The Disruption Narrative We Keep Getting Wrong
Whenever a significant AI capability emerges, the dominant conversation in leadership circles gravitates toward scale: industries transformed, jobs automated, competitive moats redrawn. Gayatri Agrawal, writing about her family, offers a quieter counter-argument — and it lands harder precisely because of its smallness.
Agrawal describes her grandfather, a 70-year-old she refers to as dadaji, as someone who, by her account, genuinely enjoys staying productive. He tracks accounts, maintains records, and has spent decades navigating successive waves of technological change — from pen and paper to laptops. The friction, she notes, was never motivation or capability. It was the keyboard.
Removing the Bottleneck, Not the Person
Rather than attempting to accelerate her grandfather's typing, Agrawal chose a different intervention: she eliminated typing as a requirement. She installed Wispr Flow, a voice-to-text tool, and configured it to activate with a single key. According to Agrawal, her grandfather can now dictate naturally — in whichever language feels comfortable — and the software handles transcription, error correction, and formatting across applications including WhatsApp Web, Tally, and document editors.
The follow-on step is equally instructive. Agrawal then installed Claude Cowork on the same machine and connected it to his accounts folder. Instead of manually searching through files, he can now simply speak queries and interact with his own data conversationally. The system retrieves and surfaces what he needs without requiring him to remember file structures or navigation paths.
This is not a story about a technology-resistant elder being reluctantly brought into the modern age. It is a story about the wrong interface being the only barrier between an engaged, capable person and full participation in digital work.
"Ye Toh Jaadu Hai"
Agrawal recounts the moment her grandfather looked at her and said, "Ye toh jaadu hai" — roughly, "This is magic." She describes that reaction as the point at which a larger realisation crystallised for her.
The observation she draws is precise: we talk extensively about people having to adapt to technology, but her grandfather's experience represented something different — technology adapting to him. He did not change his language, his cognitive style, or his way of explaining things. The tools restructured themselves around his existing behaviour.
We spend so much time talking about AI replacing jobs, changing industries, and transforming businesses. But sometimes the most interesting thing about technology is much simpler.
What This Reveals About AI Adoption Strategy
For leaders thinking about how to embed AI into organisations, Agrawal's framing carries a practical implication that is easy to miss. The standard adoption playbook focuses on training — upskilling employees, building digital literacy, running change management programmes. All of that assumes the human must move toward the tool.
The emerging alternative — already visible in voice interfaces, natural language queries, and contextual AI agents — is designing for the irreducible human: the person whose value lies in judgment, domain knowledge, and experience, not in their willingness or ability to learn yet another interface convention.
Agrawal does not claim this transition is universal or complete. But the grandfather scenario functions as a proof of concept: a 70-year-old with decades of accounting knowledge, previously slowed by a keyboard, is now more operationally capable than before — without being asked to become someone different.
The Mindset Shift Worth Watching
What Agrawal is articulating, whether she frames it this way or not, is a quiet inversion of the usual technology adoption contract. For most of computing history, humans have been required to speak the machine's language. The question her grandfather's experience raises is whether the most durable competitive advantage of current AI is that it finally speaks ours.