Indonesia to use agentic AI for smarter, faster policy analysis

The Indonesian government will start using agentic artificial intelligence (AI) tools for policy analysis, according to the Ministry of State Secretariat. 

This was expected to strengthen evidence-based decision-making for the Indonesian President and Vice-President, at the highest level of government. 

Unlike generative AI (GenAI), which only responds to user prompts, agentic AI can act autonomously, from scanning data sources, verifying information, and generating contextual analysis without direct human instruction. 

Speaking at a webinar titled Strategy for implementing Agentic AI in Government institutions on October 7, the Ministry of State Secretariat’s Head of the Agency for Technology, Data and Information, Gogor Oko Nurharyoko, noted that till now, policy analysis at the Ministry has been conducted manually. 

“Agentic AI is not meant to replace humans, but to strengthen the decision-making process,” he added. 

Nurharyoko said this capability could change how the government works from a traditional reactive model to a proactive one.  

With systems that automatically detect emerging trends and patterns, policymakers could receive analyses that not only describe current situations but also forecast future developments. 

“Imagine if the President received reports that include predictions and scenario modelling, policymaking would become far more proactive and targeted,” he added. 

This initiative aligns with the Ministry’s broader mission to reinforce evidence-based policy analysis, ensuring the President and Vice-President receive timely, data-driven insights to guide national governance. 

Real-time insights for presidential decisions 

According to Nurharyoko, agentic AI would enable real-time and cross sectoral insights that are crucial for policies that evolve rapidly on the ground.

He pointed to free nutritious meal programme – one of President Prabowo Subianto’s flagship initiatives – as an example. Traditionally, progress reports were compiled through bureaucratic processes that took time, even as citizens shared live updates on social media every hour.  

With agentic AI, data from social media and other public sources could be automatically filtered, verified, and translated into policy recommendations that reflect real-world conditions. 

“If we continue relying on the old methods, our analysis will always lag behind reality,” he noted. 

However, he cautioned that the path toward automated analysis would require careful groundwork, as one of the key challenges is data digitalisation. 

“This technology depends on digitally formatted data, while a large portion of government documents still exist in physical form,” he explained, adding that the Ministry is accelerating efforts to digitise archives, policy reports, and administrative documents to enable integrated analysis. 

To subscribe to the GovInsider bulletin, click here. 

Building the foundations for AI governance 

In the next session of the webinar, University of Indonesia’s computer science expert, Professor Yudho Giri Sucahyo, highlighted three essential foundations for adopting AI in government: data readiness, computing infrastructure, and human resources.

“Once the data, infrastructure, and talent are ready, only then can we talk about implementation,” he said. 

He emphasised that data was the most critical foundation, as AI could function without robust and reliable datasets.  

Prof Sucahyo underscored the importance of the One Data Indonesia initiative, which connects multiple databases through the government’s API gateway and establishes data exchange standards across institutions. 

These standards, he noted, were vital for defining which data is public, personal, or confidential – a long-standing issue that has hindered interoperability between agencies. 

The second foundation was infrastructure. Given AI’s high computational requirements, the government would require powerful and secure data centres to process and train AI models effectively. 

The third was human capability. Civil servants, he said, should not be passive users of technology, but active participants in its design, oversight, and ethical governance.  

To support this, universities and research institutions are ready to expand Indonesia’s AI talent pipeline. 

“This initiative will help boost the supply of skilled professionals amid rising demand for AI specialists in Indonesia,” he added. 

The University of Indonesia plans to launch a dedicated AI degree programme next year to accelerate skills development in this field. 

Ethics and human control remain paramount 

During the Q&A session, a participant from a government agency asked who would be held accountable if an agentic AI system made an incorrect decision.

Prof Sucahyo responded that maintaining the human-in-the-loop principle was critical. AI systems must never operate fully autonomously without human oversight. 

He emphasised that risk management, legal frameworks, and ethical guidelines were top priorities for almost every country. Indonesia too will soon issue a Presidential Regulation on the National AI Roadmap and AI Ethics Guidelines. 

“Ethics are just as vital as technical readiness. Without proper oversight, AI could lead to bias, privacy breaches, or unequal access to technology.”  

Another participant asked whether the government could use private cloud services for AI processing and how security could be guaranteed. 

AWS Indonesia Government Account Lead, Muhamad Yopan, responded that such concerns often arise from misconceptions. 

“What matters is not where the data is located, but how its security and compliance are managed. Cloud providers are generally better prepared to handle cyberattacks because they undergo continuous audits and security updates,” he said. 

He added that under the latest regulations from the Ministry of Communication and Digital Affairs, government data classified as restricted could now be processed in private clouds under strict supervision, enabling innovation without compromising security. 

Closing the discussion, Prof Sucahyo emphasised the importance of synergy between policy, bureaucracy, and technology. Often, the problem does not lie in the system itself, but in silo mentalities and leadership. 

“In that context, AI is not a threat, but a test for the bureaucracy to transform into a more adaptive and trustworthy institution.”