New Salesforce research indicates that AI at work remains predominantly a solo activity. A staggering 68% of executives report unshared AI insights weekly, 66% observe colleagues unknowingly duplicating AI-driven solutions, and 59% see work stalled due to AI tools trapping information. Employees spend an average of 3.6 hours per week switching between AI and other applications, amounting to 22 lost business days annually. In an ideal scenario, 72% of executives would design AI for collaborative team use with shared context.
Introduction: Interface Isolation and the Need for Multiplayer AI As businesses increasingly rely on AI, a Salesforce survey of 250 U.S. executives reveals a bottleneck of interface isolation. 68% see AI insights go unshared, 66% report duplicated efforts, and 59% experience work stalls due to isolated AI tools. Employees waste 22 business days annually toggling between AI and other apps, a problem that intensifies in larger organizations. 72% of executives envision AI that fosters team collaboration with shared context, moving from a 'single-player' to a 'multiplayer' era of AI. What one person learns with AI rarely reaches the team Most AI use remains individual, leading to siloed insights. 68% of executives observe AI outputs not being shared or reused, causing stalled company output. Discoveries, like efficient prompts that can halve review time, often stay confined to one person's workflow, hindering broader team benefit and collective intelligence. Teams are solving the same problems twice The lack of shared AI knowledge results in duplicated efforts across organizations. 66% of executives report colleagues unknowingly solving the same problems independently with AI at least once a week. This not only wastes hours but can also lead to conflicts when disparate AI-assisted outputs need reconciling, as exemplified by engineers debugging identical errors due to unshared chat histories and uncoordinated AI coding causing major systems conflicts. Work stalls when it’s out of view Siloed AI use significantly contributes to work delays and inefficiencies due to a lack of shared visibility. 59% of executives state that work stalls or gets delayed weekly because it's trapped in a single person's AI tool, unseen by the rest of the team. This problem is more acute in larger companies, with 73% of enterprise executives reporting weekly stalls compared to 37% at SMBs. More employees mean more private chat histories, leading to blind spots and increased time spent troubleshooting preventable issues, such as debugging AI-generated code that could have been avoided with clear visibility. Few companies have connected AI across the business The implementation strategies for AI tools contribute to existing usage and adoption challenges. While 49% of organizations are deploying AI agents capable of autonomous tasks and workflows, a mere 15% have integrated these agents seamlessly across multiple systems and interfaces to function organization-wide. This indicates that most companies are still in early stages of broad AI integration, limiting their ability to leverage AI's full potential efficiently. Toggling between AI and other work tools costs about 22 business days a year Ineffective coordination of AI efforts leads to substantial time wastage. Executives estimate that employees lose an average of 3.6 hours per week switching between disparate AI tools and other work systems. This accounts for nearly 10% of every workweek, accumulating to roughly 181 hours or 22 business days lost annually per employee. Only a small fraction (11%) of companies report having their AI tools within a single, integrated ecosystem, eliminating context-switching time. Executives want AI that works across the team When envisioning ideal AI usage, executives strongly favor collaborative approaches. 72% prefer AI that operates across teams with shared context, enabling value to compound and build upon itself. In contrast, only 26% would opt for individual AI use where value remains isolated. Leaders believe this cooperative framework enhances cross-functional efficiency, preventing redundant prompt engineering and accelerating project delivery. Slack GM Rob Seaman highlights that agents working in shared spaces can learn from collective conversations, allowing subsequent individuals and agents to build upon past interactions. Methodology The research data was gathered from a double-blind Salesforce Pulse Survey conducted between September 14 and 22, 2026. The survey included 250 U.S. executives at director level or above, all possessing decision-making authority over AI investments within their organizations. Average weekly hours lost were calculated using the lower end of response ranges and annualized over a 50-workweek year. Percentages were rounded, and SMBs were defined as companies with under 250 employees, while enterprise businesses had 2,500 employees or more.
Introduction: Interface Isolation and the Need for Multiplayer AI
As businesses increasingly rely on AI, a Salesforce survey of 250 U.S. executives reveals a bottleneck of interface isolation. 68% see AI insights go unshared, 66% report duplicated efforts, and 59% experience work stalls due to isolated AI tools. Employees waste 22 business days annually toggling between AI and other apps, a problem that intensifies in larger organizations. 72% of executives envision AI that fosters team collaboration with shared context, moving from a 'single-player' to a 'multiplayer' era of AI.
What one person learns with AI rarely reaches the team
Most AI use remains individual, leading to siloed insights. 68% of executives observe AI outputs not being shared or reused, causing stalled company output. Discoveries, like efficient prompts that can halve review time, often stay confined to one person's workflow, hindering broader team benefit and collective intelligence.
Teams are solving the same problems twice
The lack of shared AI knowledge results in duplicated efforts across organizations. 66% of executives report colleagues unknowingly solving the same problems independently with AI at least once a week. This not only wastes hours but can also lead to conflicts when disparate AI-assisted outputs need reconciling, as exemplified by engineers debugging identical errors due to unshared chat histories and uncoordinated AI coding causing major systems conflicts.
Work stalls when it’s out of view
Siloed AI use significantly contributes to work delays and inefficiencies due to a lack of shared visibility. 59% of executives state that work stalls or gets delayed weekly because it's trapped in a single person's AI tool, unseen by the rest of the team. This problem is more acute in larger companies, with 73% of enterprise executives reporting weekly stalls compared to 37% at SMBs. More employees mean more private chat histories, leading to blind spots and increased time spent troubleshooting preventable issues, such as debugging AI-generated code that could have been avoided with clear visibility.
Few companies have connected AI across the business
The implementation strategies for AI tools contribute to existing usage and adoption challenges. While 49% of organizations are deploying AI agents capable of autonomous tasks and workflows, a mere 15% have integrated these agents seamlessly across multiple systems and interfaces to function organization-wide. This indicates that most companies are still in early stages of broad AI integration, limiting their ability to leverage AI's full potential efficiently.
Toggling between AI and other work tools costs about 22 business days a year
Ineffective coordination of AI efforts leads to substantial time wastage. Executives estimate that employees lose an average of 3.6 hours per week switching between disparate AI tools and other work systems. This accounts for nearly 10% of every workweek, accumulating to roughly 181 hours or 22 business days lost annually per employee. Only a small fraction (11%) of companies report having their AI tools within a single, integrated ecosystem, eliminating context-switching time.
Executives want AI that works across the team
When envisioning ideal AI usage, executives strongly favor collaborative approaches. 72% prefer AI that operates across teams with shared context, enabling value to compound and build upon itself. In contrast, only 26% would opt for individual AI use where value remains isolated. Leaders believe this cooperative framework enhances cross-functional efficiency, preventing redundant prompt engineering and accelerating project delivery. Slack GM Rob Seaman highlights that agents working in shared spaces can learn from collective conversations, allowing subsequent individuals and agents to build upon past interactions.
Methodology
The research data was gathered from a double-blind Salesforce Pulse Survey conducted between September 14 and 22, 2026. The survey included 250 U.S. executives at director level or above, all possessing decision-making authority over AI investments within their organizations. Average weekly hours lost were calculated using the lower end of response ranges and annualized over a 50-workweek year. Percentages were rounded, and SMBs were defined as companies with under 250 employees, while enterprise businesses had 2,500 employees or more.
