Share with your CHRO
Infosys CHRO Shaji Mathew and Reliance Industries CHRO Ashwani Prashara made a pointed case at the ETHRWorld Annual Conclave 2026 that AI in HR demands process demolition before automation. Mathew cited Infosys’s 100-plus HR applications across a 300,000-person workforce as evidence that layering AI onto legacy complexity compounds problems rather than solving them. His concrete example, collapsing a multi-day offer-letter process to minutes by stripping redundant validation steps, is the model. Prashara pushed further, arguing the conversation has moved past automation into organizational design.
What this means for your business
The offer-letter example is doing more work than it appears. Compressing days into minutes wasn’t an AI story, it was a process audit that AI then accelerated. CHROs sitting on sprawling HR tech stacks, the average large enterprise runs dozens of point solutions, face a specific trap here. If your AI roadmap is a list of automation use cases mapped onto existing workflows, you’re likely encoding the inefficiency rather than removing it. The question worth asking before any vendor conversation is which processes would you eliminate entirely if you were redesigning from scratch today.
Prashara’s reframe from “automation” to “organization design” is the sharper claim in this piece. When he says the challenge is decomposing work into components and deciding which belong to intelligent systems versus humans, he’s describing something most HR functions aren’t structurally equipped to do. That kind of work-decomposition analysis sits at the intersection of industrial engineering, behavioral science, and technology architecture, none of which have traditionally lived in HR. The convergence argument, that business architecture, technology architecture, and workforce architecture can no longer be planned independently, places CHROs in conversations they’ve historically been excluded from, which is both the opportunity and the credibility risk.
The “high touch” warning from Prashara is where I’d push back slightly on the framing. Calling the future role “Chief Humility and Respect Officer” is a conference applause line, but the underlying concern is real and measurable. Organizations that digitize employee experience without investment in manager capability consistently see engagement scores drop even as transaction times fall. The leading indicator to watch isn’t how much HR automation your org has deployed, it’s whether your manager-to-employee conversation frequency has held steady or declined as self-service adoption rises. If it’s declined, the automation dividend is being consumed by disengagement, not returned as business performance.
Concept deep-dive: Process decomposition
Process decomposition means breaking a workflow into its smallest distinct steps to evaluate each one on its own merits, rather than treating the whole process as the unit of improvement. Think of it as the difference between renovating a house room by room versus asking which rooms shouldn’t exist at all. In an AI context, it matters because intelligent systems perform well on bounded, well-defined tasks and poorly on ambiguous, judgment-heavy ones. Mapping which steps are which is the precondition for any automation decision that actually changes outcomes.
Based on reporting from Revolutionizing HR: The Impact of AI on Human Resources, ETHRWorld, originally published 2026-09-14 05:51:00.
