Share with your CHRO
MakeMyTrip is betting that AI workforce readiness starts with problem identification, not tool adoption. Group CHRO Yuvaraj Srivastava structured the company’s capability push around an internal “Agent-a-thon” where employees surfaced nearly 200 use cases from their own workflows, then worked in six-week sprints pairing functional staff with engineers. The organizing principle is deliberate sizing: match the AI resource to the problem’s actual scope, cost, and consequence, rather than defaulting to the most powerful model available.
What this means for your business
Most enterprise AI training programs are, functionally, tool catalogs with a completion certificate attached. The recurring failure mode looks like this: L&D buys licenses, mandates modules, tracks completion rates, and reports adoption numbers that don’t connect to any measurable business outcome. Srivastava’s framing cuts through that pattern cleanly, and the organizations most exposed are the ones whose AI readiness metric is still “percentage of employees trained” rather than “number of workflows measurably improved.”
The Agent-a-thon structure is worth examining as a model, not just an anecdote. Asking functional teams to originate use cases rather than receive them from a central IT or transformation office does two things at once. It routes problem-identification to the people with actual workflow knowledge, and it makes outcome ownership impossible to dodge. An engineer paired with a finance analyst who defined the problem and set the success metric can’t ship a technically impressive solution that misses the point. That accountability structure is what most hackathon-style programs get wrong, and it’s the variable that determines whether the output becomes a pilot or a slide deck.
The deeper bet here is that “skills intelligence,” knowing when a simpler, cheaper tool is sufficient and when it isn’t, becomes a durable competitive advantage as AI tooling commoditizes. If every organization has access to the same frontier models, the edge goes to the ones whose employees can calibrate spend to problem size without a centralized approval gate slowing every decision. CHROs who are still measuring AI fluency by tool exposure should weigh whether their current capability framework would even surface that skill, because it almost certainly won’t appear on a learning management system completion report.
Concept deep-dive: Skills Intelligence
Skills intelligence, in this context, means knowing not just what a tool can do but whether it’s the right tool for a specific problem’s size, cost, and consequence. Think of it as clinical judgment applied to AI tooling: a doctor doesn’t prescribe the strongest available drug for a mild headache. For CHROs, this reframes capability building away from breadth of tool exposure and toward the diagnostic habit of matching solution complexity to problem complexity before committing resources.
Based on reporting from Not every problem needs the most powerful AI model: Yuvaraj Srivastava, MMT’s group CHRO, originally published 2026-07-31 04:20:00.
