Salesforce survey: AI insights stay trapped with individual workers

Most AI use at work still happens one person at a time, and that is slowing companies down, according to a Salesforce survey of 250 U.S. executives. Of those surveyed, 68% say insights from someone's AI use go unshared, and 72% say they would prefer AI that works across the team with shared context.
- Salesforce surveyed 250 U.S. executives: 68% see AI insights go unshared and 59% see work stall in one person's AI tool.
- Executives estimate employees lose an average of 3.6 hours a week toggling between AI and other apps, about 22 business days a year.
- Weekly stalls affect 73% of enterprise executives versus 37% at small and medium-sized businesses.
- Only 15% have embedded AI agents across multiple systems and interfaces.
- 72% would choose AI that works across the team with shared context.
Insights stay with the individual
Salesforce describes the problem as "interface isolation": AI tools are in wide use, but what one person learns in a session rarely reaches colleagues. According to the survey, 68% of executives see AI insights go unshared, 66% say colleagues unknowingly use AI to solve the same problem, and 59% observe work stall because it is trapped in one person's AI tool.
The duplication is frequent. Sixty-six percent of executives say two colleagues use AI to solve the same problem separately, without realizing it, at least once a week. One IT director quoted in the research said engineers' time was wasted when they prompted AI to fix the exact same error because separate chat histories prevented them from sharing the solution. Another respondent said uncoordinated AI coding had caused duplicated work and major systems conflicts.
Reconciling separate AI-assisted outputs can also create additional work, the report says.
Larger companies are hit harder
Weekly stalls grow with organization size. Among enterprise executives, 73% report them, compared with 37% of executives at small and medium-sized businesses. Salesforce explains this by simple arithmetic: more employees means more private chat histories and more places for work to go unseen. In the study, SMBs are companies with fewer than 250 employees, and enterprises have 2,500 or more.
For technical teams, these blind spots can mean time spent troubleshooting avoidable problems. One accounting director said clear visibility into AI-generated code would have saved developers hours of debugging.
Few have connected their AI, and switching costs time
Most organizations are past experimentation, but few have tied their AI together. Forty-nine percent are deploying AI agents that complete tasks or workflows with some autonomy. Only 15% have embedded AI agents across multiple systems and interfaces so they work seamlessly across the organization.
The cost shows up in lost time. Executives estimate employees lose an average of 3.6 hours a week switching between AI tools and other systems, which Salesforce says is nearly 10% of every workweek. Over a 50-week year, that adds up to about 181 hours, or roughly 22 business days per employee. Only 11% say their AI tools sit in one integrated ecosystem and cost no time at all.
Executives want multiplayer AI
Asked to imagine an ideal setup, 72% of executives say they would choose AI that works across the team with shared context so value compounds and builds on itself. Twenty-six percent would have each person use AI on their own, with the value staying with that individual. One financial services director said a shared team workspace with collaborative AI context prevents redundant prompt engineering and accelerates cross-functional project delivery.
Rob Seaman, EVP and GM of Slack, framed the shift in Salesforce's terms: "The single-player era of AI made people faster. The multiplayer era is what makes a whole company smarter, together." He argued that an agent working in a shared space, rather than a private chat, picks up how a company actually operates through the conversations a team is already having.
What it means and how the survey was run
The findings point to coordination, not just adoption, as the next challenge for companies using AI, particularly larger ones. Salesforce, which owns Slack, positions shared context as the answer. Readers should note that the figures reflect executive perceptions and estimates, not measured time use.
The data come from a double-blind Salesforce Pulse Survey of 250 U.S. executives at director level and above who have decision-making authority over AI investments. Data was collected from September 14 through 22, 2026. Weekly hours lost were calculated using the low end of each response range and annualized across 50 workweeks.

