By Hervé Legenvre
This article examines how Vodafone Procure & Connect (VP&C) has built and deployed AVA, an AI-driven sourcing and procurement platform, across its global operations. We highlight the platform’s architecture, independently audited performance outcomes, and the co-development model. We also explore how VP&C has approached adoption and change management, and we conclude by reflecting on what the VP&C experience reveals about the future of AI in procurement. Vodafone Procure & Connect is the 2025 winner of the Industry Shaper & AI Pioneer EIPM Peter Kraljic Award for Excellence.
In a world of permanent change, regulatory complexity, and data proliferation, procurement must operate with a greater speed, insight and consistency. Artificial intelligence offers a response to this challenge, not merely by automating tasks, but by reshaping the nature of procurement work itself.
Looking at AI in procurement, promise and practice remain separated by a persistent gap. Many documented deployments remain narrow in scope, limited to specific use cases and activities, and few organisations have confronted the full complexity of implementing AI at enterprise scale. Scale is not simply a matter of adding more users. It requires resolving questions of data, process, people engagement, and change that simple pilots never encounter.
Scale is not simply a matter of adding more users. It requires resolving questions of data, process, people engagement, and change that simple pilots never encounter.
VP&C is one of the rare cases where the question of scale has been confronted directly and systematically. VP&C manages €25 billion of annual spend across thousands of suppliers and multiple operating companies. Its recent deployment of the AVA platform provides an unusually rich opportunity to examine what AI-enabled procurement transformation looks like when it is pursued with the ambition, rigor, and commitment required to deliver results with significant business impact.
The Procurement Imperative: Why AI and Why Now
VP&C identified converging forces that made a digital-first procurement model an operational necessity. Market volatility demands agility. Regulatory and ESG obligations require continuous monitoring across a significant supplier base. Data complexity has grown beyond what traditional processes can handle. And cost pressures require teams to deliver faster, more measurable outcomes.
VP&C’s response is articulated through a clear digital vision: “connected and digital everywhere”. The vision is structured around three capabilities: access to data, making data speak, and making data act. What distinguishes VP&C’s approach is that it was designed from the outset as a platform serving the full breadth of a global procurement operation.
Performance Outcomes: Impact at Scale
The financial and operational outcomes of VP&C’s AI deployment are substantial and independently audited. Since April 2025, AVA has processed ~600 sourcing projects and handled 3 billion euros of competitive new sourcing. Within the first year, the platform delivered 15% additional savings on top of business-as-usual projections, confirmed by two independent finance audits. This result is particularly significant because it was achieved above an already strong baseline: VP&C’s cost savings already ran ahead of world class benchmarks before the AI deployment began.
Adoption metrics are equally compelling. All new sourcing events now take place through the AVA platform, with over 10,000 suppliers using the system globally, and thousands of Vodafone users actively engaged. This matters because adoption is where most digital transformation programmes fail. Technology that sits unused cannot deliver value.
Operational speed has also transformed markedly. Sourcing cycle times have nearly halved. Contracting time has reduced by 43%. Supplier onboarding and risk assessment, a process that once took seven days, now takes two. Across all procurement activities, VP&C monitors over 35 speed metrics digitally, creating a discipline of continuous measurement that keeps improvement visible and accountable. The aggregate impact is a 56% faster delivery time driven by the impact of automation and AI.
“Within the first year, the platform delivered 15% in additional savings above projected targets, confirmed by two independent finance audits.”
The Ava Platform: Architecture and Use Cases
The AVA platform is VP&C’s proprietary AI and digital procurement platform. It is scalable, integrated across the full sourcing-to-payment cycle and designed to work with different ERP systems and procurement tools e.g. SAP S4 Hana or Ariba Guided Buying. Rather than relying on a single monolithic technology suite, VP&C built AVA as an open architecture that can that allows it to retain ownership of the intelligence layer.
The platform currently supports more than 40 AI use cases in development or live deployment, from a broader roadmap of more than 200 identified by procurement teams. These include guided RFP generation, supplier Q&A automation, automated scoring and contract management. Each use case targets a specific point in the sourcing process where AI can improve speed, consistency, or insight.
VP&C structures use cases according to a risk-effort matrix with three categories. In “assist” use cases, AI provides recommendations while people retain decision authority. In “augment” use cases, AI supports more complex decisions such as supplier evaluation. In lower risk areas, greater automation can be used for repeatable tasks such as purchase order processing or catalogue management. This architecture helps ensure that AI is deployed with governance and human oversight.
The breadth of the roadmap is striking. Across demand management, sourcing, contracting, procurement operations, catalogue management, supplier management, stakeholder management, and analytics, VP&C has mapped specific AI applications with expected impact ratings. High-impact use cases include RFP generation, contract redlining, purchase requisition assistance, supplier risk management, and the “Happy Circuit” an automated process that re-negotiates spot buy quotes in real time. The ambition is not incremental improvement, but a more strategic redesign of how procurement is done.
Co-Development With a Start-Up: A Model for Procurement Innovation
One of the most instructive aspects of the AVA programme is how it was built. VP&C did not rely solely on a large technology integrator. Instead, it entered a co-development relationship with a specialist start-up, combining Vodafone’s procurement expertise and scale with the start-up’s agility and technical depth. Crucially, VP&C retained ownership of the intellectual property.
This model created a different innovation dynamic. Large integrators often optimise for scope management and risk mitigation. Start-ups, by contrast, optimise for learning, and iteration and speed. By co-developing AVA with a start-up, VP&C gained access to rapid development cycles while embedding structured governance around the process.
Managing Change: Building Confidence and Capability
Technology alone does not transform organisations. VP&C’s change management approach has therefore been as important as the platform itself. It focuses on three layers: engaging users to drive adoption, building capability through targeted training, and co-creating AI solutions with procurement teams.
Category managers were also closely involved in shaping the platform. They created blueprints for their categories that shaped how the AI workflows operate. As a result, AVA was not simply introduced, it was built with them. That co-creation helped strengthen adoption and ensured the platform reflected real operational needs.
The mechanics are disciplined. Weekly AVA review meetings assess progress and adoption. Net Promoter Score (NPS) tracking measures satisfaction from among users and suppliers. A structured change request process has already processed 380 user requests within the first six months, showing that practitioners are actively shaping the tool around their work. The NPS indicators trend positively, confirming that the platform is not merely tolerated but genuinely valued by its users.
The change programme also creates alignment between VP&C and local execution. VP&C secured formal commitment from country managers to local targets, helping ensure that the benefits of AI are embedded in day-to-day procurement operations rather than remaining a central programme ambition.
A key operational benefit is the reduction of manual effort. VP&C expects digitalisation to reduce repetitive manual workload by around 30% across the function, particularly in sourcing and contracting. Automation does not eliminate procurement professionals. It creates capacity for higher-value activities that require human judgment: stronger supplier negotiations, innovation, co-development and stakeholder management.
Scaling the Use of the Platform Across the Full Company
AVA has been deployed across all Vodafone operations for local long-tail sourcing, extending the platform’s reach to the full diversity of local market contexts. The purchasing platform is also live for over 9,000 purchasing users across all local markets, enabling approximately 90,500 purchase orders annually.
These deployment decisions have implications that extend beyond operational efficiency. They signal that VP&C has moved from being a procurement function to being a platform with procurement at its core. AVA is a potential competitive offering, a procurement-as-a-service infrastructure that other organisations can access.
Deploying AI for a single use case is different from deploying it across a global function.
The two-year ambition VP&C has set for itself is ambitious but grounded. It targets a further 30% reduction in manual effort, a 40% increase in speed of operations, and an additional 15% in savings attributable to GenAI across IT, corporate services, networks, and commercial services. These targets are not arbitrary. They follow from the trajectory already established by the first year’s results and from the pipeline of use cases that have already been designed and prioritised.
The Future of AI in Procurement: What Vodafone Procure and Connect’s Experience Tells Us
VP&C’s experience offers several lessons seeking to scale AI in procurement.
First, scale requires architecture. Deploying AI for a single use case is different from deploying it across a global function. The technology, governance, and change management approach need to be designed with scale in mind from the beginning.
Second, co-development with practitioners improves adoption and outcomes. Category managers who help design AI workflows understand them more deeply, trust them more readily, and use them more effectively.
Third, independent validation matters. VP&C’s decision to commission two finance audits of its AI-generated savings, strengthens credibility in an environment where AI benefit claims are often treated with scepticism. Audited outcomes give the programme weight and help sustain commitment.
Fourth, procurement and technology strategy are increasingly connected. By building an owned AI platform, co-developing with a start-up and extending AVA into broader procurement contexts, VP&C has positioned procurement as a source of strategic capability, not just operational delivery.
Ultimately, transformation is measured not by deployment alone, but by outcomes. The ~600 sourcing events processed, the 3 billion euros of competitive sourcing executed, the audited savings delivered, the 10,000+ suppliers engaged are not technology metrics. They are business results. Vodafone Procure & Connect has demonstrated that AI in procurement can deliver measurable value at scale, and that efficiency, when combined with the right governance and adoption model, can strengthen rather than diminish the role of procurement.
