(Bloomberg) -- Salesforce Inc. is unveiling a pivot in its artificial intelligence strategy this week at its annual Dreamforce conference, now saying that its AI tools can handle tasks without human supervision and changing the way it charges for software.
The company is famous for ushering in the era of software as a service, which involves renting access to computer applications via a subscription. But as generative AI shakes up the industry, Salesforce is rethinking its business model for the emerging technology. The software giant will charge $2 per conversation held by its new “agents” — generative AI built to handle tasks like customer service or scheduling sales meetings without the need for human supervision.
The new pricing strategy also seeks to protect Salesforce if AI contributes to future job losses and business customers have fewer workers to buy subscriptions to the company’s software.
Salesforce is even leaning into the employee-replacement potential of the new technology. Its new AI agents will let companies increase their workforce capacity during busy periods without having to hire additional full-time employees or “gig workers,” Chief Executive Officer Marc Benioff said Tuesday during a keynote speech at the company’s annual Dreamforce conference.
Despite a breathless focus on AI since early 2023, application software makers like Salesforce, Workday Inc. and ServiceNow Inc. have little to show for their efforts. Revenue and valuation gains from AI have largely flowed instead to makers of hardware such as Nvidia Corp. or cloud infrastructure like Oracle Corp.
Many software vendors have launched AI assistants that can summarize or draft written content — the most well-known of which is Microsoft Corp.’s Copilot. Generally, customers haven’t been quite ready to pay for these additional features.
“I think the results that people saw from the copilots did not meet their expectations,” said Chief Operating Officer Brian Millham in a late August interview. “I think the hype was a little ahead of the results on the copilot front.”
The new iteration of Salesforce’s AI products is meant to run without supervision “in contrast to now-outdated copilots and chatbots that rely on human requests and struggle with complex or multistep tasks,” the company said in a statement. For example, publisher John Wiley & Sons Inc. said it was able to boost the number of customer service claims it resolved using Salesforce “agents” without employee interaction.
This represents a change for Salesforce, which until recently focused on building human-assisting AI tools. “We have a principle called human in the loop — we don’t know enough yet to introduce technology that is fully autonomous,” Patrick Stokes, an executive at the company, said in September 2023.
At Dreamforce this week in San Francisco — where 45,000 in-person attendees are expected — the company will make the case for its new strategy. Its venture arm also announced a new $500 million fund oriented for AI startups.
Benioff on Tuesday argued the new AI products will be highly accurate and secure because Salesforce already holds a massive amount of customer data. That’s in contrast to what he called “those nasty copilots.”
The pivot by Salesforce addresses another investor fear that job losses from AI could hurt the software-as-a-service business model.
One of the main pitches of AI is labor efficiency. A company that uses AI tools for customer service will need fewer human representatives to serve the same base, for example. But slower-growing corporate workforces will dent revenue growth at software companies, which largely charge based on the number of workers authorized to use the products. Wall Street analysts have spent recent earnings calls sparring with software company management teams over this risk.
By pricing its new AI features based on outcomes rather than the number of employees using it, Salesforce is insulating itself from customer job cuts.
Millham, the COO, gave an example of a 5,000-person call center needing 30% fewer workers within five years. Others may just elect to hire fewer going forward, he added.
“We like to think, ‘Let’s re-skill and move people up and and do more complex work,’” Millham said. “But there are companies certainly who are looking through the lens of — ‘Hey, I think AI can do a lot of the work of the humans today.”
(Updates with comments from CEO beginning in the fourth paragraph.)
©2024 Bloomberg L.P.