According to the CEO of Catalyst, True AI integration has four possible non-negotiables.
“When you think about converting companies with AI, you actually have to do four things right,” Hemant Taneja said in an episode of the 20VC podcast released Monday.
Taneja said it was “very, very difficult” to achieve all four things and ensure that AI adoption exceeds the prototyping of OpenAI or human models.
“So these things are hitting the wall,” he added.
The VC said companies must first prepare their data infrastructure for AI adoption. Data infrastructure includes components such as servers, databases, cloud platforms, and networking equipment that work together to ensure data is used safely.
The second essential is a big language model that understands your business.
“You need to train these models in the secret sauce, in the context of your business,” he said.
Secondly, businesses need to think about changing the workforce as humans and AI work together.
“Some people are trying to manage AI agents. Some AI agents are trying to manage humans,” he said. “Imagine how your organizational chart needs to be changed.”
“The fourth is that for all of this to work, it actually takes courage on the top.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Catalyst's major investments include Airbnb, Windsurf, Mistral AI and more. The company did not respond to requests for comment from Business Insider.
Taneja is on a list of technology and business executives who have spoken about the importance of leaders being deeply involved in AI.
On a podcast in May, John Chambers, former CEO of VC and Cisco, said that most leaders have not reinvented themselves.
“As an AI leader, in my opinion, I have to reinvent myself every year,” he said.
This is because AI is running “5 times the speed” and “Internet terminology” delivering three times the results.
Nvidia CEO Jensen Huang talks regularly about how AI agents and humans will work together in the future, but says he uses AI as a tutor every day.
“In a fairly new field for me, I might say, 'I'm going to start by explaining like I'm 12 years old.' And over time, you should go up to PhD level and do your job,” he said at a meeting in May.
The ability of AI to quickly collect, analyze and communicate information could bridge technology gaps, Huang said.
