Microsoft India president Puneet Chandok describes a human-AI 'barbell' and prescribes concrete actions for graduates, managers, and enterprises: learn AI skills, keep human judgement, focus on ROI, and redesign workflows around always-on automation.
Treat AI skill as table stakes while sustaining human judgement: technical ability plus empathy and sense-making form the human‑AI barbell.
Enterprises must link AI projects to clear ROI, redesign processes for token-based automation, and build 'frontier professionals' who combine automation with judgement.
The barbell model separates two non-negotiable capabilities. On the technical end, people must master agent orchestration and other AI skills. On the human end, empathy and sense-making grow in value as automation advances. Chandok argues that tools like copilots can sharpen analytical reasoning—because they invite unfiltered interrogation—yet organisations must still create human-to-human interactions so emotional and judgment skills do not atrophy.
Advice for graduates and early-career engineers
Chandok says today's graduates have an advantage as AI natives, but they must treat AI skills as table stakes. Practical guidance:
Enterprise operating changes to accelerate AI adoption
Chandok identifies three shifts shaping corporate operating models:
1) ROI focus: Organisations and boards demand a clear path to ROI. AI pilots without ROI get deprioritised. Projects should be scoped and measured against returns.
2) Workflow redesign for tokens: Don't just layer models on broken processes. Rethink workflows around token-based automation that operate 24/7 without human constraints. Example: Air India rebuilt customer service to handle 13 million queries with 97% accuracy.
3) Building frontier professionals: Firms need teams who truly use and interrogate automation rather than merely discussing it. Microsoft's study found 32% of AI users in India qualify as 'frontier professionals'—double the global average.
Chandok stresses that upskilling can't stop at entry level. Managers must be capable of setting ROI-aligned priorities, redesigning processes, and cultivating frontier behaviour. Practical actions for leaders include measuring outcomes, validating that AI reduces costs or improves quality, and insisting teams ship iteratively with clear judgement checkpoints.
Chandok describes using a copilot to plan a morning session: the tool analysed a report, the user debated with it, and pushed back where it missed angles. He calls the copilot a 'guardian angel' that amplifies performance while still requiring the user's judgement. The point: use AI as an interlocutor to sharpen reasoning, not as a crutch.
Chandok's blueprint is operational rather than ideological: master AI skills, preserve and exercise human judgement, prioritise projects with measurable ROI, redesign processes for continuous automation, and build teams who actually use and test these systems. For graduates, that means shipping work and demonstrating iterative judgment. For enterprises, it means measurable deployments and new operating models that pair token capital with human oversight.
Open the app view to save this story, compare related coverage, and continue from the same source.