At Zycus Horizon EU & UK, two independently commissioned research studies landed on the table simultaneously – one authored by Forrester Consulting, one by The Hackett Group, both commissioned by Zycus.
Different methodologies, different sample sets, the same essential conclusion: at the precise moment agentic AI is reshaping the source-to-pay landscape, most procurement organisations have not yet established who is accountable for how that technology gets deployed.
This is not a story of procurement versus IT. Both functions have indispensable roles. It is a story of clarity of ownership – and of what happens when that clarity is absent. It is also, critically, a story for chief executives and boards: the governance norms your organisation sets now will determine whether AI in procurement becomes a source of competitive advantage or a source of strategic drift.
Zycus, which commissioned both studies and built the Merlin AI Agentic Platform that sits at the centre of this conversation, has made a deliberate bet: that the procurement profession needs not just better technology, but better frameworks for leading it. The research is the evidence base. The May 13 summit is where that evidence becomes action.
The confidence gap
The Zycus – Forrester study — Don’t Delegate AI: Why Procurement Leaders Must Personally Shape, Not Surrender, AI-Driven Decisions – surveyed 261 global procurement leaders and found a structural disconnect at the heart of AI adoption.
| 69% Confident in their AI vision | 31% Confident in their team’s ability to execute | 38pts The execution gap |
That 38-point gap is where strategic AI ownership tends to migrate quietly — not because IT reaches for it, but because procurement hasn’t yet claimed it. IT’s involvement in agentic AI is not only appropriate but essential: this is a foundational platform area, like ERP or cloud infrastructure, where technology expertise, security standards, and integration architecture all require IT leadership. What the research cautions against is the CPO abdicating strategic judgement — allowing the ‘what’ and ‘why’ of AI to be determined by those who cannot fully see the commercial context.
“AI doesn’t just automate procurement. It encodes procurement judgement into systems.”
— Don’t Delegate AI — Forrester Consulting, commissioned by Zycus
When AI determines which suppliers get flagged, which contracts trigger review, which spend gets scrutinised – it is executing someone’s judgement at scale. The Forrester research, commissioned by Zycus to ground its platform development in practitioner reality, makes clear that judgement must belong to procurement. The top three value-leakage areas – requirement scoping, post-award compliance, and obligation tracking, each affecting 48% of organisations – are precisely where agentic AI is expected to deliver returns. But only if CPOs define the rules of engagement.
The barriers to scaling are equally telling: budget constraints at 49%, governance gaps at 43%, competing priorities at 43%. Forrester’s verdict is direct: these are not technology problems. They are leadership and ownership problems.
The adoption index: Where organisations actually stand
The Hackett Group’s Agentic AI in Procurement Adoption Index 2026, developed in partnership with Zycus, adds ground-level precision to Forrester’s strategic argument. Drawing from organisations where 78% report revenues above $1 billion and 60% manage more than $1 billion in annual spend, it maps enthusiasm against reality.
58% of procurement leaders expect agentic AI to have a high to very high positive impact. Only 2% anticipate little to no effect. The technology has believers. What it lacks, in most organisations, is the governance architecture to match the ambition.
Six findings every CPO should know
1. Awareness hasn’t cascaded. 75%+ of executives believe in AI’s value. Only 60% are confident their teams share that understanding.
2. Control confidence is low. Fewer than 50% of CPOs feel confident in their ability to monitor and control agentic AI systems.
3. IT and procurement both lead strategy. IT guides strategy at 58% of firms; procurement at 53%. Collaboration is the norm – but accountability must be explicit.
4. Workflows beat point tools. 65% prefer integrated agentic workflows over individual agents. Orchestrated autonomy, not isolated automation, defines the era.
5. Scaling is imminent; readiness is not. Deployment sits at 20–30% across S2P processes. Scaling within 12 months is expected. Governance maturity lags.
6. Tail spend is the proving ground. 86% are likely to deploy AI agents for tail spend negotiation autonomously. Only 36% are satisfied with current approaches.
The Hackett Group also identifies the highest-risk implementation areas: fraud detection, autonomous negotiations, contract management, payables management, and stakeholder management. Contract management appears on both the highest-value and highest-risk lists – which is not a contradiction. It signals that the areas of greatest potential are exactly where thoughtful CPO-led governance design is most critical.
The investment signals confirm that scaling is coming regardless of readiness. Study respondents expect to direct approximately 12% of their technology budgets toward agentic AI within three years. Spend analytics, purchase order processing, market and pricing intelligence, supplier onboarding, and contract management top the list of expected returns. Zycus has built the Merlin AI Agentic Platform specifically to address these S2P process areas – with pre-built, interoperable agents designed to work as orchestrated workflows rather than isolated point tools, directly reflecting the 65% preference Hackett Group surfaced in their research.
On governance – Hackett Group/ Zycus Research
The top governance concern: loss of human oversight and control, driven by unclear cross-functional accountability and poor data quality. The top compliance risks: data privacy, data security, and legal/regulatory compliance. These require a whole-of-enterprise response, set from the top.
A message for the C-suite and board
Both studies carry a signal that goes beyond the CPO’s desk. AI in procurement will shape how billions in third-party spend are managed, how supplier risk is monitored, how commercial negotiations are conducted, and how contract obligations are enforced across the supply base. These are not departmental technology decisions. They are enterprise-level stakes that belong on the board agenda.
CEOs and boards need to actively set the expectation that AI deployment in commercially sensitive functions operates under clear cross-functional accountability – not in a grey zone where strategy and execution belong to different owners by default. That means defining, at the enterprise level, who is accountable for what AI does, how decisions escalate when AI operates at the edge of its guardrails, and how performance against outcomes gets measured, reported, and reviewed.
The Hackett Group research is explicit that organisations have far greater maturity in general technology management than they do in strategy, governance, change management, and talent development. The infrastructure instincts are there. The organisational architecture to govern what AI does inside that infrastructure is, in most cases, still being built. Boards should be asking whether that architecture is being designed intentionally – or emerging by accident.
Organisations that will lead in this era are those where the C-suite creates the conditions for procurement and technology to collaborate with shared objectives, mutual accountability, and a governance framework robust enough to earn the trust of regulators, suppliers, and stakeholders.
The Agentic AI Procurement Summit · 13 May 2026 · Global Virtual
The research has been published. Now comes the practical question: what do I actually do on Monday morning? On 13 May 2026, Zycus hosts the first and only global Agentic AI Procurement Summit – a virtual event with dedicated streams for US, EMEA, and APAC audiences. The conversations that began in Vienna at Zycus Horizon EU+UK will now open to thousands of procurement executives worldwide.
| Aatish Dedhia, Founder & CEO, Zycus | Christopher Sawchuk, Principal & Global Procurement Advisory Practice Leader, The Hackett Group | Jeffrey Rajamani, Senior Analyst, Strategic Sourcing & Procurement, Forrester Research |
Christopher Sawchuk and Jeffrey Rajamani – the analysts who authored the defining research of this moment – will jointly map a practical adoption roadmap for procurement leaders at every stage of readiness. This is not a repeat of the Vienna conversations for those who missed them; it is the next chapter, taking the research findings into implementation territory. Practitioners from EY and IBM will address real-world implementation realities, sharing how governance frameworks, change management programmes, and technology integrations are being built inside large enterprises right now. A live demonstration of the Zycus Merlin AI Agentic Platform will show autonomous source-to-pay workflows in action – not slides, not hypotheticals. And delegates will leave with concrete frameworks for taking ownership of the AI agenda within their own organisations, applicable regardless of where they currently sit on the adoption curve.
ALSO AVAILABLE FREE — AGENTIC AI ADOPTION INDEX TOOL BY ZYCUS
Before the summit, find out exactly where your organisation stands. The Zycus Agentic AI Adoption Index benchmarks your readiness against real industry data from The Hackett Group study – giving CPOs a clear-eyed view of where they are and where leaders are pulling ahead. Access it free at zycus.com.
The bottom line
The Forrester and Hackett studies, both commissioned by Zycus to ground its platform development in practitioner reality, converge on a message that is urgent without being alarmist. Agentic AI is already inside the procurement function. It is reshaping spend analysis, purchase order flows, contract management, and increasingly, the strategic decisions that define commercial outcomes. The question is no longer whether to adopt. It is whether the adoption will be led or inherited.
The organisations that will succeed are not those who deploy AI fastest. They are those who deploy it with the clearest accountability, the most thoughtfully governed workflows, and the strongest alignment between the commercial judgement of procurement and the technical expertise of IT. That alignment requires leadership from the top.
“Successful agentic AI isn’t about blindly following the robots. It’s about humans and AI working together in partnership.”
— Jeffrey Rajamani, Senior Analyst, Forrester Research — Zycus Horizon EU+UK, Vienna
The roadmap is being written. Zycus – through its research partnerships, the Merlin AI Agentic Platform, and the May 13 summit – is providing the tools, the evidence, and the community to help procurement leaders write it on their own terms. The question is whether your organisation will help shape that roadmap, or inherit someone else’s version.