True AI integration has four non-negotiables, according to General Catalyst’s CEO.

“When you think about transforming an enterprise with AI, you actually have to do four things correctly,” Hemant Taneja said on an episode of the 20VC podcast released on Monday.

Taneja said that it is “very, very difficult” to achieve all four things and ensure that AI adoption goes beyond prototyping an OpenAI or Anthropic model.

“That’s why these things are hitting a wall,” he added.

The VC said that companies need to prepare their data infrastructure for AI adoption first. Data infrastructure includes components like servers, databases, cloud platforms, and networking equipment that work together to make data secure and usable.

The second must-have is large language models that understand your business.

“You have to train these models in the context of your secret sauce, your business,” he said.

Next, companies need to think about workforce transformation, since humans and AI will work together.

“Some humans are going to manage AI agents. Some AI agents are going to manage humans,” he said. “Imagine how the org charts have to change.”

“The fourth, for all this to work, you actually need courage at the top. The CEOs need to really get behind it to drive it.”

General Catalyst’s major investments include Airbnb, Windsurf, and Mistral AI, among others. The firm did not respond to a request for comment from Business Insider.

Taneja joins a list of tech and business executives who have spoken about the importance of leaders getting deeply involved in AI.

In a May podcast, John Chambers, a VC and Cisco’s former CEO, said that most leaders do not reinvent themselves.

“As a leader in AI, you have to reinvent yourself, in my opinion, every year,” he said.

This is because AI is moving in “internet terms” at “five times the speed” and delivering “three times the results,” Chambers said.

Nvidia’s CEO Jensen Huang, who regularly speaks about how AI agents and humans will work together in the future, said he uses AI as a tutor every day.

“In areas that are fairly new to me, I might say, ‘Start by explaining it to me like I’m a 12-year-old,’ and then work your way up into a doctorate-level over time,” he said at a conference in May.

AI’s ability to rapidly collect, analyze, and communicate information could close the tech gap, Huang said.