From gate to shaper — designing AI that delivers

How, at what cost, with what outcome

Why this matters

Most organisations are past whether to incorporate AI. The question on every CPO's desk is how the function should — and what that says about procurement's place in the corporate AI agenda. Across EBG's three procurement surveys 2025–2026 the pattern holds: most sit at "developing" maturity, only around 5% have integrated AI into decisions, and few report material hours back. Spending is up, outcomes are modest. The CPO-level question isn't "did this investment work" — it's whether procurement is the function that asks the design questions early, makes the trade-offs visible, and can tell the CEO what we're getting back. Behind it sits an operating-model choice: control with procurement, autonomy with the business, and where the function-defining decisions actually get made.

Discuss with peers

• How is procurement positioned in your organisation's AI agenda right now — leading it, supporting it, or being routed around? And what's your intent for the next 12 months?
• Across the AI spend you've approved or shaped this year — what's the pattern in what's delivered, and what does it tell you about how to make these calls going forward?
• When the CEO or CFO asks what we're getting back — what answer do you want to give in twelve months, and what has to be true now to give it?

JSModerator: Jatin Singhal, Global Head of Category — Technology, The Adecco Group

From digitalization to decision-making

Designing risk solutions for the moment all the data becomes available

Why this matters

Orkla has built the structure — seven siloed risk processes consolidated into one system across 45,000 legal entities — and is now upgrading it with AI: triage of incoming supplier forms, automated certificate and document validation, supplier monitoring. The focus is how to organise the supplier-risk process — which applies well beyond risk — and how to apply AI to manage it at scale. As Orkla's own team notes, the field moves fast: what's true in May may have shifted by October. That is exactly why this is a room to compare design choices in real time, not a fixed playbook.

Discuss with peers

• For one specific risk type, which approval design have you chosen — centralised, delegated, local, or central assessor with local approval — and what made it the right fit for how your business actually works?
• Where have you turned, or would consider turning, a risk task into a working AI solution — triage, validation, monitoring — and what did designing it teach you about ownership and follow-up?

NAICModerators: Nils Holm Andersson, Director Digital Procurement & Development, and Isabel Carlebom, Head of Sourcing Process — Orkla Procurement

Predictive procurement — building the structure, finding the value

Contract intelligence in practice — a case-led session

Why this matters

Risk mitigation and cost efficiency are now hygiene. The harder question is how procurement shows value going forward — and how much of that value rests on individual depth versus what sits in the structure so it survives the person who built it. Dnyanesh Sarang is a sharp case on exactly that. Now Head of Global Supply Chain at Optinova, he built a working structure for predictive procurement on top of the contract estate: a method written as four pillars of risk and resilience, and an AI-enabled diagnostic he built himself that reads the estate and flags exposure ahead of time. In one case it surfaced a regulatory pricing escalation buried in a supplier agreement fifteen weeks before the supplier's own notification window — terms restructured before the increase landed, roughly fifty million dollars of revenue protected, visible on the finance scorecard rather than in a procurement report. Case-led by design: short case, the structure shown with its actual artefacts, the diagnostic shown being built, then structured Q&A. A lab like the others — just run case-led rather than open-floor, so you know what you're walking into.

Discuss with peers

• Where in your own function would predictive — risk, price, supplier, regulatory — actually unlock value going forward, and where are you still operating reactively?
• What part of that value rests on individual depth that would walk out the door tomorrow — and what is genuinely carried by structure and process that would survive?
• If you wrote the next role profile knowing this case, what would you ask for so the person hired can build both: the structure that lasts and the value that lands on someone else's scorecard?

DSModerator: Dnyanesh Sarang, Head of Global Supply Chain, Optinova

Compliance: define it, implement it, prove it

Who owns regulatory work when the rules keep moving

Why this matters

The regulation tells you to comply. It never tells you who decides what it means, who sets the scope, or who is left holding it when the deadline moves. And the deadlines do move: the 2026 Omnibus pushed CSDDD out to 2029 and cut the framework by two-thirds, the deforestation rules slipped again, and PPWR starts to apply on 12 August — two months before this room meets — with part of its detail still unwritten. Chase each rule on its own timeline and you never catch up. The leaders getting ahead answer a different question first: who defines and scopes this, who implements it, who verifies it — and where, honestly, the answer today is "nobody decided, someone just did." Not a chemicals problem, and not a sustainability-versus-cost fight. An ownership question that normally sits across procurement, sustainability, legal and regulatory affairs at once — and usually gets discussed in four separate rooms. Here it gets one.

Discuss with peers

• For one regulation you live with (CSDDD, PPWR, CBAM, REACH, internal due diligence) — who defines what it means for your business and sets the scope? A method, or a call someone made that no one has revisited?
• Are you chasing regulations one deadline at a time, or have you built a foundation — supplier, contract, material data — you can layer each new rule onto? If you started Monday, where would you begin?
• For one requirement you've put into a contract — do you verify suppliers actually meet it, or is it "in the contract and hope"? If you closed that gap, who would own the checking?

ABModerator: Alisha Bloodworth, Vice President Global Procurement, Perstorp

Who do you hire in 2027?

Writing the ad for the transformation AI holds as a promise

Why this matters

Picture procurement in autumn 2028. Agents inside your S2P stack run sourcing on their own — RFQ to award, no humans in the loop on standard categories. Your suppliers' bots negotiate with your bots. Live market intelligence on raw materials, geopolitics and supply risk lands in your category manager's inbox before she asks for it. Your team isn't bigger. It's fundamentally differently composed. The vendors are shipping the technology now; the analysts say mainstream adoption hits inside 24 months. If they're right, your next hire is the one who has to make that future work. So — who is that, and what do you write in the ad to find them?

What you walk out with

• Four to five capabilities you'd genuinely put in your next real job ad — each tied to the value it delivers, each something you could test.
• A sharper sense of what "good with AI" actually means in practice when you read it on a CV.
• A shared read across the room on where Nordic procurement thinks it's heading — and where it isn't.

TLModerator: Tobias Lindberg, Head of Regional Sourcing EMEA, Sandvik SRP

Foundation, speed, value — developing procurement tech at enterprise scale

Everyone says data is key, everyone says AI — what does that actually take in a large, complex organisation?

Why this matters

Every keynote, vendor and board deck says the same two things: data is the foundation, and AI is the future. But inside a large, complex organisation, actually delivering on them is a far more layered job than the conversation lets on. You put years of rigour into getting master data genuinely usable — and the moment it is, the message becomes "now move fast with AI," while the pilots meant to prove the case keep failing to land (MIT's 2025 research put that at 95% of enterprise GenAI pilots). The conversation has run ahead of reality almost everywhere: in EBG's 2025 survey, 74% still struggle with poor data and not one organisation called itself "leading." This isn't foundation-versus-speed, or pro- or anti-AI — it's a sequencing question few have answered well at scale: how much foundation before you move, and how to find value along the way so it earns its keep before it's ever "finished." There's an honest disagreement in it too — build the foundation first, or pilot forward and let value pull it along. Both have a point. Which holds at your scale is the conversation — the honest read from inside the most demanding environments, not the version polished for a stage.

Discuss with peers

• You've put years into the foundation — data, master data, the connected systems. Where has it actually unlocked speed or value, and where is it still "in progress" no matter how much you invest?
• Everyone is told to "do AI." For a company your size, what has actually delivered value — and what looked good in a pilot but never made it into how the organisation runs?
• Build the foundation first, or pilot forward and let value pull the foundation along? At the scale and complexity you operate in, which has actually worked for you — and what did it cost to find out?

KPModerator: Kristina Prilepchanska, Head of Procurement Solutions, Electrolux Group

Start from the process, not the tool

Where AI actually changes procurement — and the gap from here to there

Why this matters

Most procurement leaders meet AI as a tool question — which platform, which feature. The more useful question starts one step earlier, in the process itself: where in strategy, category work, source-to-contract or procure-to-pay does AI change what is actually possible — and what is the gap between where a function is now and where it is heading? Across EBG's surveys most rate their digital maturity as "developing," and few have concrete cases in production, which makes a real, working example the most useful starting point. Marjo Jokinen opens with one: an AI assistant for contracting at Yara, now scaling across contract types. From there the room maps a process stage of its choice — where each function sits today, where AI moves the needle, and what stands in the way. The focus stays on the process and the gap, not on a catalogue of tools.

Discuss with peers

• Part 1. Pick a process stage — strategy, category strategy, source-to-contract, or procure-to-pay. Where are you now, what are the short-term opportunities where could AI change what's possible, and where do you want to be in two to three years or in the long-term consideration?
• Part 2. What's blocking the move — data, skills, ownership, mandate? If you started Monday, what's the first thing you'd change?

MJModerator: Marjo Jokinen, VP Technical Procurement, Yara International

Contract Intelligence: Value Already Sitting in Your Agreements

From pre-negotiation standardization to post-execution value capture

Why this matters

Most procurement teams focus contract value creation solely on the initial negotiation, leaving substantial operational and financial gains behind once the ink is dry. True contract intelligence starts before execution with structured negotiation playbooks that pre-align best-case and minimum acceptable terms with major Business Stakeholders and Legal. Standardizing these boundaries across categories drives contract harmonization, accelerates cycle times by removing redundant legal reviews, and elevates baseline terms. Once executed, transforming these structured agreements into dynamic assets allows teams to continuously capture hidden margins, enforce indexation, and track performance commitments.

This roundtable focuses on practical strategies for operationalizing contract intelligence—from pre-negotiation standardization to post-execution value capture—before layering on advanced tools and AI.

What we’ll discuss

• How is your organisation standardizing negotiation playbooks across categories to pre-align legal boundaries, reduce cycle times, and improve contract quality?
• What concrete financial value (e.g., missed rebates, unapplied price indexations, missing penalties) are you currently extracting from existing agreements?
• How do you move your team’s culture from treating contracts as static legal safety nets to managing them as dynamic, value-generating assets?
• Where does AI and advanced tech fit into this operational foundation to scale contract intelligence across your entire portfolio?

KTModerator: Kathrin Thiesen, Senior Director Program & Production Procurement, BioNTech

Building the next procurement professional

A table for the next generation — when the entry-level path changes

Why this matters

Operational and analytical work have always been the two entry points that build a procurement professional from the ground up. They're also the first work AI absorbs — and in EBG's 2026 survey, 61% still develop their people the old way while only 5% have started building something new. Procurement hasn't been remade yet; the shift is only beginning. But the generation entering now isn't waiting for it to happen. They are the ones building the agents, cleaning the data and training their colleagues — carrying the freshest training and feeling the shift earliest. What has been missing is a room in the Nordics where they compare notes across companies and categories. This is that room: not a youth panel, not a place to vent, but a constructive read on what actually builds a procurement professional now, from the people already building it. Since 2010 EBG has asked good people what matters most to them right now — in 2026 that question goes to procurement's next generation.

Discuss with peers

• What has actually changed in your own work over the last twelve months — tools, tasks, what you're asked for — and is the shift real where you work yet, or still only talked about?
• What did your training and your way of working already see that the function hasn't caught up to — and what would you bring in first?
• What do you bring in that the function doesn't have yet — and what are you learning that the tools don't give you: judgment, validation, context? If the old ground-up path is changing, what builds a procurement professional now?

The whole Next Gen group at one table in the first lab run — the second run is your own pick. If the group is large, or enough of you want a second Next Gen round, the table runs twice · Group led, opened by one of the participants

Scaling without slowing down: what can we learn from each other?

One process, one digital solution, different realities

Why this matters

Most large procurement organisations have three groups touching the same process: the people who develop it, the people who build the digital tools that run it, and the line that lives with it every day. Where the mandate sits between them decides whether anything scales. Inter IKEA is working on exactly that line: process development and digital enablement now sit in one scope across the group's purchasing entities, with the line organisation implementing rather than inventing. Behind it are years of local entrepreneurship, where every team tweaked its own ways of working and harmonisation never landed.

Mårten Wennersten and Ankit Shah open with where Inter IKEA is on that journey: one process across categories, one digital solution, and the different realities it has to work in. Then the room compares how others have drawn the same lines, what it takes to change behaviour in two directions at once, in the line that wants to keep designing its own solutions and in the leadership above that keeps assigning it projects, and a newer problem: when the process is fully integrated and switching supplier is one click, who still checks whether the answer is right?

Discuss with peers

• When the process cannot have clean handovers, because specification, sourcing and supply keep iterating with each other, how do you structure the collaboration so it stays efficient? And do your KPIs measure the process, or still the function?
• Changing this means working in two directions at once: the line that wants to keep designing its own ways of working, and the leadership above you that keeps assigning it projects. Which direction has been harder, and what has actually moved behaviour?
• The more integrated and simplified the process becomes, the less the people running it feel whether the result is right. Where do you see that already, and what are you doing to keep the understanding behind the button?

MWASModerators: Mårten Wennersten, Business Process Owner, and Ankit Shah, Business Process Partner — Inter IKEA