Two types of organizations tend to end up in the same place when AI goes wrong.

The first moves fast. It deploys AI tools across hiring, forecasting, content production, and customer management, measures success by implementation speed and adoption rate, and treats any caution as resistance to change. The second moves slowly, unsure whether the technology is reliable enough for the decisions it’s being asked to support. It watches from the sidelines, falls behind on deployment, and loses ground to faster-moving competitors.

Both get into trouble. Different timelines, same core problem: neither starts with the question of whether the managers overseeing these systems understand enough about them to govern them responsibly.

Technology executive Phaneesh Murthy, whose advisory work spans technology companies in the U.S. and India, has diagnosed both failure modes with the same phrase: “Blind faith in AI is as dangerous as blind resistance to it.” His argument focuses on the quality of judgment that AI adoption requires, with pace as a secondary concern.

The Phaneesh Murthy Case for AI Fluency Over AI Speed

Murthy’s framing of the problem is specific. Organizations build technical teams to construct AI systems. They build compliance functions to manage regulatory risk. They don’t build management literacy to govern how AI outputs are used in real decisions.

That gap — between the teams constructing AI systems and the management layers deploying their outputs against consequential decisions — is where AI-related failures tend to originate.

When an AI screening tool produces systematically biased hiring outcomes, the engineering team that built the model shares responsibility. So does the management team that deployed the tool without reviewing what the model was trained on, what patterns it learned from historical data, and what the demographic distribution of its outputs looked like. These are governance questions. They require judgment. They belong at the management layer, and most management layers aren’t equipped to ask them.

“Technology scales intent,” Murthy has said. “If your intent lacks responsibility, the scale will magnify that flaw.”

What Phaneesh Murthy’s AI Fluency Model Requires of Enterprise Leaders

Fluency, in Phaneesh Murthy’s framework, rests on three things.

The first is capability recognition. AI is genuinely good at certain tasks: processing large datasets to find patterns, automating high-volume repetitive work, surfacing statistical correlations that human cognition would miss. Probabilistic forecasting from structured inputs is another genuine strength. Managers who know this can make AI deployment decisions that reflect actual organizational value.

The second is limitation recognition. Generative models produce false information confidently. Training data carries past biases into present decisions. Output quality follows input quality, and most organizations aren’t rigorous about their data pipelines. Managers who don’t know this end up overtrusting results they have no basis for evaluating.

The third is strategic contextualization. AI generates options. It doesn’t determine which options serve the organization’s actual goals, values, and constraints. A manager who treats the option with the highest AI-generated probability score as a finished decision has ceded something that’s not delegable — the judgment about where this organization should go, why, and what trade-offs that direction entails relative to actual organizational values and constraints.

“The goal is not to be first with technology,” Murthy has said. “The goal is to be wise with it.”

How Phaneesh Murthy Rethinks Human Work in AI-Assisted Organizations

AI’s practical effect on task allocation is real and ongoing. Reporting, scheduling, preliminary analysis, content drafting, customer segmentation: all of these can be AI-supported at a level of quality and volume that would have required substantial staffing commitments a few years ago.

The management question this creates centers on what human effort should be redirected toward once AI absorbs those tasks.

Murthy is direct about where he thinks that effort belongs. “The manager’s role is not to compete with AI. It is to elevate people beyond what AI can do.” That means moving human capacity toward synthesis, ethical evaluation, and relationship-dependent work. His advisory positions, including his role at InfoBeans and Opus Technologies, reflect this organizational philosophy in practice.

Why AI-Literate Leadership Becomes the Differentiating Credential

The credibility gap between AI-literate and AI-unfluent managers is widening. Teams working with AI-literate managers engage more substantively with AI-generated analysis — they push back on outputs that don’t account for known data quality issues, raise questions about model training populations before deployment decisions are finalized, and bring a level of critical engagement to AI-supported forecasts that passive tool users never develop. They catch more problems before those problems compound.

Boards with AI-fluent executives build stronger AI governance structures. They make better deployment decisions. They have more productive conversations with technical leadership about what AI can and can’t be expected to deliver.

Murthy has framed the organizational requirement plainly: “Leadership today requires technological awareness. Ignorance is no longer neutral.”

The managers building fluency now are filling a present gap, and the organizations they lead are better positioned for it.