Chander's claim is that the transformation business has stopped promising replacement and started selling coexistence. Mainframes and thirty- to forty-year-old core systems are not going anywhere; what changes is consumption, so the work is pulling data out, marrying it, and exposing it as APIs and mobile experiences on top. His sharpest example is American healthcare, where a price-transparency mandate forced payers to publish rates — the same MRI can cost $1,000 with one provider and $100 with another — and the data sat in mainframes that could not expose it, state by state, in different formats. On generative AI he is deliberately cool: the post-ChatGPT frenzy is settling into the trough, and the value actually showing up is unglamorous — shift-left engineering, code scaffolding, test automation, code conversion, and making sense of unstructured data. A telco network copilot that spots a failure, teleprompts the technician up the tower and auto-remediates the fault is people-plus-AI, not AI instead of people. The contrarian note is that regulated industries adopt faster than anyone credits them for, provided the compliance work is done first — the perception that they are slow, he says, is exactly backwards. And the through-line for anyone working inside this business is relevance with a short half-life: a CTO who still writes code says he would be irrelevant in six months if he stopped reading.
Worth your time if you are
CIOs choosing between replatforming and wrapping the mainframe
Services leaders selling AI into regulated industries
Engineers whose skills have a six-month half-life
GCC and delivery heads hiring in tier-2 India
