For decades, procurement has been urged to move up the value chain. And we‘ve made a lot of progress. Two decades ago, much of the function relied on paper, phone calls and email and was mired in operational tasks like reconciling invoices, processing purchase orders, and approval workflows. Today teams are much slicker. Cloud platforms and automation have freed procurement teams from administrative drudgery, allowing them to focus on supplier relationship management, risk mitigation, and increasingly, sustainability commitments that would have been afterthoughts in the early 2000s. Data analytics now informs category strategies and negotiations in ways that were simply impossible when procurement operated on spreadsheets and instinct.
Yet many procurement organisations are still lumbered with huge backlogs and stuck in firefighting mode. They’re running sourcing events, managing exceptions, chasing stakeholders, and battling systems that don’t talk to each other and require too much manual intervention. Our customer Invesco shared on a recent podcast that their procurement inbox was receiving 12,000 emails a year, with 80% being internal messages from team members simply trying to ‘move processes along manually’ or dealing with ad hoc cries for help.
Agentic AI is the first tech breakthrough to come along that truly helps solve these issues by handling complex, multi-step operational work, accurately, at scale. AI isn’t a magic fix, though. Many core procurement processes were designed 40–50 years ago for a different reality, and those legacy structures must be transitioned away from in stages.
Introducing a new framework for AI transition in procurement
As with any transformation, procurement requires a practical framework to evolve beyond constant “firefighting” toward the holy grail of strategic value creation. Based on our work with some of the world’s leading G2000 enterprises, we have devised an AI Maturity Model that serves as a blueprint for this transition without necessitating additional headcount or increasing procurement workloads.
First stage: Foundation and enabling phase
One of the core pain points of many organisations today is capacity. If procurement is to elevate into that truly strategic part of a business, driving growth and top-line performance, it starts with creating some space. The way to create that space is not necessarily about doing everything differently. It’s about asking: how can we handle the existing pipeline or workload in a smarter, more automated way?
Therefore, the first order of business is tackling those notorious backlogs; a problem all procurement teams share and one ideally suited to AI. As our customer, Cristian Martin, Director of Procurement, at the world-renowned London School of Economics said in a recent interview: “I can’t look at every piece of paper that goes through the finance department, as the scale is just too much,” adding that they would need 4,000 staff to do so manually.
In this transformation stage, AI is deployed mainly to handle foundational efficiency and automation. AI starts to handle routine administrative processes like data collection and document management, without altering existing workflows. These initial automations shorten cycle times and free up staff to oversee more company spend having to hire more people. These early improvements to data quality and productivity start to convince stakeholders that change is worthwhile. Indeed, LSE reported early on in their transformation that they could process twice as many tenders at a 15% cost saving.
It’s worth noting that implementing AI differs from traditional enterprise software, where the norm is often “test until it’s perfect before going live.” With AI, the opposite is true: you deploy early so the system can learn, adapt, and improve through iteration. Data doesn’t need to be perfect upfront. As Rhonda Spraker Griscti, Executive Director, Digital Strategy & Global Process Lead at BMS, noted in our webinar: “Don’t wait until you have perfect data. If you wait, you’ll never get started.” You only discover the data structure you truly need once the AI is in action.
Second stage: Advanced optimization
Stage two is where it starts to get really interesting, because you’ve created some space and can start to think a little differently about category strategy, sourcing strategy, and how the procurement function is adding value to the business.
With the procurement team’s time freed up, the critical question is: “What should the humans focus on?” This is where it pays to forget history and consider where human intelligence adds the most value. One key area is applying an “investment mindset” to the supplier base. AI can equip procurement teams to evaluate a far broader range of variables than just cost. This gives them the data and confidence necessary to strategically partner with suppliers and ecosystems that can foster sustainable growth, enhance brand reputation, and mitigate risk.
In practice, this means teams can move from wrestling with spreadsheets to asking the questions that lead to consistently better, more defensible decision making. Advanced optimisation and AI make it possible to express intent in plain language and translate it into models that weigh factors like cost, risk, ESG, and compliance side by side. Scenarios become easier to explore, trade-offs easier to explain, and governance easier to embed. Some decisions can be safely automated; others keep a human in the loop.
Third stage: Transformation and business partnership
The third stage of procurement transformation is about achieving true business partnership and value creation. With agentic AI insights, sourcing teams can start anticipating market shifts and guiding stakeholders with prescriptive, data-backed recommendations. Category strategies become living models, shaped by real-time signals, changing risks, and evolving market conditions. The aim is to get beyond just better decisions in the moment and build greater organisational resilience and sustainability over time. In this stage, procurement starts to deliver measurable strategic value, positioning itself as a core contributor to long-term innovation and stability.
The greatest threat to Stage 3 maturity is the enterprise tendency to add new AI tools without switching off old manual processes. This process layering occurs when staff continue to perform portions of the work “the old way” despite having automated alternatives. True maturity requires that technology, process, and organization evolve in tandem.
If you think achieving true business partnership is a pipe dream, consider this evidence. Grocery giant Tesco viewed the savings they achieved through agentic AI procurement so strategically that they called it out in their 2024 annual report. It stated that their “Save to Invest” program used the £1.2bn in savings not just to cut costs, but to “offset inflation and create headroom to fund investments.” Similarly, HP reported that their indirect procurement team is generating $0.08 EPS (Earnings Per Share) annually, driven by their ability to use AI to gain control over “previously unmanaged tail spend.”
Modern agentic AI procurement can truly evolve from being firefighters to strategic partners. These powerful examples illustrate that with the right structured approach to transformation, procurement can help deliver on a range of business goals from efficiency and productivity gains, to reinvestments, to boosting shareholder value.
Peter Wetherill
Peter Wetherill is VP EMEA at Globality, where he is part of the Customer and Go-To-Market team. Globality brings transformational automation and intelligence to the enterprise sourcing process for Global 2000 companies.
