Business adoption of AI agents triples as Salesforce index shows measurable ROI
Salesforce's 2026 Agentic Enterprise Index shows the average number of AI agents per organization nearly tripled to 13, with creation time down 53% and measurable ROI emerging as agents handle more customer service autonomously.
The research, based on aggregate usage data from Salesforce's Agentforce platform across 400 businesses and a survey of roughly 5,000 respondents in nine markets, shows that agent capabilities improved by 350% and can now handle complex, multi-step tasks. The time needed to create a new agent dropped 53%, from an average of four days in early 2025 to 1.9 days today. The average number of unique actions per agent doubled from two to four, with retail peaking at nine.
Employee trust in AI agents has deepened sharply. Weekly initiated agent sessions tripled since February 2025, and employees now average nearly eight sessions per week with agents. Autonomous agents handle 7 in 10 customer service sessions, up from zero in early 2024. By August 2026, AI agents had handled more than five million conversations at help.salesforce.com, while human agents handled 2.4 million.
The index also found that agents are becoming more proactive and cross-functional, with the share of secondary functions rising from 1% to 6% in the past year. In April 2026, 734 million Agentic Work Units were consumed, representing a 15% month-over-month increase in the action-calls-to-output-token ratio.
Industry adoption varies. Consumer-facing sectors such as retail and travel lead in deployment speed and overall agent use, especially during peak seasons. Regulated industries have been slower to adopt, but their agents are often more sophisticated. The report notes measurable returns: sales increased four-fold when shopper agents were present.