Episode 2 - AI in Procurement: What’s Real, What’s Next, and What Actually Works in 2026

AI is everywhere in procurement—but what’s actually real, what’s hype, and what’s delivering measurable value today?

In this episode of The Procurement Brief, we cut through the noise and take a clear, management-level look at how AI is reshaping procurement operating models in 2026. This isn’t theory. This is what’s working inside real organizations right now.

We explore the shift from traditional digitization—ERP, P2P, sourcing, and CLM platforms—into true decision intelligence. Procurement is no longer just automating workflows; it’s becoming a strategic function powered by data, predictive insights, and AI-driven execution.

You’ll learn:
 • Where AI is already delivering impact across sourcing, contract management, supplier risk, and spend analytics
 • How leading organizations are reducing manual workload while increasing savings capture
 • The difference between “AI features” and true AI-enabled operating models
 • What procurement leaders must do now to stay relevant over the next 3–5 years
 • Practical, no-nonsense guidance on where to start—and where not to waste time

We also break down the real-world constraints: data quality, adoption challenges, integration complexity, and why most organizations struggle to move beyond pilots.

This episode is designed for procurement and finance leaders who want to move beyond buzzwords and understand how AI can drive measurable enterprise value—today, not someday.

If you’re thinking about the future of procurement, this is the conversation you need to hear.

Transcript

SPEAKER_00 0:00

Hello and welcome back to the procurement brief. I'm Patrick Bliss. Today we're going deeper than the typical AI conversation. Because right now, every procurement technology provider is talking about AI. Every conference is filled with gen AI demos. Every platform claims to be intelligent. But here's the real question. What is actually working inside procurement operations today? And what will meaningfully change by 2027 and beyond? This episode is built from the lens of real implementations, system integrator experience, and what leading organizations are deploying right now. We're going to walk through AI across the full procurement lifecycle, the platforms actually delivering results, where ROI is real and where it's still early, and how procurement leaders should be thinking about adoption over the next 24 months and beyond. Let's get into it. A journey of digitization. And that journey has been important because before digitization, procurement was largely manual, email and spreadsheet driven, relationship driven, but not always data driven. So organizations invested heavily in technology to bring in structure, control, and visibility to the process. Let's break down what that actually looked like. First, ERP systems, platforms like SAP, Oracle, and Microsoft Dynamics. They became the financial backbone of the enterprise. They brought standardized purchasing processes, general ledger integration, basic supplier records, purchase order control, and financial compliance and auditability. For the first time, companies could answer questions like: how much did we spend? Who did we pay? Was it approved? But ERPs have limitations. They were designed for record keeping, not intelligence. They told you what happened, but not what to do next. Then came procure-to-pay or P2P platforms. Solutions like Coupa, Ariba, Oracle Procurement Cloud, and many, many others. These improved the user experience and expanded control. They introduced guided buying, catalog-based purchasing, automated approval workflows, invoice matching and processing, and supplier portals. This was a major step forward. It reduced Maverick spend, it improved compliance, it increased visibility into transactions. But again, these systems were built and are built around workflow automation. They make the process smoother, but they don't make decisions. They still relied on humans to interpret data, to intervene, and to manage exceptions. Next came contract lifecycle management or CLM platforms. Tools like ISERTIS, Agiloft, Ironclad, Conga, and others. These digitized contracts, which had historically been one of the most fragmented areas in procurement. CLM introduced centralized contract repositories, version control, clause libraries, approval workflows, and obligation tracking. This was critical, especially in regulated industries. For the first time, organizations could find contracts quickly, standardize legal language, track renewals and obligations. But again, the limitation, CLM systems stored and organized contracts. They didn't understand them. You still needed a person to read the contract, interpret risk, identify deviations, and drive negotiation strategy. Then came e-sourcing platforms like Ariba Sourcing, Jagger, GEP, IVA, Workday Scout RFP, and others. These tools digitize competitive bidding. They enabled RFP creation and distribution, supplier response collection, bid comparison, auction capabilities, and basic scoring models. This improved transparency, it improved competition and documentation. But sourcing still required heavy manual effort. Reading proposals, summarizing responses, comparing suppliers, building evaluation decks. Again, structured process, but not intelligent insight. Now don't get me wrong, I use these tools every day and I think they're great. But if you step back, there's a clear pattern. Digitization gave us structure, control, visibility, and standardization. But it did not give us intelligence, prediction, recommendation, and decision support. And that's the inflection point we're in right now. Because what organizations are realizing is this. And this is where AI comes in. According to Gartner, by 2027, more than 50% of procurement organizations will be using AI-enabled decision intelligence platforms. The Hackett group adds to this by stating world-class procurement teams are already using AI to reduce operational workload by up to 45% while simultaneously increasing savings capture. This is not incremental improvement. This is not just better tools. This is a fundamental shift in how procurement operates. Now what's actually changing? Well, we're moving from process automation to decision automation. Not just executing steps faster, but enabling systems to recommend and in some cases execute decisions. We move from data visibility to decision intelligence. It's no longer enough to see the data. Now the expectation is: what does the data mean? What actions should we take? What risk exists? And what opportunity or opportunities are we missing? System workflows turn into adaptive learning systems. Traditional systems follow rules. AI systems learn patterns. They improve over time. They adapt based on behavior, outcomes, and new data. So digitization was the foundation. It got procurement organized, visible, controlled. But intelligence, that's that's what really turns procurement into a true strategic function. And that's exactly where we're going next. Now that we've established the shift from digitization to intelligence, let's bring this to life across the procurement lifecycle. Because AI is not one capability, it's being embedded piece by piece into every stage of procurement, from intake to sourcing to contracting to negotiation and to payment. And each one is evolving at a different pace. Let's start at the front door of procurement, intake. Historically, intake has been one of the biggest friction points. A business stakeholder needs something, but they don't know where to go. So what happens? They email someone, they bypass procurement, they engage a supplier directly. And procurement, well, they find out later. Digitization tried to fix this with guided buying. P2P platforms introduced catalogs, preferred suppliers, approval workflows, and intake forms. And it helped to a point. But the reality is users still had to understand the system. They had to know which category to select, which form to use, and what process to follow. Now, enter AI. Platforms like Zip, Coupa, SAP, Ariba, and others are fundamentally changing intake. Instead of navigating a system, the user simply states intent. I need a consulting firm for a six-week project. And AI translates that into category classification, risk tier, required approvals, suggested suppliers, budget alignment, and workflow routing. This is what's called intake orchestration, and it's powerful because it removes friction at the very beginning. According to Spend Matters, AI-driven intake can improve routing efficiency by 30 to 50% and significantly increase procurement channel compliance. So instead of procurement chasing demand, demand flows correctly from the start. That's a fundamental shift. Now let's move into sourcing. Traditionally, sourcing has been one of the most labor-intensive activities in procurement. Creating the RFP, building requirements, collecting supplier responses, reading proposals, summarizing differences, aligning stakeholders. It is high-value work, but it is extremely time consuming. Digitization gave us e-sourcing tools. They structured the process, which was important, but the work itself remained manual. Now AI is changing that. Platforms like Evalua, GEP, Jagger, and others. They're embedding AI directly into sourcing workflows. Here's what that looks like in practice. Instead of building an RFP from scratch, AI can generate a first draft using historical sourcing events, suggest requirements based on category best practices, preload evaluation criteria, and recommend suppliers based on past performance. When responses come back, AI can summarize each proposal, highlight key differences, flag risks or missing responses, and score suppliers against weighted criteria. So instead of spending days reading and comparing proposals, the sourcing manager starts with a structured, summarized, prioritized view. According to Deloitte's Global CPO survey, AI-enabled sourcing can reduce event cycle time by up to 60%. And here's the real impact. It doesn't replace sourcing professionals. It elevates them. They spend less time reading and more time thinking, less time compiling, and more time negotiating. Next, let's talk about contracts. Historically, contracts have been one of the biggest blind spots in procurement, even after digitization. Yes, CLM systems centralized contracts, but they didn't make them intelligent. So what did that mean? Well, contracts were stored but not easily understood. You still needed someone to read the document, interpret the clauses, identify risk, and compare deviations, tracking obligations manually. Now AI is transforming CLM. Platforms like iSertis, DocuSign, and Sirion are embedding Gen AI into contract workflows. Now AI can extract key clauses automatically, identify non-standard language, compare contracts against playbooks, suggest fallback language, and flag risk exposure in real time. Even more importantly, AI can track obligations after signature, things like renewal dates, SLAs, pricing escalators, compliance requirements. According to McKinsey, AI-enabled contract analysis can reduce review effort by 50 to 80% while improving risk detection. This is one of the highest ROI areas in procurement today because contracts are everywhere and they're historically underutilized. Now let's move into negotiation because this is where procurement becomes an art. And historically it's been driven by experience, instinct, preparation, and sometimes incomplete data. You walk into a negotiation knowing what you paid last time, what the supplier proposes, maybe some benchmark data, but rarely the full picture. But AI is beginning to change that, and embedded within platforms like Coupa and GEP, AI can provide market pricing benchmarks, historical negotiation data, supplier performance insights, total cost modeling, and scenario analysis in real time. Imagine being in a negotiation and having real-time intelligence that tells you you're 12% above market. The supplier has conceded on this clause before. There are viable alternatives with lower risk. According to Gartner, AI augmented negotiation can improve realized savings by 3-5 percentage points. Yeah, I know that sounds moderate, but at an enterprise scale, that could be massive. This doesn't replace negotiation skill though. It enhances it. It gives procurement leaders better information in the moments that matter most. Let's move on to discuss supplier management. Traditionally, supplier risk has been reactive. A disruption happens, and then we respond. Digitization improved visibility but relied heavily on periodic reviews, manual scorecards, and lagging indicators. Now AI is flipping that model. Platforms like ECOVADIS, Pre-Wave, and RISC methods are using AI to monitor suppliers continuously. They analyze financial signals, ESG compliance, news and media, regulatory violations, cybersecurity posture, and geopolitical exposure. So instead of quarterly reviews, you get real-time alerts. IDC reports that AI-driven supplier monitoring improves disruption detection by up to 35% and significantly reduces response time. Procurement becomes proactive, not reactive. Finally, let's talk about accounts payable because this is one of the most mature AI use cases today. Historically, invoice processing has been manual, error prone, labor-intensive, it matches invoices to POs, it resolves discrepancies, and it chases approvals. Digitization helped, but it still requires heavy human intervention. Now AI is automating this end-to-end. Platforms like AppZen and VIC.ai are enabling intelligent invoice capture, automatic GL coding, duplicate detection, fraud identification, and autonomous approval of low-risk invoices. According to Deloitte, AI-driven AP automation can reduce invoice processing costs by 60 to 80%. This is where we're already seeing near autonomous operations. And it's a glimpse of what's coming across procurement more broadly. So when you look across the life cycle, AI is not one transformation, it's many, happening at once. Some areas, like AP, are already mature. Others, like sourcing and negotiation, are rapidly evolving. And some, like autonomous procurement, well, they're just beginning. But the direction is clear. Procurement is moving from manual execution to intelligent orchestration. And then in the next section, we're going to separate what truly works today and what's still hype. Now let's step back for a moment, because if you've been anywhere near procurement technology conversations lately, you've probably heard a lot about AI. Every platform is AI powered, every demo is transformational, and every roadmap promises automation. But here's the reality: not all AI in procurement is equal. Some capabilities are delivering real value today, some are emerging quickly, and some there's still more vision than reality. So let's break this down clearly. What's actually working right now? There are several areas where AI is not only real, it's already delivering measurable results. Spend classification and analytics is one of the most proven use cases. AI can now classify spend with over 90 to 95% accuracy across millions of transactions. It identifies duplicate suppliers, pricing inconsistencies, tail spend opportunities, and contract leakage. And it does it continuously, not just quarterly. This is foundational because if you don't understand your spend, you can't optimize it. Contract intelligence is arguably one of the highest ROI use cases today. AI can extract clauses, compare contracts against playbooks, identify risk exposure, and suggest fallback language. And most importantly, it dramatically reduces review time. Imagine if your attorneys and the investment costs they bring could focus on actual solutions instead of finding issues. We're seeing organizations move from weeks of contract review to hours. In regulated environments, this is a game changer because it improves both speed and compliance. Moving on to invoice automation and AP intelligence, as mentioned, this is one of the most mature areas today. AI is already coding invoices automatically, detecting duplicates, flagging anomalies, and identifying fraud patterns. In many organizations, low-risk invoices are now processed fully autonomously. This isn't future state, this is happening today. What about intake and workflow orchestration? AI-driven intake is gaining real traction. It simplifies how demand enters procurement. Instead of users navigating systems, AI interprets intent and routes requests correctly. This improves user adoption, compliance, speed, and most importantly, it ensures procurement is involved early. How about supplier risk monitoring? AI-driven supplier intelligence is also very real. Continuous monitoring of financial health, ESG performance, cybersecurity posture, and regulatory exposure. This moves procurement from reactive risk management to predictive risk management. So these areas, they're not experimental, they are operational. So what's emerging but not fully mature yet? Let's talk about the next layer. These capabilities are real, but still evolving. AI-driven sourcing execution can assist with RFP creation, proposal summarization, and supplier scoring. And it's delivering real-time savings. But is it fully autonomous sourcing? Not quite yet. Because sourcing still requires stakeholder alignment, business context, and strategic decision making. AI helps, but humans still lead. Negotiation intelligence is one of the most exciting areas. AI can provide market benchmarks, supplier history, pricing models, and scenario analysis. But here's the reality. Negotiation is still deeply human. It involves timing, tone, relationships, and strategy. AI enhances negotiation, but it doesn't replace it. Autonomous procurement for low-risk spend is coming and it's coming fast. We're starting to see auto-approved purchases, system-generated POs, and supplier recommendations. But fully hands-off procurement across categories? Yeah, still emerging. Because organizations are still building trust, governance, and risk frameworks. This is going to scale, but it's going to take two to three years. Now, let's be honest, what's overhyped? Here are areas where the market is getting ahead of reality. Fully autonomous procurement. You hear messaging like procurement runs itself, AI replaces buyers, and end to end automation. Come on. That's not where we are. And in most organizations, that's not where we should be. Procurement involves risk, judgment, relationships, and regulatory oversight and interpretation. You don't automate that completely, you augment it. Another interesting factor is AI replacing procurement professionals, and this comes up a lot, and it's simply not accurate. AI is not replacing procurement individuals. It's replacing manual work, repetitive tasks, and data gathering. In fact, as AI adoption increases, the demand for skilled procurement professionals will actually increase because the role now shifts to strategy, influence, decision making, and value creation. Another misconception is perfect AI decision making, that AI always gets it right. Well, it doesn't. AI can misinterpret data, it can hallucinate outputs, it can miss context entirely and over-generalize patterns. That's why governance matters. It's why human oversight still matters. This may be the single most important thing I can say in this entire podcast. AI is a tool. It is not an authority. So if we step back, here's the truth. AI and procurement is very real in structured, data-heavy processes. It's rapidly evolving in strategic areas, but it's also overhyped when it comes to full autonomy. And the organizations that are winning right now, they're not trying to automate everything. They're focusing on high volume, high friction, data heavy processes. And they're using AI to reduce effort, increase speed, and improve decision quality. So the takeaway is simple. Don't chase the hype. Focus on what works. Build confidence in the areas delivering value today and expand from there. Because AI is not a single transformation, it's a progression. And the organizations that approach this pragmatically are the ones that will scale it successfully. So let's shift the conversation. Because up to this point we've been talking about technology, but the real impact of AI and procurement is not the tools, it's the operating model. Now, yes, I know I just said AI is a tool. But that's the technology side of it. The company's operating model is super important. Because AI doesn't just make procurement faster, it can change how work gets done, who does the work, where decisions happen, and what skills actually matter. And this is where many organizations are getting it wrong today. They implement AI, but they keep the same structure, the same roles, and the same processes. And when that happens, they get incremental improvement, not transformation. At its core, procurement is shifting from executing transactions to orchestrating outcomes. Historically, procurement teams were built around processing requisitions, running sourcing events, managing contracts, handling supplier issues. A lot of doing. But AI is removing much of that execution burden, which means the role of procurement is evolving into interpreting insights, guiding decisions, managing risk, influencing stakeholders, and driving enterprise value. This is a fundamental shift. And it changes the roles. So let's look at this from the role level. Traditionally, category managers spent a significant amount of time on building RFPs, collecting bids, analyzing responses, and managing supplier negotiations manually. Now AI is handling much of the data-heavy work, which means the category manager's role has shifted to strategy development, stakeholder alignment, supplier relationship management, value engineering, and risk evaluation. They've become less of a sourcing executor and more of a business partner and advisor. Procurement operations teams have historically focused on requisition processing, PO creation, invoice resolution, and workflow management. With AI-enabled P2P and AP, much of this becomes automated. So what happens? Well, the operations teams evolve into exception management, process optimization, data quality ownership, and system governance. Instead of processing everything, they manage what needs attention. And how about supplier management teams? Well, this has traditionally been periodic: quarterly reviews, manual scorecards, and lagging indicators. AI can change that completely. Supplier data can become continuous, real-time, and predictive. So supplier managers shift from reviewing performance to actively managing risk, opportunity, and innovation. They become risk advisors, performance strategists, and relationship leaders. How about procurement leadership? Because this is where the biggest shift has happened. Historically, procurement leaders were measured on cost savings, spend under management, and compliance. Now, these still matter and I'm still measured on them. But AI enables something bigger. Procurement leaders now have the opportunity to become enterprise value architects. They'll drive cost optimization. Personally, I call it spend optimization, risk mitigation, supplier innovation, ESG alignment, and speed to market. According to the Hackett Group, organizations adopting AI and procurement can increase strategic capacity by 20 to 30 percent without increasing headcount. That's capacity and bandwidth. That is throughput. It's not efficiency, it's transformation. Now here's something we're seeing more and more. Leading organizations are creating procurement AI centers of excellence. These teams are responsible for AI governance, use case prioritization, model oversight, data quality standards, and cross-functional alignment. This is a capability that needs ownership. Without that ownership, you get fragmentation, you get shadow AI usage, inconsistent outputs and risk exposure. The COE needs to become the control tower. Let's talk about talent. Because AI doesn't reduce the need for people, it changes the skills required. The future procurement professional needs to be data literate, comfortable interpreting AI outputs, skilled in prompting and interacting with AI systems, strong in business acumen, and strong in stakeholder influence. In many ways, and this is the focus of my podcast, procurement becomes more strategic, more analytical, and more consultative, and less transactional. So what happens if you don't evolve the operating model? Well, let's be direct. If an organization implements AI and doesn't evolve its operating model, here's what happens: tools are underutilized. Teams resist change. Value will not be realized, and AI becomes just another system. We've seen this before with ERP, P2P, and CLM and some of their limitations. Technology alone does not transform procurement though. Operating model change really drives it. The winning model going forward and the organizations who are getting this right are doing a few key things. They're automating high-volume, low-value work. They're redeploying talent to strategic activities. They're building centralized governance around AI. They're aligning procurement closely with finance, legal, and IT. And they're measuring outcomes, not just activity. They're designing procurement to operate lean, intelligent, and highly strategic. So the real takeaway here is this: AI is not just improving procurement, it's redefining it. And organizations that recognize that and evolve their operating model accordingly are the ones that will lead. Because the future of procurement is not about doing more work, it's about delivering more value with smarter systems and more strategic people. Now let's talk about something that often gets overlooked in AI conversations, and that's risk. Because while AI creates enormous opportunity, it also introduces new categories of risk that procurement leaders need to understand and manage proactively. And this is especially important in industries like financial services, healthcare, and insurance, and any organization operating under regulatory oversight. Because in these environments it's not enough for AI to be powerful. It has to be explainable, auditable, controlled, and compliant. The reality is AI is not risk-free. There's a perception in the market that AI is inherently smart and therefore reliable. But that's not entirely true. AI systems can misinterpret data, generate incorrect outputs. We talked about hallucinating responses, missing context, and reinforcing bias in underlying data. And when you apply that to procurement, the risks become very real. Let's look at the major categories of risk. First, let's start with contract risk. AI can generate or suggest contract language, but if not governed properly, it can insert non-compliant clauses. It can deviate from approved legal standards, and it can misinterpret fallback or replacement language. In regulated environments, this is critical. A single clause deviation can create financial exposure, regulatory risk, or legal disputes. So now let's talk supplier risk. AI models rely on data. If that data is incomplete, outdated, or biased, the output will be as well. Remember the old adage, garbage in, garbage out? Well, it applies here. This can lead to poor supplier selection, unintentional bias in evaluation, and overlooking critical risk signals, which directly impacts resilience, compliance, and reputation. How about financial risk? In procure to pay an AP, AI can misclassify spend, approve incorrect invoices, miss duplicate payments, and misinterpret pricing terms. Without proper controls, this can translate directly into financial loss. Some new key elements that are super important include data privacy and security risk. AI systems often can process supplier data, contract terms, pricing structures, and personally identifiable information or PII. If not properly governed, this creates exposure to data breaches, regulatory violations, and confidentiality risks. What about lack of auditability and explainability? Well, this is one of the biggest concerns in regulated industries. If an AI system has to make a recommendation, you need to be able to answer why did it make that decision? What data was used, and what rules were applied? If you can't answer those questions, you don't have control. For regulatory alignment, frameworks like the OCC Third Party Risk Management Guidelines and the Interagency Third Party Risk Guidance, they make one thing very clear. Organizations are responsible for the outcomes of their third-party relationships, including technology providers and AI systems. That means if an AI system makes a decision, your organization owns that decision. There is no the system did it defense. So how do leading organizations manage this? Well, they build governance into the foundation. Keeping humans in the loop and the controls is super important. AI should augment decisions but not replace accountability. For critical activities like contract approvals, supplier selection, high-value sourcing awards, there must always be human review and approval. How about policy and playbook integration? AI should not operate in isolation. In fact, it never should. It should be grounded in approved contract clauses, risk frameworks, category strategies, and approval thresholds. In other words, AI should reflect your procurement policy, not invent its own. We've mentioned auditability and traceability. Every AI-driven decision should be logged, traceable, and explainable. You should be able to reconstruct what happened, why it happened, and who approved it. And this is essential for audit, compliance, and internal trust. What about data governance and quality? Well, it's been said before, AI is only as good as the data behind it. Leading organizations invest heavily in clean supplier master data, in structured contract metadata, in accurate spend classification, and controlled data access. Because poor data leads to poor decisions and even faster with AI. What about vendor risk management? Well, this is critical and often overlooked. AI vendors are third parties. Don't forget that. Which means they too must be evaluated under third-party risk frameworks, security and privacy standards, model transparency expectations, and service level agreements. You need to understand how their models work, how data is used, where data is stored, and what controls are in place. So what's the role of a procurement AI governance model? Well, this is where leading organizations are investing heavily. They're establishing formal AI governance structures within procurement, often through a center of excellence or a cross-functional governance body, with representation from procurement, legal, technology, risk, and compliance. This group defines the acceptable use of AI, the risk thresholds, the approval frameworks, and the implementation standards. Because AI is an enterprise capability. So here's the key message. AI in procurement is incredibly powerful, but it is not plug and play. It requires discipline, governance, oversight, and alignment with enterprise risk frameworks. The organization that gets this right will move faster because they can move with confidence. The organizations that ignore this will slow down, or worse, create exposure. So the goal is not to slow AI adoption, the goal is to enable it responsibly. Because when AI is governed properly, it doesn't increase risk, it actually can reduce it. And that's where procurement can lead. Now let's start to bring this all together because one of the most common questions procurement leaders ask is this where do we actually stand today and what does good look like? And the best way to answer that is through a maturity model. Because AI adoption and procurement is not binary. You're not either using AI or not using AI, you're somewhere along the curve. And understanding where you are on that curve is critical to knowing what to do next. So let's walk through the five basic levels of AI maturity and procurement. Level one is manual reactive procurement, and this is where many organizations started, and some are still here today. This environment looks like heavy reliance on email, using spreadsheets to drive analysis, limited visibility and to spend, contracts stored in shared drives, and reactive supplier management. Work is manual, fragmented, and time consuming. Procurement is often seen as a processing function, a cost center, and a gatekeeper. At this level, well, AI is not even part of the conversation yet. Level two is digitized procurement, and this would be where most organizations are today in 2026. Core systems are in place, ERP, P2P, e-sourcing tools, CLM. Processes are standardized, structured, and documented. You have visibility and dispend, controlled workflows, and improved compliance. But here's the limitation. Decisions are still manual, insights are still reactive, and the systems, well, they're only as effective as the people operating them. This is digitization, it's not intelligence. Level three is AI augmented procurement. And this is where leading organizations are today. AI is embedded into key processes: spend classification, contract analysis, invoice automation, intake orchestration, and supplier monitoring. What changes here is not just efficiency, it's decision support. Procurement teams now have faster insights, better visibility, improved risk detection, and more time for strategic work. Work begins to shift from manual execution to intelligent guidance. This is the inflection point because once you reach level three, you start to see the real transformation. Level four, let's call it semi-autonomous procurement. This is where we're heading over the next 24 to 36 months. At this level, AI doesn't just assist, it begins to execute within defined boundaries. You'll start to see autonomous processing of low-risk purchases, AI-driven supplier recommendations, contract redlining aligned to policy, real-time negotiation intelligence, and automated sourcing workflows. Human involvement is shifting to oversight, exception handling, and strategic decision making. The system handles repeatable, data-driven, and low-risk activities. And this is where procurement becomes faster, leaner, and significantly more scalable. Level five is autonomous procurement ecosystems, which is a future state. So let's look ahead. At level five, procurement operates as an intelligent ecosystem. AI can predict demand, initiate sourcing, evaluate suppliers, negotiate within parameters, execute contracts, and monitor performance. All with minimal human intervention. Now, let's be clear, most organizations are not here today, not even close, and they won't be for several years. Because level five requires high data maturity, strong governance frameworks, deep organizational trust in AI, fully integrated systems, and AI technology that, well, to be honest, it doesn't even exist today. But this is the direction of travel. And the key question becomes: where should you focus? For most organizations, the priority is not jumping to level five. It's moving from level two to level three, preparing for level four. That means embedding AI into core workflows, building confidence in AI outputs, establishing governance, training teams, and cleaning data. Because once you do that, scaling becomes much easier. So what's the biggest mistake organizations are making today? Well, let me call it out really clearly. The biggest mistake is trying to leap ahead, trying to implement autonomous sourcing, fully automated procurement, and end-to-end AI without the foundation. When that happens, trust breaks down, adoption slows, and value is lost. AI maturity is built, it's not installed. So this maturity model, it isn't theoretical, it's a practical tool, and it will adjust over time. Leaders can use it today to assess their current state, align stakeholders, prioritize investments, build a roadmap, and measure progress. And most importantly, to set realistic expectations. The takeaway from this is simple. AI and procurement is not a switch that you turn on, it's a journey. And the organizations that understand where they are and move forward deliberately are the ones that will capture the most value. Because the goal is not to chase the future, it's to build toward it step by step. So at this point, we've covered a lot. We've talked about what AI Can do today, what's coming next, what's real, and what's hype, how the operating model is changing, and how maturity evolves over time. But now the question becomes: what should you actually do with all of this? Because understanding AI is one thing. Executing on it, well, that's something very different. And this is where many organizations are getting stuck. They see the opportunity, they're excited, but they don't know where to start. So let's break this down into a practical roadmap. Step one is fixing the foundation. Data comes first. Before anything else, you have to address your data. AI is only as good as the data behind it. If your data is fragmented, inconsistent, unstructured, or outdated, then AI will not create value, it will amplify confusion. So the first priority is clean your supplier master data. Standardize your spend taxonomy, structure contract metadata, and align system of systems of record. This isn't exciting work, but it is essential because without it nothing else can scale. Step two is to standardize your playbooks and your policies. You have to define how procurement makes decisions, because AI needs guidance, it needs boundaries, it needs rules to operate within. That comes from contract fallback libraries, risk tiering frameworks, approval thresholds, category strategies, and supplier evaluation criteria. If those don't exist or they're not standardized, AI has nothing consistent to apply. So before you automate decisions, you have to define them. Step three, and this is really, really important because I'm guilty of it, a lot of people are guilty of it. Start small, but start smart. One of the biggest mistakes organizations make is trying to do too much at once. They try to implement AI across sourcing, contracts, P2P, supplier management, all at the same time. And what happens? Well, complexity increases. Adoption slows and value gets diluted. Instead, start with one or two high-impact use cases. Some of the best starting points are contract intelligence, spend classification, invoice automation, and intake orchestration. Why? Because they're high volume, they're data-rich, and they're relatively structured, which makes them ideal for AI. Prove the value there first, then expand. Step four is build AI capability within your team. Technology alone is not enough. I've said it before, AI is a tool. Your team has to evolve with it. Procurement professionals need new skills: data interpretation, analytical thinking, AI interaction and prompting, and risk evaluation in an AI-driven environment. This doesn't mean turning procurement into data scientists, but it does mean becoming more analytical, more strategic, and more comfortable working alongside intelligence systems. Because the future of procurement is human plus AI, working together. Step five is establish your governance early. Don't wait to think about governance. Build it in from the beginning. That includes defining where AI can be used, setting approval thresholds, establishing human in-the-loop controls, creating audit and traceability standards, and aligning with legal, IT, and risk teams. This is especially critical in regulated industries because governance doesn't slow AI down. It enables it to scale safely. Step six is align procurement with the enterprise. AI and procurement does not exist in isolation. It touches finance and legal, IT, risk, and all business stakeholders. So alignment matters. You need to ensure data flows across systems, that policies are consistent, that risk frameworks are integrated, and that technology decisions are coordinated. Procurement becomes part of a broader enterprise AI strategy, not a standalone initiative. Step seven is measure what actually matters. Too many organizations focus on adoption metrics, system usage, number of AI tools deployed, but that's not what matters. What matters are outcomes. Are you seeing faster cycle times, higher savings realization, reduced risk exposure, improved stakeholder experience, and greater team capacity? Those are the metrics that define success. According to McKinsey, organizations that combine AI adoption with operating model change capture two to three times more value than those that focus on technology alone. Two to three times. Wow. What does this look like in practice? Well, if you step back in your own organization and look, a successful AI roadmap in procurement kind of looks like this. You start with clean data, clear policies, and focused use cases. You build capability within your team and governance around your processes. You align across the enterprise, and you measure real business outcomes. And maybe most importantly, there's a mindset shift required. AI isn't a project, it's not a one-time implementation, it's an evolution, a transformation. Leaders need to think in terms of continuous improvement, iterative deployment, scaling over time. Because the organizations that win are not the ones that move the fastest. They're the ones that move deliberately and consistently. So the path forward is clear. Start with the foundation. Focus on what works, build capability, govern responsibly, and scale over time. Because AI will not transform procurement overnight. But over the next two to three years, it will fundamentally redefine it. And the leaders who act now will be the ones shaping the future. So, what does this all mean in the long run? For decades, procurement has fought to be seen as strategic, moving from order taker to sourcing partner to value driver. AI helps accelerate that journey because when you remove manual work, data gathering, and process friction, what's left is the part of procurement that actually matters the most: judgment, strategy, influence, and relationships. AI doesn't replace procurement. It reveals what procurement can be. The procurement leader of the future is not defined by process expertise. They're defined by their ability to interpret data, their ability to influence decisions, to manage risk, to unlock value across the enterprise. They're part strategist, part operator, part technologist, and part advisor. And increasingly, they're expected to lead in areas like AI adoption, data governance, supplier innovation, and enterprise risk alignment. Procurement will no longer be a support function. It's becoming a central decision engine within the enterprise. I talk about organizations that win, and those that will win are not the ones that simply buy AI tools. They're the ones that rethink how procurement operates. They invest in their people, they build strong data foundations, they govern intelligently, and they adopt AI deliberately. They don't chase hype, they build capability and they scale it. Now on the other side of that, there's a real risk. Organizations that delay, that take a wait-and-see approach, that assume AI will mature before they act, well, they're gonna find themselves behind. Because what's happening right now is not slowing down, it's accelerating. Vendors are advancing rapidly, competitors are adopting quickly, expectations from stakeholders are increasing, and procurement is expected to keep up. But here's the balance. You don't need to transform everything overnight. You don't need to implement every tool, you don't need to jump to full autonomy. What you need to do is start. Start with one use case, one process, one capability, build confidence, demonstrate value, and expand from there. That's how transformation actually happens. So if you're leading procurement today, your role is not just to adopt AI. Your role is to guide your organization through this transition, ensure it's done responsibly, and position procurement to lead, not follow. Because procurement is uniquely positioned. We sit at the intersection of spend, suppliers, risk, contracts, and business demand. There are very few functions with that level of visibility and influence. And AI amplifies that position. So here's the bottom line: AI is not the future of procurement. AI is the present. And over the next few years, it will define how procurement operates, how value is created, and how organizations compete. Thanks for joining me for this episode of the Procurement Brief. I have to admit this was one of the more fun episodes that I had in mind. If you found this valuable, please like the video, subscribe to the channel or the podcast platform, and share it with others in your network. And I'd love to hear from you. Where is your organization on this journey and what challenges are you seeing? What's working and what's not? Until next time, I'm Patrick Bliss.