Salesforce (CRM) touted its new suite of autonomous artificial intelligence agents ahead of its annual Dreamforce conference Thursday, saying its Agentforce platform represented “the third wave of the AI revolution.”

Speaking to reporters, CEO Marc Benioff said the deployment of digital agents marked the start of a hybrid future in which humans and AI agents work side by side.

“This is the biggest and most exciting piece of technology we have ever worked on,” Benioff said. “We are just starting.”

Agentforce allows Salesforce’s enterprise customers to build and deploy AI-powered agents customized to their business needs. It utilizes all the data stored in the company’s Data Cloud platform, which combines information from multiple sources.

The autonomous assistants stand in for human customer service representatives, making decisions and acting on information. Unlike traditional chatbots that handle pre-programmed questions, AI agents can reason and tackle multiple tasks at the same time. The software runs on top of Salesforce’s existing business apps.

Salesforce claims that integration and deployment of the technology require little to no coding, reducing the time and cost of building the agents.

“It's a difference between asking an AI in level one to write a sales email versus asking the agent to help you hit your number, hit your quota for the year, and then the agent figures out, OK, what are the steps that I need to take to hit the quota?,” said Clara Shih, CEO of Salesforce AI. “The human supervisor looks at it, and then the agent actually performs those actions. It’s a completely next level of automation.”

Agentforce builds on the company’s existing suite of AI products, including Einstein Copilot, and marks a strategic shift for a company known for customer relationship management software.

Salesforce extols AI's efficacy, but concerns linger

Salesforce is looking to convince a skeptical market that its AI offerings can compete in an increasingly competitive enterprise AI space led by Microsoft (MSFT).

"I think they need to show us real innovation around AI, that they are actually investing aggressively to be AI-first," RBC Capital Markets analyst Rishi Jaluria said in an interview with Yahoo Finance prior to Thursday’s announcement.

Salesforce stock rose slightly on Thursday and in early trading Friday but remained off by 3% year to date, compared to an 18% gain for the Nasdaq Composite (^IXIC) index.

Benioff has maintained that Salesforce has an outsized advantage because of the sheer scale of customer data the company has, pointing to a trove of 250 petabytes in a quarterly earnings call earlier this year. He touted Agentforce as the biggest revenue driver moving forward and singled out the platform 39 times in Salesforce’s second quarter earnings call.

Shih said Salesforce’s utilization of specific customer data also means its AI technology is less prone to hallucinations.

"Having specific context about what the AI is trying to get done helps make that AI more accurate," Shih told Yahoo Finance. "And then you add in all of the information, all the data that we have about who's using it and the customer they're talking to — it gets to a level of personalization where you can actually start to deploy these autonomous agents."

A handful of companies, including ADP (ADP), OpenTable, and Wiley, have already deployed these agents. Healthcare provider Kaiser Permanente is using Agentforce to handle patient inquiries and reported resolution rates of more than 90%, according to Benioff. Disney (DIS) has deployed agents to augment tour guides at its parks.

"[Disney] said it was giving them twice the accuracy of the AI that they were already using," Benioff said.

The agents are equipped to tackle various roles out of the box.

For example, one of Salesforce's service development agents communicates with new buyers, while a sales coach agent helps human customer representatives sharpen their sales pitches and coaches them during actual calls. There are also agents that help develop marketing campaigns.

Data safety and ethics surrounding AI remain a major concern, as there are few regulations and standards. Shih said Agentforce is grounded in data that’s specific to individual companies and allows those firms to set their own guardrails, depending on the tasks agents are assigned.

Any new role or new function requires human approval, she said.

"We have these guardrails in place that have been battle-tested by years of production, used by your customer service employees, by your salespeople, by your marketers," she added.

Click here for the latest stock market news and in-depth analysis, including events that move stocks

Read the latest financial and business news from Yahoo Finance