The long-awaited Procurement Act is scheduled to come into force in February next year promising a more flexible regime. Professor Albert Sanchez-Graells picks out some of the challenges it will pose for those working in public-sector procurement.
After a recently announced 4-month delay to its ‘go live’ date, the Procurement Act 2023 is now due to enter into force on 24 February 2025. This provides additional time for public buyers, private vendors and their advisors to prepare for the changes brought by the Act. Many of those changes are of direct relevance to tech procurement. And many pose as many challenges as they overcome rigidities or perceived problems with the current EU-derived regime. In this piece, I highlight the changes and challenges at the top of my list.
Ambition behind the new Act
Procurement reform has been long in the making. It was kick-started in December 2020 by the Johnson government in a context mired in Brexit-related promises to have a bonfire of red tape. Almost four years on, the functional goals of the reform have slowly evolved and crystallised around three main pillars: flexibility and innovation, transparency, and raising standards. A system of procurement regulation based on increased discretion to maximise the flexibility of new (minimal) rules adoption of an open data approach to procurement transparency, and increased professionalism and commercially oriented practices is envisaged. This direction of travel has not persuaded everyone, and Scotland will retain its current EU-derived rules, thus marking the start of significant regulatory divergence across the UK’s four nations.
Several implementation challenges and open questions arise from the tension between that vision and the current state of procurement practice in most of the UK. At present, the vast majority of contracts are awarded through centrally-managed commercial vehicles such as framework agreements, and the procurement function is under-resourced and suffering from a shortage in digital skills. Whether the cultural (and resourcing) shift required to reap the benefits of increased flexibility and discretion will take place remains to be seen. I harbour significant doubts, not least due to the limited training and upskilling programme that underpins the rollout of the Act. For most public sector buyers this is limited to 10-hours of self-learning and a dense training manual, while there is no provision other than minimal knowledge drops for the private sector.
The policy goals of the reform are also in a state of flux. While the previous Conservative governments articulated them through two consecutive National Procurement Policy Statements (NPPS, in 2021 and in 2024), the new Labour government thinks the vision they encapsulate ‘does not meet the challenge of applying the full potential of public procurement to deliver value for money, economic growth, and social value.’ A new ‘mission-oriented’ NPPS is thus in the making to ensure that ‘the new regime in the Procurement Act goes live with a bold and ambitious Statement that drives delivery of the Government’s missions’: hence the 4-month delay to the initial go live date. In my view, this move exaggerates the importance of the NPPS, and perhaps underplays the many implementation challenges required to adopt a ‘mission-oriented’ approach to procurement.
There are few publicly available details on the now much-awaited refreshed policy orientation, but my impression is that the tech procurement underpinning the digitalisation of the public administration and public services will have a prominent role to play in relation to several strands of policy delivery. Earlier this year, the National Audit Office reported widespread experimentation and piloting with AI in government and this trend seems set to accelerate under the policy direction of the Department for Science, Innovation and Technology: they have already singled out AI adoption to support delivery across government and to transform citizens’ experiences of interacting with the state as two key areas for the future AI Opportunities Action Plan.
So how will the Act facilitate innovative tech procurement, and which challenges remain?
Flexibility for tech procurement under the Act
Procedural flexibility and deregulation are some of the key planks of the Act. The Act moves away from the predefined tendering procedures currently in place under the Public Contracts Regulations 2015 (and other procurement laws for utilities and concessions) and limits competitive tendering to two procedures: open and competitive flexible. It is the latter which brings all the novelties. Competitive flexible procedures come to replace (or bring under a single umbrella) restricted, competitive with negotiations, competitive dialogue and innovation partnership procedures under the current rules.
Crucially, as stressed in guidance,
‘The competitive flexible procedure allows the contracting authority the freedom to design its own procedure. The contracting authority may choose to incorporate numerous processes into the procedure, such as negotiation, dialogue or a demonstration stage.’
In short, the competitive flexible procedure is very much a ‘do it yourself’ placeholder that allows each contracting authority to tailor procedural design to each specific project. While it seems unlikely that contracting authorities will move away from current practices in most cases, tech procurement could be an area where different approaches can be trialled.
There will, though, be significant challenges in implementing a compliant procedure that involves new approaches or high levels of flexibility and discretion, as the Act places significant disclosure obligations around the procedure, phases, and criteria to be applied from the beginning. This is an area where creating added value by facilitating ‘more commercial’ (or agile…) approaches to tech procurement will have to be carefully balanced with ensuring the robustness, integrity and legal soundness of the procedures. The overall design of competitive flexible procedures will need to remain proportionate to the object of the procurement and the participatory burden on tech vendors—to avoid the disproportionate participation costs and complexity that ultimately favour large operators in industries where there is already a tendency towards concentration and limited competition. Issues around tech prescriptiveness, organisational conflicts of interest, and the avoidance of commercial capture will, in my view, require particularly careful consideration.
Unaddressed challenges for tech procurement, especially of AI
One ever-present challenge in tech procurement concerns technical design and evaluation. The rules on technical specifications and award criteria tend to be open-ended and leave significant areas of discretion (and associated uncertainty) unaddressed. In practice, this tends to be minimised by making use of technical standards, where they exist, or ‘state-of-the-art’ clauses where permissible. The Act does not introduce significant changes on these issues (see eg guidance on technical specifications), although it does omit the reference to technical standards in an apparent push towards setting performance and functional requirements. While this omission may increase, or at least keep unchanged, the level of technical flexibility afforded to tech procurers, it can be problematic in some contexts. In particular, AI procurement.
Despite the quick acceleration in AI adoption (or experimentation and piloting towards it) and the likely further push in this policy direction, the Act and the associated guidance do not provide any new rules or criteria. Public buyers are still left to rely on the generalities of the Digital, Data and Technology Playbook, the insufficient Guidelines on AI Procurement, and a growing but scattered body of related guidance, such as the Guide to Using AI in the Public Sector and the more recent Introduction to AI Assurance. Depending on the use case, public buyers also need to consider the ICO’s Guidance on AI and data protection. Crucially, there is still limited cross-over between procurement as the ‘tech gateway into the public sector’ and public sector duties and obligations, especially when the tech is to be deployed in citizen-facing services and operations. This poses an additional challenge for the ‘responsible’ adoption of tech in the public sector which, in my view, carries increasing risks of mass harms.
Moreover, the existing guidelines, and those offered by private standard-setting organisations such as ISO or IEEE, are formulated at rather high levels of abstraction, are not easy to implement, and there is no single point of reference facilitating structured decision-making prior to launching the procurement of AI or automation solutions. This is a flawed regulatory strategy in my opinion; a more considered approach to public sector controlled technical standardisation and granting use case permissions is required.
For now, initiatives such as SCL’s own Artificial Intelligence Contractual Clauses can support some of the more technical and commercial aspects of such structured decision-making and careful consideration of which standards could be used (eg amongst those included in the AI Standards Hub). However this approach to tech procurement still places significant emphasis on a need to ‘reinvent the wheel’ every time. Given the technical complexity and uneven technical and commercial capability among buyers, the public sector remains exposed to regulatory capture and commercial determination.
Final thoughts
The Procurement Act 2023 will enter into force shrouded in the promise of flexibility and expectations it will play a key role in the process of public sector digitalisation to unlock the opportunities of digital technologies, such as AI. However, maximising the flexibility embedded in the new Act will require high levels of commercial and digital competency, a shift in culture and a willingness to take (limited) risks to do things differently. It will also need industry buy-in to these new approaches. From a broader governance perspective, and in the absence of a much-needed different regulatory approach, ensuring the public interest embedded in tech procurement is properly safeguarded will require foresight and probity by all those involved, including external and in-house legal and technical advisors. This appears to be an area where readers of this Magazine can make a difference for the better.
Albert Sanchez-Graells is a Professor of Economic Law at the University of Bristol Law School. He specialises in public procurement, competition, and digital regulation. Albert is currently a member of the Cabinet Office’s Open Contracting Steering Group and the NHS Independent Patient Choice and Procurement Panel. The standard disclaimer applies.
