Salesforce CEO Marc Benioff said his company will likely spend $300M on Anthropic tokens

• Salesforce CEO Marc Benioff had high praise for AI coding agencies and Anthropic.

• It allows reps to be more efficient and faster at product iteration, the CEO said.

• He said Salesforce will also introduce technology that makes coding easier in Slack.

Marc Benioff He’s on board with the AI ​​coding agency craze and is ready to spend big on it for his company.

In an episode of the All-In podcast released Friday, the Salesforce CEO said his company is estimated to spend $300 million on tokens from Anthropic.

Tokens are pieces of information, such as parts of a word, that a user types and a model processes. AI companies track token usage to bill users and large organizations.

“These coding agents are great. Anthropic is great,” he said. “I’ll probably use the $300 million Anthropic (token) in Salesforce this year. Coding. Everything will be cheaper to build.”

Benioff said on the podcast: artificial intelligence agents He brought “unprecedented” levels of efficiency to his company, pointing to areas such as service, support, distribution and marketing.

Last August, the CEO announced that AI agents enabled Salesforce to reduce its support staff from 9,000 to 5,000.

Along with the productivity gains he’s seen in coding reps, Benioff explained in the podcast that Salesforce is also working on ways to make coding easier. Slack, productivity platform Salesforce has owned it since 2021.

“We are even working on the technology inside Slack to make it easier for everyone to code,” he said. “You’re going to see great things with Slack and code, I’m not ready to talk about it yet. But there’s no doubt we’re in a new era of coding.”

While Benioff plans to spend $300 million on Anthropic this year, he said he doesn’t believe every token used by a company employee should go to a cap model like Anthropic’s Claude.

The CEO called for a “middleware” that could determine which future token inputs should go to Anthropic and which might be suitable for smaller models.

Read the original article Business Content