Autonomize CEO says healthcare shows why enterprise AI projects fail
Ganesh Padmanabhan argues companies need governed AI systems, not more disconnected agents, citing costly health-plan approval delays.
By Daniel Okafor · Business Editor
3 min read
Autonomize AI founder and CEO Ganesh Padmanabhan says enterprise AI is often being applied to the wrong part of the problem, with health care showing the cost of that mistake. In a Fortune commentary, he argued that companies are automating fragments of old workflows rather than redesigning how decisions get made and checked.
Padmanabhan, who spent more than a decade at Dell Technologies before later working at CognitiveScale and founding Autonomize AI in 2022, said the problem is clearest in health-plan operations. According to his account, a new drug approval can trigger a coverage review involving six or seven specialists and take 60 to 90 days at a cost of about $100,000 per drug.
For patients, he wrote, the delay can mean waiting in coverage limbo. Padmanabhan cited schizophrenia drugs as one case where interrupted treatment can lead to hospitalizations, which he said can cost health plans $8,000 to $15,000 per admission and add up to $4 million to $7 million across several hundred admissions at a large plan.
Padmanabhan said Autonomize AI is focused on reducing administrative work in health care by coordinating AI systems around complex operational tasks. Fortune’s author biography said the company works across three of the five largest U.S. health enterprises and was named a World Economic Forum Technology Pioneer in 2026.
AI pilots face scrutiny
Padmanabhan pointed to broader signs that corporate AI projects are struggling to show value. He cited MIT research, reported by Fortune, that examined more than 300 enterprise AI deployments and found 95% of generative AI pilots produced no measurable return.
He said the issue was not only model quality. In his view, organizations often attach AI to existing processes without changing the structure of the work, limiting the payoff.
He also cited Gartner’s forecast that more than 40% of agentic AI projects will be canceled by the end of 2027. Gartner attributed that expected pullback to factors including inadequate risk controls, cost and unclear value, according to Padmanabhan’s commentary.
Health care as an early test
Padmanabhan said health care exposes AI governance problems faster because decisions are regulated, data is fragmented and clinical expertise is scarce. He cited the Association of American Medical Colleges’ projection of a physician shortage of up to 86,000 by 2036.
He also cited the American Medical Association’s 2024 physician survey, which found that prior authorization takes an average of 13 hours of physician and staff time each week. The AMA survey also found that 93% of physicians said prior authorization delays patient care.
At one health enterprise that works with Autonomize AI, Padmanabhan said 600 nurses focus mainly on prior authorization and payment integrity. He argued that such work keeps clinicians tied to administrative processes rather than patient care.
Padmanabhan said Autonomize’s coordinated AI approach can reduce a drug-coverage assessment from months to four to eight hours, with a clinical pharmacist reviewing the output instead of producing it from scratch. He said that approach cuts direct labor costs by 97% and leaves a record of the AI agents’ work for later compliance review.
His broader argument is that companies need systems to manage, audit and govern AI agents. Without that, he wrote, enterprises risk repeating earlier technology cycles in which disconnected tools sped up work without improving the decisions behind it.
This story draws on original reporting from Fortune.