Salesforce bets on "Agent Albert" to prove AI won't kill enterprise software Key Points - Salesforce CEO Marc Benioff pushes back on fears that AI agents will make traditional enterprise software obsolete, arguing the technology is actually a massive growth opportunity. - Customer reports show that Salesforce's current AI tool Agentforce handles routine tasks well but struggles with complex requests. A new product called "Agent Albert" is expected to launch by the end of the year. - Salesforce also introduced a new metric called "Agentic Work Unit" to quantify AI's impact. Wall Street worries AI could make traditional enterprise software obsolete.
Salesforce CEO Marc Benioff is pushing back with a new AI product and a homegrown metric. Wall Street is betting that AI agents could make traditional enterprise software obsolete, and Salesforce is squarely in the crosshairs. The company's business model relies on per-seat pricing, which looks shaky if customers start cutting headcount as AI agents take over more work. On top of that, companies might just "vibe code" their own software instead of paying for it - at least that's the fear driving the sell-off. Salesforce stock is down 28 percent since the start of the year. CEO Marc Benioff told the Wall Street Journal that the so-called "SaaSpocalypse" theory gets it backwards.
In his view, AI makes Salesforce more valuable, not less, and homegrown solutions are too risky when it comes to data security and compliance. He argues the opportunity has never been bigger, even as critics see the company cornered. Agentforce's mixed results raise questions about Salesforce's AI strategy Benioff's answer is a new AI product codenamed "Agent Albert," slated for release by the end of the year. The platform is designed to analyze users automatically and take actions on its own. But the track record of its predecessor, Agentforce, is a warning sign. Since launching in late 2024, only 23,000 of Salesforce's 150,000 customers have adopted it, according to the report.
It works well for routine tasks - education company Pearson saw a 40 percent jump in customer inquiries resolved automatically. On more complex jobs, though, the system falls short. Jewelry maker Pandora told the WSJ that Agentforce can't deliver reliable recommendations when customer requests are vague. To make AI's contribution easier to quantify, Salesforce rolled out a new metric called "Agentic Work Unit" (AWU), designed to capture how AI capabilities translate into concrete outcomes like resolved inquiries. In its most recent quarterly report, the company logged 2.4 billion AWUs, up 57 percent.
It was, however, the first time Salesforce had ever published the number, making outside comparisons impossible. AI News Without the Hype – Curated by Humans Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section. Subscribe now.