Governments Turn to Agentic AI, but Data Gaps Hold Back Progress Only 21% of organisations currently have sufficient, high-quality data to train and fine-tune AI models
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
Agentic artificial intelligence (AI) is set to gain wider use across government organisations in the coming years, according to new research from the Capgemini Research Institute.
Agentic AI, which can make autonomous decisions, is seen as a natural extension of this momentum as governments seek to streamline operations, enhance decision-making, and respond more effectively to complex social challenges.
The study reveals that 90 per cent of public sector organisations intend to explore, pilot, or roll out agentic AI within the next two to three years.
Meanwhile, nearly two-thirds (64 per cent) of public sector bodies are experimenting with or deploying generative AI (Gen AI) to strengthen service delivery.
Data as the weak link
Despite enthusiasm, data readiness remains a pressing hurdle. Only 21 per cent of organisations currently have sufficient, high-quality data to train and fine-tune AI models. While leaders see the potential for AI to address rising citizen demands with limited resources, weak data foundations continue to slow progress.
"With rising citizen demands and stretched resources, public sector organisations recognise the ways in which AI can help them do more with less. However, the ability to deploy Gen AI and agentic AI depends on having rock-solid data foundations," said Marc Reinhardt, Public Sector Global Industry Leader at Capgemini.
The study highlights that in the next two to three years, 39 per cent of organisations will assess feasibility, 45 per cent will begin pilot programmes, and 6 per cent expect to scale existing agentic AI initiatives. Sectors such as defence (82 per cent), healthcare (75 per cent), and security (70 per cent) are particularly active in adopting generative AI solutions.
Security, trust, and skills concerns
Adoption, however, is not without hesitation. As many as 79 per cent of executives cited data security as a leading concern, while 74 per cent expressed doubts about trusting AI-generated outputs. The report suggests that governments must create stronger governance frameworks and data infrastructures to build confidence in these technologies.
The research also points to slow organisational progress in developing AI skills and promoting data sharing. Only 12 per cent of respondents said they were very advanced in activating data effectively, and just 7 per cent rated themselves mature in building AI-related skills.
Although every organisation surveyed reported current or planned data sharing initiatives, most efforts (65 per cent) are stuck at the planning or pilot phase. Challenges related to cloud infrastructure, sovereignty of AI systems, and privacy concerns continue to limit progress, even though data sharing is central to improving model performance and policy outcomes.
Leadership roles expand
Governments are also turning to specialised leadership roles to address these challenges. The report notes that 64 per cent of public sector organisations already have a Chief Data Officer in place, while another 24 per cent plan to appoint one. Similarly, 27 per cent currently employ a Chief AI Officer, and 41 per cent intend to establish this position in the future.
Reinhardt noted, "Going forward, governments can be more agile and effective as AI augments the work of government employees to source information, conduct policy analysis, make decisions, and answer citizen queries."
