Certainty of outcomes: building confidence in delivery

Certainty is something we all crave. The ability to predict and control our environment gives us confidence and a sense of safety – valuable qualities in a world that’s unstable and ever-changing.

But what happens if you have no choice but to operate in an environment where certainty feels elusive? All too often, that can be the case for digital and AI transformation in the public sector. Even when teams do their due diligence, establishing clear objectives and outcomes from the outset, the reality is that on average it takes up to seven years to deliver a digital project on the government’s Major Projects Portfolio (GMPP).¹ In such timeframes, technology evolves, requirements shift, and external circumstances can change dramatically.

Perhaps wisdom from ancient Greece offers a valuable perspective here. The Stoic philosophers understood the importance of focusing our energy on the things we can change and accepting what we cannot. This insight may hold the key to managing modern digital and AI transformation projects effectively.

With anticipated rapid advancements in AI over the next decade, as reasoning and predictive capabilities continue to progress and surpass current milestones, it is not far-fetched to imagine that the uncertainty we face today may seem stable in comparison to the potential disruptions AI could bring in the future. While certain aspects, such as human nature and behaviour, will remain unchanged, there are areas in which we will retain the ability to exercise control.

How can public sector leaders navigate this challenge, balancing the need for certainty with the realities of long-term digital change?

1. Investing the time upfront in a strong procurement strategy pays off in the long run 

A great metaphor used by the Stoics to bring their philosophy to life was that of an archer trying to hit a target. As soon as the arrow leaves the archer’s bow, hitting the target falls outside of their control. However, in the lead up to that moment, there are so many elements that are within the archer’s control – including selecting the best bow and arrow. In digital and AI transformation, taking the time upfront to select the right supplier, route to market, and procurement outcomes is one way to select the best bow and arrow, so to speak.

Procurement strategies should be as much about creating desired behaviours as they are about creating desired outcomes. The greater the level of change and uncertainly, the more that behaviours will matter more than any originally conceived outcome.

Recent analysis by the National Audit Office (NAO) reveals that, on several large digital change programmes, the approach taken to working with suppliers was a major factor in the programme running into difficulties, and contributed to more than £3 billion of cost increases. The same report found that the public sector makes considerably less use of outcomes-based contracts than the private sector, which limits suppliers’ ability to provide solutions to underlying business problems.²

To overcome these challenges, and evolve to an approach that enables government to gain greater certainty and value from their supplier relationships, we make three recommendations:

• Aligned incentives and learning objectives. Creating a less adversarial relationship between the public sector and suppliers should be a priority. Instead of relying on short-term, fixed-price contracts, government should explore partnerships that encourage long-term collaboration and closer alignment with suppliers. Suppliers should be encouraged not just to deliver initial requirements but to continuously improve and innovate, with government evaluating the value added by suppliers more holistically. This could include the potential for contract extensions if certain criteria – including service quality, level of improvement over the life of the service, and overall value for money – are met. In our age of data and AI, aligning on how both parties learn and what it is they want to learn and improve as the programme progresses is of paramount importance.
• Focus on what can be controlled and what trends are already in play. It’s important to recognise that designing and delivering new digital and AI solutions involves considerable complexity, which is rarely fully understood upfront. Instead, organisations should focus on what they can control, such as problems to be solved and outcomes to be delivered, and allocate ongoing funding to achieve those. AI will reduce the marginal cost to deliver a line of code, exactly how much and by when is less important than knowing the direction of travel. Such trends need to be factored into procurement strategies that will last multiple years.
• Blend digital and commercial expertise. Commercial teams often have insufficient digital and AI expertise to drive optimal terms, while digital teams tend to lack the capability to properly manage vendors. By bringing together a team where these skillsets are represented and managed effectively, organisations can bring greater transparency and control to the procurement process, ultimately leading to better outcomes.

2. Make use of platform-based technology where appropriate

Previous government Digital, Data and Technology (DDaT) manuals and the Technology Code of Practice (TCoP) advocated for the use of open-source technology and avoidance of vendor lock-in. However, taken to an extreme, this can create a technology landscape that is highly duplicative, with some 190 different authorisation services currently used across government,³ for example. 

We believe that public sector departments should carefully consider the trade-offs between a platform-based architecture and a pure open-source approach when implementing digital transformation projects. While both approaches have their merits, a balanced strategy leverages the strengths of each. When used appropriately, platform-based design enhances a digital transformation project's ability to respond to changing circumstances and business requirements. Modern platforms offer scalability and modular architectures, enabling independent development, testing, and deployment of components. This leads to faster iteration cycles and quick adaptation to new business needs. We recommend a hybrid approach, based on the following priorities:

• Adopt platform-based solutions for core functionalities. Leverage established platforms like Salesforce or Microsoft Dynamics for critical business processes where standardisation and reliability are paramount. This can accelerate implementation and ensure ongoing support.
• Use open-source components for customisation. Integrate open-source solutions for specific, unique requirements that platforms may not adequately address. This allows for greater flexibility and innovation where needed.
• Develop an interoperability strategy. Ensure that platform-based and open-source components can seamlessly integrate through well-defined APIs and data standards. This approach aligns with the UK Government's Interoperability Standards.
• Invest in internal capabilities. Build in-house expertise to manage both platform-based and open-source solutions, reducing dependency on external vendors and fostering innovation and learning.
• Regular evaluation and adaptation. Continuously assess the effectiveness of the chosen architecture, being prepared to adapt as technology evolves and government needs change.

Consideration needs to bear in mind the transformative impact generative AI will have on the balance of buy versus build. As the marginal cost to deliver a line of code trends towards zero, what can be done in house and what should be done in house will both change.

3. Good delivery management is a valuable skill – and government departments need more of it. 

The rapidly changing nature of technology, complex requirements and user needs, and integration with legacy systems often mean that digital and AI projects in the public sector start with a higher number of “unknown unknowns” (issues or challenges that couldn’t have been anticipated at the start), compared to traditional non-technology projects.

Strong delivery management plays a crucial role in ensuring the smooth progress of projects in situations like these. It enables the early identification of risks and development of mitigation plans to avoid disruptions. Delivery managers also facilitate collaboration between cross-functional teams, drawing on expertise from a range of disciplines to effectively solve problems when they arise.

Yet, delivery management skills are often in short supply within public sector organisations. As highlighted in the recent State of Digital Government Review, digital and technology teams across the public sector are struggling to build the in-house skills they need to manage delivery effectively. Of the £26 billion public sector digital and data spend in 2023, less than 20% went to permanent public sector staff, while 55% was spent on externals.⁴

For digital and AI transformation initiatives to succeed, government must do more to attract and retain skilled specialists. Relying on external contractors isn’t just expensive, it means that organisations often lack the knowledge and capabilities to deliver and utilise digital and AI effectively. Attracting and retaining digital and AI delivery management skills within government will require the same attention to compensation and career progression as other digital roles, as outlined in the recent State of Digital Government report. However, one aspect that is particularly important for delivery management is the need to truly recognise and champion it as a distinct craft. Often, delivery management is mistakenly conflated with traditional PMO roles, and its scope can vary greatly depending on the specific role and context in which it is applied. As a result, defining the role of a delivery manager can be challenging. Yet, anyone with experience working on a digital or AI project can easily identify the absence of effective delivery management and the impact it has. The DDaT professional framework offers a strong foundation for defining the role, but some of the structured, formulaic approaches to recruitment within the civil service may hinder government departments from attracting top-tier delivery management talent. In delivery management roles, experience, instinct, and mindset can be more critical than knowledge of specific methodologies or toolsets – and civil service recruitment practices may need to be updated to reflect this.

The advent of enterprise AI also means that the opportunity cost of failing to build internal knowledge and data through programme delivery activity will have a compounding detriment to organisations, whilst further empowering retained third parties.

4. Build on the experience and knowledge of others

When government departments share their experiences and insight, it can have far-reaching impact. The State of Digital Government Review cites how the West Midlands Police Force shared their approach for adopting AI to respond to 101 calls with other forces to great effect.⁵ However, initiatives like these are still the exception, rather than the rule, in the public sector. In the age of enterprise AI, any failure of the public sector to pool data and learnings will represent even greater lost opportunity than it has to date.

What if there were a way for government departments to access data and insights from others’ experiences of delivering similar digital and AI transformation initiatives? With information, for instance, on how long a particular project took, the costs, pitfalls, and lessons learned, organisations could approach their own transformations with much greater confidence and certainty. Reliable prediction relies on structure, consistency and correlation in the underlying data. The public sector has an incredible opportunity to act as one to create delivery datasets of prestigious depth and breadth.

The newly formed Digital Centre for Government has an ambition to make this a reality. It aims to enable government departments and teams working on digital and AI programmes to feel “confident to work in the open, being open about the challenges and learning in public”.⁶ The new centre had a clear remit for setting cross departmental standards and bringing together previously disparate parts of government. Furthermore, the government’s focus on ‘missions’ is intended to reduce previous silos in government, and provide ‘a new way of doing government that is more joined up, pushes power out to communities and harnesses new technology’.⁷

At Baringa, we see this as a promising step forward in the sharing of knowledge and expertise across government departments. The opportunity to exchange knowledge, build meaningful connections, and collaborate to solve challenges is something that our clients regularly tell us makes a difference in achieving successful project outcomes

Getting closer to certainty

Rather than dreaming about how things might be in an ideal world, the Stoic decides to engage with the world as it is. The AI practitioner engages with the world as it could be, by laying the foundations for data gather than will enable generational learning and compounding of experience.  Whilst there is (and always will be) uncertainty in the delivery of digital and AI transformation in public sector, by focusing on the things that can be controlled – like procurement practices, technology choices, data recording and ownership, and delivery expertise – organisations can lay the foundations for successful change.

Stay tuned for our final article of this series, where we’ll explore how public service organisations can achieve lasting digital change, by focusing on the continuous optimisation of outcomes.  

To learn how Baringa can help you gain greater certainty and control over public sector digital and AI transformation efforts, contact us.

Find out how to build confidence in digital and AI transformation for the public sector

^{1. Major projects in government – Institute for Government}

^{2. Government's approach to technology suppliers: addressing the challenges - NAO insight}

^{3. State of digital government review - GOV.UK}

^{4. A blueprint for modern digital government (HTML) - GOV.UK}

^{5. Mission-driven government – The Labour Party}