8. Workforce and Culture

Towards broad tech fluency, elite talent, and high-performance culture.

1. How It (Doesn’t) Work Today

The structures, cultures and workforce practices of the public sector have not kept up with the increasing dynamism of modern organisational forms:

Outdated career models and rigid HR structures: Government HR systems often prioritise lifetime careers, rigid job classifications based on credentials rather than agile skills, and slow, bureaucratic hiring processes. This makes it difficult to attract, develop, and retain talent with expertise in cutting-edge digital, data, and AI expertise, who are often undervalued and underpaid compared to the private sector. Career progression for technical specialists outside of traditional managerial tracks is often limited. 

Cultural inertia and misaligned incentives: A risk-averse, process-driven culture often stifles the experimentation and iteration that are essential for AI development. Political and bureaucratic incentive structures rarely reward long-term technology transformations.

Structural immobility: Unlike the private sector, where companies face constant competitive pressure to evolve, public sector agencies are infrequently reorganised and almost never eliminated. Institutional legacy accumulates. This allows institutional legacies to accumulate and outdated structures to persist, even when they hinder progress.

Alongside this structural stasis, governments around the world are contending with an aging workforce. In some countries, the average age of civil servants is in the mid-40s or 50s, with a large cohort nearing retirement over the next decade. This demographic shift presents a dual reality: While this opens an opportunity to redesign jobs as they turn over, it also means a lot of senior personnel did not come of age in a digital environment and may lack fluency with modern technologies, let alone AI.

2. A Vision for a Modern, Agent-Empowered Workforce

The future of work in an agentic government is one in which humans and AI collaborate in new and dynamic ways within organisations that are redesigned to be agile, to facilitate continuous learning and to create public value. This vision encompasses several interconnected transformations:

Democratising tech skills: Generative AI is dramatically lowering the skills required to engage directly with data, develop simple applications, and automate routine tasks. Natural language interfaces, AI-assisted coding, and intuitive data analysis tools make data-driven decision-making and rapid prototyping widely accessible. 

Attracting elite talent: Alongside democratisation, there is an increased need for highly skilled specialists to design, build, govern, and orchestrate complex core agentic AI systems. Attracting, developing, and retaining this scarce, ‘hyper-productive’ talent is a critical priority for any organisation, but particularly challenging and expensive for the public sector. 

Blended teams of AI and human colleagues: Agentic systems will not only augment, but also significantly transform and, in some cases, automate many existing public sector job roles. The latter is particularly true for those involving routine information processing, content generation, and standardised decision-making. This presents an opportunity for slimmer organisations but also means a massive challenge for retraining and reorganisation. All public servants will need to develop skills in ‘working with AI’, such as critically evaluating AI outputs, providing effective feedback to AI systems, understanding how to use AI ethically in their context, and collaborating seamlessly with AI teammates.

Transforming organisational forms: As agents take on more complex tasks, the structure and culture or organisations will evolve. While some functions may approach full automation, most will be hybrid, with human-AI teams working side by side. Competence matrices…

AI co-workers and agentic systems, potentially designed with insights from high-performing tech companies, can act as catalysts, encouraging public sector teams toward more agile, data-driven, and startup-like routines and norms. The aim is not to replace present working culture wholesale, but to use AI to accelerate the adoption and reinforcement of high-performance practices, such as rapid iteration, transparent data sharing for decision-making, and continuous feedback loops.

c. Key Questions 

How do we reskill, redeploy, or release staff at the speed of automation, while protecting the human experience of transition? What are realistic throughput goals for upskilling, and how can micro-credentials, transition pathways, and social protections ensure no one is left behind?

What talent strategy (and compensation package) can attract world-class AI engineers without alienating unions or budget-conscious voters? Can purpose, flexibility, and influence offset the public sector’s structural limitations on pay and perks? What alternative incentives, from impact visibility to intellectual ownership, can rival private-sector offers?

As AI agents take on more decision-making and information processing, how do we prevent over-reliance and preserve human judgment? How do we maintain the ‘cognitive sovereignty’ of both public servants and citizens?

How can we navigate professional gatekeeping as AI begins to challenge established roles in fields like education, medicine, and regulation? What strategies are needed to manage political resistance from professional bodies that view agentic systems as a threat to their status or authority? Should governments invest in new models for negotiation and hybrid credentialing to support the integration of AI into traditionally protected domains?