ConfX Global
InsightXOverviewReportsSector BriefingsLeadership PulseKnowledge SessionsInterviewsArticlesTools
Enterprise AI · Podcast

Cisco’s Jeetu Patel on why trust decides whether AI gets used

Jeetu Patel, President and Chief Product Officer, Cisco

Jeetu Patel is president and chief product officer at Cisco, responsible for building and taking to market its networking, security, data and collaboration products. He joined host Sam Ransbotham on Me, Myself, and AI, the podcast from MIT Sloan Management Review, in an episode released on 14 October 2025. They talked about Cisco’s part in the AI build-out, why he thinks AI is both underrated and overrated, what it means for hiring and mentoring, and how to tell a megatrend from hype.

Building the plumbing

Patel described Cisco as the critical infrastructure company for the AI era, supplying the plumbing that lets people connect, stay secure while connected and use a data platform. He accepted the picks-and-shovels comparison, saying infrastructure providers tend to do well in platform shifts because the whole underlying system has to be rebuilt. Cars, he noted, needed roads and traffic lights before they could fit into society.

Data centres, he argued, show the scale of the rebuild. Today’s facilities lack the power, compute and network bandwidth that AI needs, so they have to be rethought and re-racked, with compute, networking and switching equipment cooled in new ways.

Why the scale is hard to grasp

He ranked AI above the internet and the automobile in importance. Each wave builds on the ones before, he reasoned, so later waves tend to have larger effects, and he called AI probably the most consequential set of inventions in our lifetime. Innovation is compressing timescales so sharply, in his telling, that scientific progress could compound around a thousandfold, which makes the situation three years out impossible to predict.

Patel traced the take-off to ChatGPT’s launch on 30 November 2022. Large language models changed things because they understand human language, so people no longer have to learn the machine’s. That depended on petabytes of public internet data for training, another case of each revolution resting on the infrastructure of the last.

Recalling a conversation with Ray Kurzweil about two decades earlier, he said the lesson was that change compounds across many dimensions at once, while people tend to think in straight lines.

Trust, productivity and the bigger prize

On security, Patel said its role had reversed. It used to be seen as a drag on productivity. In the AI era, people will not use systems they do not trust, so security becomes a condition of use.

“security is going to be a prerequisite for successful adoption of AI”

People underrate AI, he argued, when they focus on productivity, which is only the first-order effect. Second-order effects include new kinds of models beyond language models, such as world models and physical models, and AI producing original insight that is not already in the human record. That would let people solve problems they could not previously imagine. He put productivity at about 10% of the value and new problem-solving at about 90%.

Where he sees AI overhyped is in the assumption that humans will become obsolete. Human instinct and judgment, gut calls included, are hard to replicate, and people are not built to sit on a beach while machines do everything. Jobs will change and people will move to higher-order work. He acknowledged that displacement will happen for a period and said that suffering should not be taken lightly. If society gets ahead of it, though, retraining could be more efficient than in past disruptions, with the public and private sectors working together.

Hiring, unlearning and mentoring in both directions

Patel strongly opposed cutting entry-level hiring. Companies that stop bringing in new people, he said, give up a supply of fresh ideas and the push those people provide to unlearn old habits. Organisations need to unlearn as they learn, which means bringing in new talent quickly and giving it creative freedom.

Cisco’s interns had taught him that age shapes how people use AI. A 28-year-old tends to use it for quick answers, while a 20-year-old treats it as a companion and brainstorming partner. His response is reverse mentoring. Newcomers should mentor leaders as much as leaders mentor them, and being inexperienced can simply mean being new to a field. He suggested leaders spend part of each intern session learning from the interns.

His own habit is to move into unfamiliar fields, which keeps him curious and free to ask naive questions. He values strong opinions, loosely held.

Telling a megatrend from hype

His rule is to work with megatrends and use them as tailwinds. The test is simple. If a benefit needs a PhD to explain, it is probably hype. If you can see five steps of benefit straight away, it is probably a megatrend.

In a closing round of quick answers, he said the technology is moving fast while adoption and real value inside organisations are slower. His best personal use of AI is research, and his frustration is being stuck in chat interfaces when he wants agents to take over tasks he dislikes.

What it means for leaders

  • In Patel’s framing, security teams decide whether AI gets used at all, so they belong in adoption plans from the first draft.
  • Leaders measuring AI only in hours saved may be tracking the smallest share of its value. His 10/90 split is a reason to set some budget aside for problems the business has written off as unsolvable.
  • Freezing graduate hiring to pay for AI tools could remove the very people most likely to show senior staff how to use them.
  • A standing reverse-mentoring slot, even one hour per intern cohort, is a cheap way to act on his advice.
  • His PhD test makes a handy filter for vendor pitches.
This page summarises "Never Fight a Megatrend: Cisco’s Jeetu Patel", published by MIT Sloan Management Review (Me, Myself, and AI) on 14 October 2025. The original and all rights in it belong to MIT Sloan Management Review (Me, Myself, and AI).Listen to the original ↗