Transforming public service delivery: AI, automation and trust in the modern Civil Service

How can the Civil Service respond to rising expectations, rapidly evolving AI capabilities and persistent organisational complexity?

Those were some of the big issues covered by ServiceNow’s Director of Enterprise Architecture for the UK and Ireland, Graham Williamson, in his keynote at Government Transformation North today (27 November), which focused attention on the operational realities behind the government’s digital front door. 

The session, Transforming Public Service Delivery: AI, Automation and Trust in the Modern Civil Service, opened with recognition of GOV.UK as a major achievement in simplifying citizen access to services. Behind that front end, however, Williamson described “a hornet’s nest of complexity”, which includes siloed teams, disconnected processes, fragmented data, legacy technology estates and a general reliance on staff to compensate for structural gaps. 

As Williamson put it: “Humans fill in the gaps and become the glue. Citizens become the glue of their own journey and experience across what are still fundamentally departmentally siloed services. Civil Servants become the glue in how they collaborate with one another”.

This operational delivery layer was presented as one of the critical areas in need of change. Around 60 per cent of the half-million civil servants work in this space, and it is where both citizen experience and staff experiences are shaped.

Williamson also addressed the fast-shifting AI landscape. The release of generative AI tools like Chat GPT in late 2022 introduced what he described as a step change in pace, capability and expectations. Technological shifts in the past have been substantial, but “there is something a bit different about what is going on in the AI space at the moment,” he said. 

He highlighted three core challenges raised by this shift: speed of change, concerns about trust and non-deterministic outputs, and the re-emergence of security vulnerabilities in new forms.

AI finds its lane in government

Despite these risks, he emphasised the growing maturity of commodity AI and the move away from bespoke or experimental systems towards scalable capabilities embedded in enterprise platforms that are supported by UK-based infrastructure. This trend is accompanied by an increase in small, domain-specific language models which reduce the need for departments to develop their own AI models in isolation. As Williamson put it: “Off the rack is quite acceptable”, particularly where budgets, skills and time are constrained.

Departments were encouraged to revisit AI strategies that may have been drafted in the early phases of experimentation. Strategies written 12 to 18 months ago are unlikely to reflect the current landscape of commodity AI, market investment or cross-government demand for consistent approaches.

Skills formed another element of the keynote. As more departments adopt common platforms and market-standard technologies, there is potential to develop digital and AI-related skills that transfer more easily across the Civil Service. Williamson said: “By investing in those skills, in your people now, there is benefit in the immediate. But it is also setting them up for success when they progress into the private sector or to their roles beyond the current situation”.

The legacy technology question

On legacy technology, Williamson drew on his years of experience working with public sector systems that remain operational long after modernisation efforts begin. He described legacy as a persistent, long-tail problem and noted that many organisations continue to treat it as a sequence of isolated technical fixes. 

Some organisations, however, have begun reframing their legacy programmes so they deliver wider transformation. Instead of simply replacing old systems, they are addressing the governance, funding structures and delivery models that originally created the legacy estate. This, says Williamson, is a way to “get more bang for your buck from the money that is already being spent”.

Given the ongoing nature of legacy challenges, Williamson outlined the need for multi-modal delivery. This involves operating existing services, modernising legacy systems, pursuing continuous improvement and enabling transformation at the same time. Doing so requires an enterprise architecture function that can bridge diverse systems and ways of working.

A number of public bodies are now establishing a digital transformation platform to support this. This platform model is intended to connect disparate systems, standardise shared capabilities and create a foundation for continuous change rather than large one-off modernisation efforts. Some organisations combine this with a digital spine or digital control tower to support data driven investment planning across transformation portfolios.

Williamson ended with five key actions for government digital leaders:

1. Address the messy middle of operational delivery. This is where service complexity is most visible to both citizens and staff.
   
   
2. Maximise the strategic value of legacy modernisation programmes. These programmes already represent major investment and can be used to drive structural improvements rather than isolated system upgrades.
   
   
3. Establish a digital transformation platform. Departments are encouraged to identify the core reusable capabilities that can join up silos and support long term change.
   
   
4. Revisit AI strategies. The rapid development of enterprise and platform based AI makes it necessary to update existing strategies and adopt a multi solution approach.
   
   
5. Strategically govern AI initiatives and invest in transferable skills. AI should be managed like any major organisational investment, with clear oversight of value and risks, while capability building must support both current and future civil servants.

ServiceNow has created an Enterprise AI Maturity Index to help organisations benchmark their capability. Find out more here