multiple sides to the platform. There’s the buyer side, procurement portal and the supplier side, and the agent works across all of those. From a buyer perspective, ultimately what they want to do is source an item or a service as quickly as possible. If the agent can do a lot of the stuff that they would have had to do manually then that’s great for user experience. From the procurement side, it’s just removing and automating many of the tasks that procurement have often struggled with. One of the things with tail spend is the dynamic of you having 20% of your spend sitting with 80% of your transactions and often 80% of your resource. A lot of that resource in enterprise organisations are involved in administrative tasks managing this landscape. If the agent can do some of that heavy lifting from a mechanical perspective then that can only be a good thing.”

Axiom is passionate about AI. The organisation has even coined the term ‘autonomic sourcing’ to describe something beyond simply AI-assisted buying. “Our belief is that every single interaction and transaction in itself contains data and insight and we want to capture, understand and use this intelligence that sits around all the activities,” explains Cane. “The term autonomic sourcing is because it’s reflective of something that we all experience every day. The autonomic nervous system in the body controls the rate we breathe, our heart rate and the amount we blink. We do it without thinking and that’s exactly where we want to take the platform  so procurement doesn’t have to worry about the data that’s coming out of the platform. Let the platform analyse, understand as well as create insight, actions and outcomes. That, we believe, is the next generation of procuretech solutions.”

Find out more about Axiom here.

At DPW New York City, the theme of ‘Recode’ challenged procurement leaders to rethink how their functions operate in an increasingly digital and AI-driven world. For Vel Dhinagaravel, Founder & CEO of Beroe, and Suketu Gandhi, Global Chair at Kearney, the conversation is about far more than process improvement. It is about fundamentally redefining procurement’s purpose and positioning it as a driver of competitive advantage.

The pair believe procurement stands at a pivotal moment in its evolution. While the profession has spent decades focused on savings and cost control, today’s technologies and market dynamics are creating an opportunity to transform procurement into a strategic function that directly influences growth, speed, resilience and profitability.

Moving Beyond the Savings Mindset

For many procurement organisations, success has traditionally been measured through annual savings targets. According to Dhinagaravel, that approach no longer reflects the realities of modern business.

He argues that savings were an appropriate metric when procurement’s primary role was consolidating spend and leveraging economies of scale across regions and global operations. However, the challenges facing organisations today require a different perspective.

“I think that’s outdated,” he adds. “We’ve got to move to a mode of measuring and striving for continuous competitiveness.”

This shift represents more than a change in reporting metrics. It requires procurement teams to rethink how they define value and how they contribute to broader business outcomes. Rather than focusing solely on reducing costs, procurement must increasingly consider how it can influence margins, accelerate growth and strengthen competitive positioning.

Reimagining Procurement’s Purpose

The rise of artificial intelligence has prompted many organisations to explore automation opportunities, but these leaders both caution against simply applying AI to existing operating models.

According to Gandhi, the real opportunity is not to optimise current processes but to reconsider why procurement exists in the first place.

“One of the most powerful things about the technology available right now is you don’t need to rethink how the processes run,” he says. “You have to completely reimagine why you exist.”

This distinction is critical. Many organisations are implementing AI solutions to accelerate existing workflows, but Dhinagaravel warns that doing so may create only the appearance of transformation.

“A lot of companies are putting an AI veneer on top of existing processes,” he says. “It’s giving them an illusion of speed because everything is running faster, but the outcomes aren’t going to be any different because the ambitions and the operating model haven’t been made more ambitious.”

Instead, both Dhinagaravel and Gandhi advocate for a more fundamental redesign of procurement’s role, one that leverages technology to deliver entirely new levels of strategic impact.

Procurement’s Growing Strategic Influence

Perhaps the clearest sign of procurement’s evolution is the changing nature of conversations taking place in the boardroom.

Gandhi notes that procurement leaders are increasingly being invited into discussions that extend far beyond sourcing and cost management. Rather than being tasked with delivering incremental savings improvements, they are being asked to solve strategic business challenges.

“Now they’re called in to say, ‘We are launching a new product. How can we fundamentally increase the speed at which the product will launch?'”

Similarly, procurement is becoming a critical partner in addressing supplier disruptions, supply chain challenges and broader operational issues that affect business performance.

This represents what Gandhi describes as a “night and day shift” from the traditional procurement model. The function is increasingly viewed as an enabler of growth, innovation and resilience rather than simply a transactional support service.

The Human Role in an Automated Future

As AI adoption accelerates, questions inevitably arise about the balance between automation and human decision-making.

Dhinagaravel believes technology should enhance procurement professionals rather than replace them. He describes AI-enabled tools, like Beroe MAX, as an “exoskeleton” that expands the capabilities of category managers, allowing them to make better decisions and manage a broader scope of responsibility.

“They should still be absolutely at the heart of the decision making,” he adds.

The opportunity lies in increasing both capacity and capability. By automating routine activities and delegating lower-value tasks, procurement professionals can spend more time focusing on strategic decisions and value creation.

Over time, this could enable procurement leaders to oversee significantly larger areas of spend and influence while maintaining high-quality decision-making.

Changing the Narrative for Procurement

While procurement’s influence has grown substantially, both Dhinagaravel and Gandhi believe perception remains one of the profession’s greatest challenges.

Dhinagaravel points to the COVID-19 pandemic as a turning point. During the crisis, procurement teams played a critical role in maintaining operations, securing essential supplies and helping organisations navigate unprecedented disruptions.

As a result, the outdated perception that procurement simply “buys things” has largely diminished.

However, he argues that procurement now faces a different challenge: effectively communicating its value.

Gandhi agrees and believes the way procurement talks about itself must change. “When you have a seat at the table, you are a leader in the company,” he says.

Instead of leading with savings figures, procurement leaders should focus on explaining how their work impacts customers, supports business strategy and contributes to financial performance. The operational details of procurement should come later in the conversation.

For Dhinagaravel, storytelling is becoming a core leadership capability. “Procurement is doing the work. They’re getting the results, but they’re not articulating it in a zoomed out fashion, which will appeal to the C-suite and to the board.”

Talent and Leadership for a New Era

Alongside perception, talent remains one of the biggest challenges facing procurement leaders.

According to Gandhi, organisations need professionals who can operate at the pace demanded by both technology and market competition.

As digital tools become more powerful, procurement teams must develop new capabilities that combine commercial acumen, strategic thinking and technological fluency.

Communication skills are becoming equally important. Leaders must be able to translate procurement activities into a language that resonates with senior executives and demonstrates business impact.

The ability to influence, communicate and connect procurement outcomes to enterprise objectives will be increasingly valuable as the profession continues to evolve.

The Shift from Cost Reduction to Margin Maximisation

Looking ahead, both leaders see enormous opportunity for procurement to expand its contribution to business performance.

Dhinagaravel believes the next stage of procurement’s evolution lies in shifting from a cost-reduction mindset to one focused on margin maximisation. Something that inspired the genesis of Beroe MAX.

Rather than concentrating solely on purchase prices, procurement can influence a wide range of factors that affect profitability – including revenue growth, pricing realisation, waste reduction and specification improvements. “All of these levers are available to procurement,” he says.

This broader perspective positions procurement as a central contributor to business success rather than a function focused exclusively on expenditure control.

A Defining Moment for Procurement

For Gandhi, the profession has never been better positioned to create value. “This is the most exciting time since the profession of procurement was invented,” he says.

With procurement increasingly involved in commercial, financial and product-related decisions, the function is becoming central to how organisations compete and grow. Combined with rapidly advancing technology and a growing appetite for transformation among procurement leaders, the conditions are in place for a fundamental reinvention of the profession.

The challenge now is not simply adopting new tools. It is embracing a new vision for what procurement can become.

“The next evolution of procurement is going to revolve around user experience and simplicity.”

Jeremy Lappin, CEO and Co-Founder of Candex, has no doubt about what lies ahead for procurement. 

Given the plethora of tools procurement has at its disposal- in the form of new and old tools, different processes and new third-party risk management requirements- the business world can seem fragmented. 

Fortunately, Candex exists and has a mission to simplify the user experience.

Candex is a tail spend management leader that focuses on making payments to small or one-time vendors easier. Founded in 2011 by Lappin and Shani Vaza, the company’s technology removes thousands of vendor records from financial systems all while delivering control, speed, tax compliance and visibility.

“We take the noise out of managing and paying small, one-time vendors,” Lappin explains. “Companies are constantly finding new vendors that are more cost-effective, innovative, or better suited to specific needs. We make it easy to engage those suppliers quickly, without getting bogged down in lengthy onboarding processes. This enables procurement teams to focus their time and resources on high-value, strategic purchases that warrant deeper due diligence, while we simplify the management of smaller, one-off purchases. The result is less operational overhead, faster vendor engagement, and a more efficient procurement process overall.”

In 2026, companies operate within a complex environment with a significant number of suppliers needing to be onboarded. This is often cumbersome and time-consuming when it comes to smaller spend, which is where Candex comes into the equation. Candex sits at the intersection of buyers, suppliers, and payments, and eliminates the hassle involved – allowing procurement teams to focus on completing the necessary compliance on the higher-value, more strategic spend, and removing bottlenecks.

And over the past few years, Candex’s solution has grown considerably in popularity within the space – quickly becoming a go-to tool that eradicates the unnecessary noise from operations. Lappin explains that once that happens, procurement teams are left with a clear path to excel. “The more people experience Candex, the more they make it a part of their core strategy.”

Today, much of procurement’s operating model is antiquated and still relies on processes that were designed decades ago. With this in mind, Candex is heavily aligned with this year’s DPW theme of Recode – which is the idea of rethinking how procurement and supply chains operate – and Lappin believes there is an increasing amount of pressure on introducing AI correctly into their organisations. “That’s likely to be a major theme over the next few years,” says Lappin. “The challenge won’t be adopting AI, it will be using it in ways that create meaningful value while balancing other business priorities. While building internally has become faster and more accessible, speed alone isn’t an advantage. Organizations must be deliberate about what they build and how they implement it. Otherwise, the rapid pace of development can easily lead to inefficiencies and unintended consequences.”

Janina Insua, Global Head of Sales at Candex, adds that Candex Programs – another product that Candex offers- feels closely aligned with Recode. Candex is redesigning, or rewriting processes, to make it easier to manage transactions that sit outside traditional procurement while still maintaining control and compliance. “These transactions that sit outside of traditional procurement are exactly what we built Candex Programs for,” adds Insua. “There are a subset of transactions where one-to-one POs are often not practical because requiring unique requisitions for every transaction of that payment type would be a huge friction point for the business”.

“Some examples that come to mind are HCPs, candidate reimbursements, legal settlements, charity donations; these often require shorter payment days, a different set of contracts, unique reporting, and they often benefit from a different onboarding experience than the one we know of for traditional vendors”.

“Candex Programs supports all of the above while still allowing our customers to preserve their pre-existing workflows and approvals, and they get full visibility into what’s being purchased, by who, from who, from where, how many times, and for how much.”

Candex Programs is similar to re-coding how legacy procurement systems were designed, reimagining them for the types of transactions that never really fit in the first place, but now with the right checkpoints and guardrails to stay compliant. According to Insua, one of the main reasons that non-traditional spend has historically never fit into the legacy procurement process is a poor user experience. “Procurement technologies and processes fail to deliver value when they’re too complex and misaligned with how people actually work,” Insua tells us. “Trying to force rigid and unintuitive processes leads to disengagement, slower cycle times, and high third-party risk, which no one wants. If companies want to bring structure and control to this space, they have to do it with the user experience in mind. Procurement teams need to be intentional with removing unnecessary obstacles or friction points from the correct behaviours and processes that they’re wanting to breed across the business.”

If a customer does not use purchase orders as part of their procurement workflow, with Candex Programs they still have the ability to use their pre-existing workflows to manage vendors and spend. Candex only engages a payee once a payment request has been approved by the buyer. Candex then collects all required details, runs compliance scans on the payee, and issues payment to the payee once the payment is made by the customer to Candex. And there are several ways that Candex brings control and governance to these payments, Insua reveals. “First, vendors are required to accept contractual terms and conditions upon acceptance of any order through the Candex platform,” she says. “This means there’s an engagement contract every time, which protects Candex and our customers. In addition, customers can set payment limits as to how much can be paid to any vendor through Candex, meaning the customer is setting the threshold for spend materiality.  Not only that, the customer decides what commodities are purchased through Candex as well. This means that if they want to exclude any high risk commodities, then they can. A number of customers also apply checkbox verifications or custom questionnaires for additional compliance.”

This list goes on. “Candex gives clients the flexibility to introduce additional approval layers when needed, helping them maintain the right level of governance over their spend. We also conduct daily vendor screening against leading sanctions and anti-money laundering (AML) databases, including watchlists and other regulatory sources. If a vendor is flagged, clients are notified immediately. To support transparency and compliance, Candex provides real-time visibility into every order, payment, and payee. Clients have access to comprehensive, audit-ready reporting and can manage reporting access by country, category, and other dimensions relevant to their business. Why is this important? Because in this world of AI evolution, having quality data will be key to successfully ‘recoding’, orchestrating, and automating existing processes to better cater to the user experience.” 

Find out more about Candex here.