By Dr. Upmith Singh (Principal Advisor for Policy & AI, CBC), Maryanka (Consultant, CBC) & Zainab Naaz Jamali (Consultant, CBC)

Introduction

India’s administrative journey towards citizen-centric service through e-governance spans nearly five decades of structured digital evolution. Beginning in 1976 with the establishment of the National Informatics Centre (NIC), the public sector has progressively integrated computing into governance, advanced through the National e-Governance Plan (NeGP) in 2006 to enhance accessibility and citizen-centricity (Geetika & Pandey, 2013) and transitioned via the e-Kranti framework to recognize that technological reform requires deep-seated Business Process Re-engineering (BPR) rather than mere digitization (Ministry of Electronics and Information Technology, 2015).

Even before AI entered the picture, research on India’s e-governance experience found that institutional readiness depended less on infrastructure than on human capital and administrative capacity to use digital tools effectively (Sebastian & Supriya, 2013)  a finding echoed in global benchmarks such as the United Nations Online Services Index. That distinction matters more, not less, as artificial intelligence enters public institutions. By capacity, we mean something broader than individual digital literacy: it is the combination of workforce skill, institutional systems, data, workflows, oversight structures  and leadership judgment needed to deploy a technology responsibly rather than merely adopt it. The question this piece asks is not whether government has adopted AI, but whether it has built this fuller sense of capacity and what that capacity must look like for government to become not just digital, but adaptive.

From Process Redesign to Role Redesign

Historically, administrative modernization focused on moving manual workflows onto digital portals to eliminate paper usage and improve operational metrics. However, true citizen-centricity requires rethinking how services are delivered, not merely relocating them online. International benchmarks, such as Portugal’s Virtual Citizen Shop, demonstrate that citizens should be freed from navigating complex bureaucratic structures to access public entitlements.

In the AI era, this redesign extends beyond process to role. Business Process Re-engineering (BPR) evolves into AI-augmented workflows:

• Operational Transformation: Initiatives such as the CPGRAMS “Samadhan Didi” chatbot demonstrate how AI-enabled speech recognition and classification streamline grievance routing between citizens and relevant departments.
• Strategic Oversight for Senior Personnel: For senior administrators and Subject Matter Experts (SMEs), AI does not replace domain expertise; rather, it automates routine information consolidation, allowing leadership to focus on high-order policy interpretation, assumption testing, and accountability.

This is not confined to senior administrators. The nature of public work is changing at every level. An official who once spent hours manually consolidating reports can now use AI tools to summarize information, identify patterns, and flag anomalies. The role does not disappear. its value moves toward higher-order cognitive tasks:

• Interpreting complex outputs and questioning the assumptions underlying them.
• Determining policy interventions based on contextual nuance and domain expertise.
• Maintaining human accountability within public decision-making loops.

The objective, therefore, is not to turn every official into an AI specialist, but to cultivate a workforce capable of critically using, questioning, and governing intelligent systems within their own domains. Redesigning processes without redesigning roles risks automating inefficiency rather than removing it,  a version of the same gap that limited NeGP’s early years, when technology adoption outpaced managerial and institutional change (Geetika & Pandey, 2013).

What Makes a Government “Adaptive”

If digitization is about moving services online and BPR is about redesigning how they are delivered, adaptiveness is a further step: a government that can sense changing citizen needs and reconfigure itself to meet them, without waiting for a new scheme or portal to be built. Estonia’s proactive government services illustrate this in practice. Rather than requiring citizens to apply for entitlements around life events such as a birth, marriage, or retirement, the Estonian system triggers eligible services automatically, shifting the administrative burden from the citizen to the state’s own workflows (Information System Authority, n.d.).

This is the distinction the AI era sharpens for India’s public institutions. A digital government makes existing services easier to access. An adaptive one anticipates needs, reallocates attention to where judgment is actually required, and revises its own processes as circumstances change, without eroding human accountability for the outcomes. Capacity, in this sense, is not a support function for adaptiveness; it is what makes adaptiveness possible at all. Digital Public Infrastructure (DPI) and Digital Public Goods (DPG)  supplies the scalable technical baseline, but it is institutional and human capability, contextual judgment, assumption-testing, and continuous learning that activates it, as the next two sections examine.

Mission Karmayogi: Operationalizing Role-Based Capacity

The Government of India launched Mission Karmayogi (National Programme for Civil Services Capacity Building) in September 2020 (Department of Personnel & Training, 2020), marking a strategic shift from a static, rule-based framework to a dynamic, role-based competency model,  an institutional attempt, in other words, to build exactly the kind of capacity described above.

Scale and Progress

The iGOT Karmayogi platform institutionalizes learning at scale, connecting professional development directly to official responsibilities:

• User Engagement: Over 1.7 crore registered users and more than 14.42 crore course completions across 23 languages (Press Information Bureau, 2026c). At this scale, the platform has clearly succeeded in making training accessible; what it does not yet show is whether that access has changed how officials actually perform their roles.
• Emerging Technologies: Over 218 specialized courses in AI and frontier technologies, recording over 7.3 crore enrolments and 3.35 crore completions (Press Information Bureau, 2026b). The scale of enrolment suggests strong demand for AI literacy  but demand and competence are not the same thing.

This is the limit of input metrics: course completions demonstrate reach, not capability. Administrative evaluation must look beyond how many officials completed a course, towards whether their workplace performance and governance outcomes actually changed as a result an evaluative gap that has long challenged large-scale public training programmes generally, not Mission Karmayogi specifically.

Institutional Readiness and Governance

AI readiness cannot rest on the individual official alone; it requires an enabling organizational ecosystem - the institutional half of the capacity this piece has argued for throughout. Training institutions are aligning through collaborative frameworks such as UNNATI,  the Unified New-Age National Training Institutions portal, a digital backbone developed by the Capacity Building Commission under Mission Karmayogi (Capacity Building Commission, 2023). UNNATI moves institutional accreditation and quality assurance away from periodic, portal-based assessment cycles towards a real-time, technology-enabled architecture, bringing training institutions’ faculty, infrastructure, and courses onto a single collaborative marketplace with continuous performance benchmarking rather than one-off evaluation.

The shift from periodic to continuous assessment is itself a small instance of adaptiveness:  an institution redesigning how it evaluates itself, rather than waiting for the next accreditation cycle to catch up. By fostering visibility, resource-sharing, and alignment with competency requirements across Central Training Institutes and State Administrative Training Institutes alike, UNNATI is intended to transform India’s training ecosystem into a connected, evidence-driven network rather than a set of institutions working in silos.

Alongside this, the Karmayogi Quality Framework (KQF) maintains rigorous course-quality standards through agentic AI applications and other tools (Capacity Building Commission, 2026a). But frameworks and platforms are necessary, not sufficient: public institutions must also maintain the leadership capability and human-in-the-loop oversight to govern the automated systems these frameworks introduce, otherwise adaptiveness in evaluation risks outpacing accountability in decision-making.

Conclusion: Toward an Adaptive State

India’s digital transformation has evolved from early computerization to robust Digital Public Infrastructure and, now, to the early architecture of AI-enabled governance. But the shift this piece has traced from digital to adaptive is not complete simply because the technology exists. A digital government can be built with infrastructure and platforms alone. An adaptive one cannot as it requires officials capable of questioning what a model tells them, institutions capable of redesigning their own workflows and evaluation cycles, and evaluation systems capable of measuring outcomes rather than activity.

Mission Karmayogi has built much of the infrastructure this shift requires  a role-based competency model, a learning platform at national scale, and institutional frameworks like UNNATI and KQF to sustain it. The open question, and the one worth carrying forward, is whether the next phase of capacity building measures itself by what officials complete, or by how government actually behaves differently as a result. That is the real test of whether India moves from a digital state to an adaptive one.

AI disclosure: The author used ChatGPT to support editing, improve clarity, and refine the structure of this article. The ideas, analysis, and conclusions are the author's own, and the final text was reviewed and approved by the author.

References

• Agência para a Reforma Tecnológica do Estado. (2025, November 13). Loja de Cidadão Virtual já está disponível com 150 serviços e apoio por videochamada. Government of Portugal.
• Capacity Building Commission. (2023). National Standards for Civil Services Training Institutions (NSCSTI) & UNNATI portal framework. Government of India.
• Capacity Building Commission. (2026a, May). Adopting the Karmayogi Quality Framework (KQF) attributes in the AI ecosystem of iGOT. Government of India.
• Department of Personnel & Training. (2020, September). National Programme for Civil Services Capacity Building (NPCSCB): Mission Karmayogi (Cabinet Approval Summary). Government of India.
• Geetika, & Pandey, N. (2013). National e-governance plan revisited: Achievements and road ahead. International Journal of Public Sector Performance Management, 2(2), 183–201.
• Information System Authority. (n.d.). Proactive government services. Republic of Estonia. Retrieved September 19, 2026, from https://www.ria.ee/en/state-information-system/personal-services/proactive-government-services
• Ministry of Electronics and Information Technology. (2015). e-Kranti: National e-governance plan 2.0. Government of India.
• Press Information Bureau. (2026b, August 5). Lok Sabha written reply on Mission Karmayogi and AI-related courses on iGOT. Ministry of Personnel, Public Grievances and Pensions, Government of India.
• Press Information Bureau. (2026c, August 12). Parliament question: Progress under Mission Karmayogi (Release ID 2298141). Ministry of Personnel, Public Grievances and Pensions, Government of India.
• Sebastian, M. P., & Supriya, K. K. (2013). E-governance readiness: Challenges for India. IIM Kozhikode Society & Management Review, 2(1), 31–42.