Bengaluru: Artificial intelligence should be measured by the time it saves people rather than the number of users adopting it, said Microsoft India and South Asia President Puneet Chandok at Bengaluru Tech Summit 2026.
Citing an AI-assisted lesson-planning tool deployed in Karnataka government schools, Chandok said the real measure of AI adoption should be its ability to reduce everyday workloads and deliver tangible benefits.
Speaking on the summit’s “AI & Beyond” theme, Puneet Chandok called for businesses to move beyond measuring AI adoption through user numbers and logins, and instead focus on its practical impact.
“Adoption is a login,” he said, contrasting it with “2 hours given back to a teacher” through Shiksha Copilot, an AI-assisted lesson-planning tool developed by Microsoft Research India in collaboration with Sikshana Foundation.
Puneet Chandok also urged business leaders not to wait for AI to “settle down” before beginning to adopt the technology.
Puneet Chandok Highlights AI’s Impact in Karnataka Classrooms
Chandok cited Shiksha Copilot as an example of how artificial intelligence can address practical challenges in education.
The tool helps teachers prepare lesson plans in English and Kannada, incorporating examples drawn from children’s surroundings to make classroom content more relevant.
He referred to a study involving more than 1,000 Karnataka teachers and recalled the experience of a science teacher with 20 years of experience, who described the tool as “burden releasing”.
The research offers a measurable account of the time saved through the initiative.
A study conducted by researchers from Cornell University and Microsoft Research involved 1,043 teachers and 23 curators. Its before-and-after survey analysis was based on 406 matched responses.
Teachers reported saving an average of 2.02 hours per week on lesson planning, providing an indication of how AI-assisted tools can reduce the time educators spend on preparation.
However, the findings relate to reported teacher workload and do not establish improvements in student learning outcomes. The researchers said they did not examine student learning outcomes.
The deployment covered students in grades 5 to 10 across Karnataka between December 2024 and March 2025.
Human Review Remains Central to AI-Assisted Lesson Planning
The Shiksha Copilot implementation also highlights the role of human oversight in deploying artificial intelligence in education.
AI-generated lesson plans were reviewed by curators before teachers adapted the material for classroom use.
Researchers also identified substantial language-related editing requirements in Kannada content, indicating that local-language AI applications continued to require human intervention.
The process illustrates the distinction between automated content generation and AI-assisted workflows that retain educators’ involvement.
Rather than relying entirely on AI-generated lesson plans, the initiative incorporated human review and adaptation before the material reached classrooms.
The example reinforced Puneet Chandok’s argument that AI adoption should be assessed through practical outcomes, including the time and effort it saves professionals, rather than simply counting how many people use the technology.
Microsoft Targets 20 Million Indians for AI Skilling by 2030
Beyond the education example, Chandok reiterated Microsoft’s commitment to equipping 20 million Indians with AI skills by 2030.
He emphasised that building computing infrastructure alone would not be sufficient to make India AI-ready, highlighting the importance of developing the skills required to use artificial intelligence effectively.







