Salesforce’s woes underscore the problems of sales AI

Salesforce’s CEO, Marc Benioff, said the company was “complete with Agentforce” when it launched in 2024, but so far, only 34% of customers have adopted it. As a result, the company has lost more than $200 billion in market value, and analysts say Agentforce isn’t ready for prime time.

So, are companies not interested in agent AI, or are they just not ready? And what does that mean for marketers?

When Salesforce launched Agentforce, it set the stage as a way for businesses to build and deploy autonomous AI agents to handle customer service, sales, and marketing tasks.

Benioff described agents as the next big revolution in business software, saying they will change the way companies interact with customers and perform routine tasks. Initial customer feedback, however, has been muted, with many users reporting that they spent as much time preparing and organizing data as they did using AI.

The debate heated up this month after KeyBanc Capital Markets downgraded Salesforce, citing slow Agentforce adoption and a warning that only about 23,000 of the company’s 150,000 customers use the platform. Bernstein issued his layoff the same day, a rare move for a company the size of Salesforce.

Customers are not ready for AI self-management

KeyBanc research points to two reasons why Agentforce’s acquisition has been slower than Salesforce’s expectations.

First is data readiness. AI agents rely on clean, structured, connected data to make decisions and complete tasks, but many businesses still struggle with disparate CRM records, disconnected systems, and inconsistent customer information.

The second is product maturity. Based on interviews with Salesforce partners and customers, analysts concluded that Agentforce remains in the early stages of adoption, with most deployments still limited to proof-of-concept projects rather than enterprise-wide rollouts. Their CIO survey also found that more organizations expect to reduce Salesforce usage over the next year than increase it.

“Partners we’re talking to are now starting to turn Agentforce’s proof of concepts into deals in development, and most CIOs in our survey expect to outsource Salesforce from their IT budgets over the next 12 months,” wrote KeyBanc analysts led by Jackson Ader in their report.

That suggests the challenge isn’t to convince companies to power the agent AI. It gives them the data and operational base needed to use it effectively.

Wall Street questions Salesforce’s AI strategy

Analysts’ concerns had financial implications. Shares of Salesforce have fallen more than 50 percent from their December 2024 peak, wiping more than $200 billion off their market value as investors question whether Agentforce could be the company’s next big growth engine.

KeyBanc sums up its concerns bluntly: “The customer data doesn’t exist to do meaningful AI work,” and “Agentforce, as a product, just doesn’t exist.”

Salesforce rejects that assessment. Benioff publicly dismissed the KeyBanc report as a “bad call” and pointed to internal metrics that show Agentforce is the fastest-growing product in the company’s history.

“People think we’re headed for the wall when, in reality, the opportunity has never been greater,” he told the Wall Street Journal.

Not every analyst shares KeyBanc’s view. Andreessen Horowitz recently reported that companies investing heavily in AI increased their median Salesforce usage by 3% in the past three months. Guggenheim upgraded the stock to Buy, while Monness, Crespi, Hardt also raised its rating, saying Salesforce shares are performing well despite current concerns.

Salesforce is also investing to address issues that slow adoption. The company added technology that automatically pulls customer data from external sources and expanded its data management capabilities through acquisitions, including Informatica, to improve data integration and management before customers deploy AI agents.

The takeaway for sellers

The debate over Agentforce is less about Salesforce than about business AI.

For marketers, that’s a key shift. Organizations hoping to automate campaigning, lead qualification, customer service, and personalization are likely to reap greater benefits from improving data quality, integration, and management than deploying multiple AI agents before their CRM data is ready.

Agentforce’s adoption rate is a measure of enterprise AI readiness. The companies that move the fastest won’t be the ones buying new AI software. They will be the ones that already form the data base that those systems need to deliver meaningful results.