For most of my career I have watched good ideas move at the speed of caution. In medicine, a new approach takes roughly 17 years to travel from research to routine practice, and only about one in seven ever completes that trip. This is not a health care habit alone. Across industries, leaders have been trained to expect that a technology will announce itself, mature slowly, and grant them a comfortable window to respond. The Gartner Hype Cycle, the planning tool of a generation of boardrooms, is built on exactly that assumption: five to ten years from novelty to mainstream.
Artificial Intelligence (AI) has quietly torn up that schedule.
AI is the fifth great disruption in human history, after the wheel, fire, electricity and the internet, and the first to reshape every sector at the same time. The pace is not incremental. It compounds. Research output in AI is now produced at more than a hundred times the rate of three decades ago. The ability of these systems to handle longer, more complex tasks has been doubling roughly every seven months. Capabilities that the Hype Cycle said would take half a decade to arrive have shown up in under two years. For the first time, I believe a planning framework built on historical adoption curves may simply mislead the leaders who rely on it.
So, I have proposed a different map, one calibrated to what AI can do rather than to how past technologies spread. I call it the Functional AI Pyramid, a maturity model that works as well for a bank or a manufacturer as for a hospital.
{{/usCountry}}So, I have proposed a different map, one calibrated to what AI can do rather than to how past technologies spread. I call it the Functional AI Pyramid, a maturity model that works as well for a bank or a manufacturer as for a hospital.
{{/usCountry}}Two levels sit before the AI age even begins. At the base are the Naives, the organisations and professionals still unaware the ground is moving. Just above them are the Watchers, who have noticed AI but remain unconvinced of its urgency, certain that full integration is a decade away. Both are gambling. The Naive risks irrelevance before realising the game has changed. The Watcher mistakes awareness for action. When adoption compresses from years to months, watching becomes its own form of risk.
Then come the five levels that define the AI age, and they climb the value chain of any enterprise, from clerical work to executive vision.
Talkers are the conversational AI we already use: Assistants, chatbots, drafting and translation tools. They take over deterministic, desk-based communication. In a hospital they hand a clinician back time with the patient. In any business they end the era of the knowledge worker as a data-entry clerk.
Thinkers move from conversation to cognition. These reasoning systems weigh options, find patterns and support decisions. They close the era of trial-and-error judgement. In medicine that unsettles pattern-heavy specialties like radiology and pathology. In finance, insurance and law it does the same to analysts, underwriters and junior researchers.
Doers act. Autonomous systems execute within defined frameworks, coordinating care, managing supply chains, processing claims and running operations with little oversight. This is the end of operational inefficiency, and of much mid-level coordination work.
Innovators create. Here AI stops executing and starts inventing: new drugs, new materials, new products and pathways. This level is personal for me. I have lived with diabetes since my early years, and not once was I offered a genetic test to find which medication actually suited my genotype. My own such test generated more than six gigabytes of data, information that only AI, aided by quantum computing, can turn into action. We have spent a century treating the phenotype. AI lets us treat the genotype too. As I have said, we will soon have more than eight billion ways to treat eight billion people. For every industry, this is the end of the generic.
Leaders, the summit, is AI as institutional intelligence, coordinating people, strategy, operations and governance as one system. In health care this is where I see the Housepital, a personalised, AI-driven home health environment that could meet perhaps four-fifths of an individual's needs without a hospital visit, dissolving the outpatient department and shrinking hospitals toward what only they can do. For organisations more broadly, it is the end of command-and-control hierarchies and of the bureaucratic friction, the billing, audit cycles and manual governance, that consumes a large share of operational effort.
This is why I describe what lies ahead in medicine as the Golden Age of Medicine, the shift from treating sickness to sustaining wellness, from health care to health. But the deeper lesson is one every Indian boardroom should absorb now, because the workforce maths is already visible.
By 2028, I expect a mature organisation to run on a 50/50 model: Half its workforce digital agents handling logic, data and routine workflows, half human, freed for the work that cannot be automated, namely judgement, ethics, creativity and trust. In my own organisation, digital team members already outnumber human ones. By 2030, I expect that to be the norm rather than the exception.
AI will first augment, and then replace what it augments. That is not a reason for fear. It is a reason to plan. The professional or company that treats this decade as an optional upgrade will be overtaken by the one that treats it as a deadline. Remember the firms that did not: Kodak, Nokia, BlackBerry, and how fast their leadership of a category disappeared once the curve bent.
There is one condition on all of this, and India cannot afford to get it wrong. The pyramid is a map of capability. It says nothing about fairness. The ethical failures of AI almost always trace back not to the algorithm but to the data it learned from, data that under-represents the very populations we most need to serve. This is why I argue for a shift from Ethical AI to Ethical Data, captured in a deliberately blunt campaign, Data First. AI Later. Get governance wrong, and we will simply automate our existing inequities at speed.
The prize, if we get it right, is large: Closing the gap on access, accuracy, affordability and accountability at once, in health and in every sector that touches a citizen's life. For a nation building toward a Viksit future, that is not a side benefit of growth. It is its foundation.
(The views expressed are personal)
This article is authored by Rajendra Pratap Gupta, chairman, Academy of Digital Health Sciences.