Copyright newstatesman
As the UK government accelerates its push into the use of AI, public sector organisations are rethinking how they deliver services, modernise legacy systems, and drive innovation at scale. Amazon Web Services (AWS) is at the forefront of this transition. Speaking to Deepak Shukla, head of AI and data for UK public sector at AWS, we unpack what’s behind the shift. UK government has increased its focus on Agentic AI. What do you think is driving this? The recent emphasis on Agentic AI is driven by the need to move from AI as a passive tool to an active, collaborative partner that can execute complex tasks autonomously. This represents a fundamental shift in capability, from answering “what” to accomplishing “how” and “do”. The government-led push for advanced AI adoption is driven by several key factors. One major shift is the evolution from simple automation to action-oriented intelligence. While Generative AI excels at producing content – such as text, code, and images – in response to prompts, it typically functions in single-turn interactions. Agentic AI marks the next step forward: systems that can autonomously plan, sequence, and execute multi-step tasks to achieve specific goals. This transition moves AI from a “tell me” model to a “do for me” paradigm. Improving public service delivery is another critical driver. Governments are under pressure to enhance service quality, boost productivity, and modernise outdated processes. Agentic AI plays a vital role by transforming time-consuming, manual workflows – like conducting regulatory compliance checks, or reviewing policy documents – into automated, efficient operations. These intelligent agents can significantly reduce turnaround times and improve responsiveness to citizens’ needs. Agentic AI also supports better decision-making through more effective data utilisation. These systems can operate across departmental boundaries (complying with existing data sharing agreements), integrating and analysing data from multiple sources. This ability to synthesize information at scale leads to more informed, timely decisions – essential for everything from community safety to economic development. Perhaps most importantly, the shift is toward AI that doesn’t just respond but proactively assists. Agentic systems can retain memory of goals and context, adapt based on experience, and navigate across various tools. With these capabilities, they function as persistent, intelligent assistants – executing complex tasks autonomously within defined safety parameters. The UK government’s focus on AI is driven by the recognition that these systems are not merely an incremental improvement, but a transformative technology capable of autonomously delivering on core government missions: to deliver growth and improve public services. How does AWS support government leaders to drive Enterprise AI adoption across departments? AWS empowers government leaders to navigate the complexities of enterprise AI adoption by providing a comprehensive strategy built on three core pillars: a secure and integrated technology foundation, deep hands-on expertise, and strategic investment in innovation. We envision a future where AI agents act as collaborative partners, orchestrating sophisticated, multi-step processes to improve decision-making and productivity. However, we recognise that building such agents—which must be secure, trustworthy, and seamlessly integrated with legacy systems—presents a significant challenge for large organisations. AWS addresses these challenges directly through a comprehensive support structure designed for the public sector. A secure and interoperable foundation is the starting point. AWS provides the platform and tools agencies need to build agents that meet stringent government standards. These agents can operate across all data sources, integrate with both proprietary and open-source systems, and utilise the latest foundation models – all while maintaining rigorous authentication, authorisation, and compliance controls. This combination of built-in security and flexibility is essential for deploying reliable agents at scale across diverse government departments. AWS also offers direct access to AI expertise. Through the AWS Generative AI Innovation Center, government teams are paired with AWS AI experts to help accelerate their journey from concept to deployment. Following two years of demonstrated success, AWS recently committed $100m globally to expand the centre – specifically enhancing support for agentic AI development. This initiative gives agencies the practical guidance needed to overcome technical barriers and deliver effective solutions. A notable example is AWS’s collaboration with Syngenta to create Cropwise AI, where agents analyse complex environmental data to support agricultural decision-making. How do you see AI agents addressing challenges in citizen services, such as long waiting times and inconsistent access to information? Government departments are under pressure to deliver high-quality, 24/7 citizen services amdist difficult circumstances. Traditional communication channels, may become stretched, leading to challenges with public satisfaction. Generative AI-powered solutions—including advanced chatbots, virtual assistants, and intelligent AI agents—present a transformative opportunity to address these inefficiencies. These technologies can understand natural language, retrieve relevant information from vast knowledge bases, execute multi-step workflows, and generate personalised responses. This allows for automated, end-to-end task completion, dramatically improving citizen access to services and reducing response times. AI agents can directly mitigate these challenges by introducing several key capabilities. Automated classification and routing allows intelligent systems to instantly analyse incoming requests and accurately direct them to the appropriate department or individual – eliminating bottlenecks caused by manual sorting. This is complemented by optimised resource allocation, where routine triage and processing tasks are automated – freeing staff to focus on complex, high-value activities that require human judgment. Accelerated response times are another benefit. By automating initial data collection and processing steps, agencies can significantly reduce the overall time it takes to fulfill a request – improving efficiency and citizen satisfaction. Finally, these systems are built with flexibility in mind. When faced with undefined or complex cases, agents can seamlessly escalate issues to the appropriate human experts – ensuring no loss in service quality while maintaining operational speed. With governments facing issues with increasingly obsolete legacy systems, what role can AI developer agents play in accelerating modernisation? AI developer agents are a pivotal force in overcoming the immense technical debt and productivity losses caused by legacy systems. They serve as a strategic accelerator for modernisation by automating the most complex, time-consuming, and costly aspects of the process. The scale of the challenge is significant. As highlighted in the State of Digital Government report, archaic technology costs the UK public sector an estimated £45bn in annual lost productivity savings, hindering the delivery of efficient public services. AI developer agents directly address this by automating high-effort, low-risk development tasks. A prime example is Amazon Q Developer, a generative AI-powered assistant for software development. Its proven impact includes: unprecedented efficiency: Amazon has migrated tens of thousands of production applications from older Java versions to Java 17 with its assistance. Also quantifiable savings: this effort represented savings of over 4,500 years of development work compared to manual upgrades. And finally, tangible value: the upgrades delivered performance improvements worth an estimated $260m in annual cost savings. For government departments, the opportunity is clear. By leveraging Al agents as part of services such as Amazon Q Developer and AWS Transform – the first agentic Al service for transforming NET, mainframe and VMware workloads – agencies can dramatically accelerate the pace of legacy application modernisation, achieve substantial cost savings by reducing thousands of person-hours of manual effort; improve system performance and security by efficiently migrating to modern, supported platforms; refocus precious developer talent on innovating new citizen services rather than maintaining outdated code. In essence, AI developer agents transform legacy modernisation from a daunting, multi-year financial burden into a manageable, efficient, and highly valuable strategic initiative. They are not just a tool for developers but a critical lever for CIOs to drive significant fiscal and operational improvements across the entire government enterprise. AI agents are increasingly being used in research, healthcare, and genomics. What opportunities do you see for public health systems to use this technology? AI agents present a transformative opportunity for public health systems to move from reactive care to proactive, predictive, and personalised population health management. The technology’s ability to autonomously analyse vast, complex datasets is unlocking new potentials in disease prevention, genomic research, and operational efficiency. Real-world implementations demonstrate this impact. In collaboration with AWS, Genomics England built an AI-powered tool that processes millions of pages of scientific literature to identify associations between genetic variants and medical conditions. This has dramatically accelerated research, with the system already identifying 20 potential clinically-relevant associations for intellectual disability faster than manual review alone, paving the way for improved genetic testing and health outcomes. Another example is the U.S. Food and Drug Administration (FDA), which leverages AI to accelerate the review of life-saving pharmaceuticals, reducing scientific review processes that once took days down to minutes. And beyond pure science, AI agents drive massive operational efficiencies. Brazil’s Federal Prosecution Service, for example, uses this technology to save an estimated 40,000 hours annually in case processing by automating routine tasks, freeing professionals for higher-value work. The overarching opportunity lies in deploying AI agents to act as force multipliers across the public health ecosystem. They can be deployed to, for example, predict and model disease outbreaks by analysing real-time health data, environmental factors, and population movement; personalise public health initiatives by identifying at-risk cohorts for targeted interventions; and manage complex, multi-step administrative and research workflows autonomously. What’s next for governments adopting Agentic AI? The next phase for governments adopting Agentic AI involves transitioning from experimental pilots to secure, scalable, and governable production systems. This shift represents a tectonic change in how software is built and interacted with, moving from tools that require explicit commands to proactive partners that act autonomously across digital boundaries. One major advancement is bridging the pilot-to-production gap. Services like Amazon Bedrock AgentCore play a pivotal role – offering a set of composable solutions that enable organisations to evolve AI agents from limited prototypes into robust, trusted applications. These solutions are designed to scale effectively – supporting millions of users and addressing the operational complexity involved in deploying reliable agents at scale. Another key innovation is the revolution in developer tools. The adoption of next-generation integrated development environments – such as Kiro – is transforming how public sector teams build with AI agents. These tools streamline the developer experience, shorten development cycles, and allow IT teams to move faster on projects that embed autonomous capabilities into mission-critical software. In essence, what’s next is the industrialisation of Agentic AI. The focus will move beyond mere capability to ensuring these systems are deployable, manageable, and auditable. Governments that embrace this next generation of development platforms and tools will be best positioned to accelerate innovation, improve productivity across departments, and deliver more responsive and effective citizen services. Agentic AI on AWS | AWS content hub | AWS One Government Value Agreement (OGVA) – Innovate with confidence
