Salesforce has released its 2026 Agentic Enterprise Index, revealing that businesses are rapidly expanding the use of AI agents beyond simple conversational tasks to perform real business actions, automate workflows, and drive measurable returns on investment.
The report analyzes aggregated usage data from the Salesforce Agentforce platform between February 2025 and April 2026, providing insight into how organizations are deploying AI-powered digital workers across industries.
According to Salesforce, businesses are following two distinct AI deployment strategies. Consumer-facing industries such as retail are prioritizing high-volume, task-specific AI agents that handle routine customer interactions at scale. Meanwhile, industries with more complex operational and regulatory requirements, including manufacturing, healthcare, life sciences, and the public sector, are deploying more sophisticated AI agents capable of managing multi-step workflows and cross-functional business processes.
One of the report’s key findings is the rapid growth in AI agent adoption. The average number of agents activated per organization nearly tripled over the past year, while the time required to begin creating agents after provisioning has fallen to an average of just two days, a 53% improvement during the study period.
To measure business impact, Salesforce introduced the concept of an Agentic Work Unit (AWU), defined as a discrete task completed by an AI agent. As of April 2026, AWU output from Agentforce agents was growing at a compound monthly growth rate of 15%, indicating increasing levels of AI-driven work across organizations.
Retail and other customer-focused sectors generated some of the highest AWU volumes, reflecting the growing use of AI agents to manage customer inquiries, recommendations, and support requests. However, these agents typically remain focused on a limited number of tasks throughout most of the year.
The report also found that AI agents are becoming more capable. The average agent can now perform six different skills, compared with only two at the beginning of 2025. During peak retail seasons, the average retail AI agent was able to execute nine skills, highlighting the growing complexity of AI-assisted customer interactions.
Salesforce noted that regulated and operationally complex industries are leading in AI sophistication. Manufacturing, financial services, and healthcare organizations are deploying agents that can retrieve and analyze information, generate communications, update records, and execute actions across multiple business systems.
The financial services sector stood out for combining large-scale deployment with advanced functionality. Despite operating in highly regulated environments, financial institutions are deploying AI agents at volumes comparable to consumer industries, particularly during periods of heightened demand such as tax season.
The study also found a direct link between AI agent adoption and business performance. Organizations using AI agents reported stronger sales growth and increased operational efficiency, particularly in retail environments during peak shopping periods.
Employee trust in AI agents is also increasing. Internal usage of AI agents grew significantly during the study period, with employees interacting with agents 300% more frequently on a weekly basis. Salesforce cited Slackbot, its AI-powered assistant within Slack, as the fastest-adopted tool in the company’s history, with 83% of employees using it regularly and saving up to five hours of work each week.
Customer confidence in AI agents is rising as well. Over the last five quarters, AI agents handled 170 times more customer service conversations than in previous years and successfully resolved seven out of ten interactions without requiring human intervention.
Joe Inzerillo, Salesforce President of Enterprise AI and Technology, said organizations are increasingly moving beyond traditional chatbots toward AI agents capable of executing real work and delivering measurable business outcomes.
The report concludes that as enterprise AI matures, organizations will no longer need to choose between deploying AI at scale and managing complex workflows. Instead, future AI agents will be expected to deliver both high-volume productivity and sophisticated decision-making capabilities, creating a more capable digital workforce across industries.
