Generative AI is becoming a routine workplace tool faster than many organizations are building rules around it. New research from leadership advisory firm idealis finds that 62% of U.S. workers now use generative AI professionally, up from 46% a year earlier, while only 40% say their employers have clear guidelines governing its use.
The numbers point to a growing enterprise AI problem: adoption is accelerating, but the organizational infrastructure needed to manage that adoption is not keeping pace.
The workplace AI conversation has moved beyond whether employees will use generative AI. Increasingly, the question is whether companies can establish enough structure around that use to make it productive, responsible and trusted.
Idealis’s latest workforce research suggests there is a substantial gap between the two. Its nationally representative survey, powered by CivicScience and involving more than 15,000 workers over several months, found that professional GenAI usage increased 16 percentage points year over year, reaching 62%.
Only 40% of respondents, however, said their companies have clear GenAI usage guidelines.
That disconnect matters because workplace AI is no longer limited to specialized technology teams. Employees are using large language models and generative AI applications for writing, research, analysis, customer communications, coding and other everyday tasks. As adoption spreads, organizations need to define what employees can do with these systems, what information they can provide to them and where human review remains necessary.
The research also complicates the assumption that greater exposure to AI automatically reduces employee anxiety.
Among workers who already use AI, 72% expressed concern about their own job security, compared with 46% among workers who do not yet use AI. Overall, 78% of workers remain concerned about AI’s impact on jobs, broadly unchanged from the previous year.
In other words, familiarity has not eliminated uncertainty.
AI Adoption Is Becoming a Management Issue
Leadership behavior may be one reason the policy gap is particularly significant.
According to idealis, 76% of leaders use GenAI at work, compared with 61% of non-leaders. That puts executives and managers in a position to influence not only whether AI is adopted, but also the expectations surrounding its use.
Idealis founder and CEO Dr. Sumona De Graaf argues that organizations should treat guidelines, training and trust as part of the adoption process rather than as follow-up measures.
That distinction is increasingly relevant as enterprises move from experimentation toward scaled AI deployment.
A company can provide employees with access to ChatGPT, Microsoft Copilot, Google Gemini or other AI systems in a matter of days. Establishing policies covering confidential information, intellectual property, accuracy, customer data, model outputs, human oversight and acceptable use is considerably more complicated.
The strongest enterprise AI programs therefore increasingly resemble technology-management programs rather than simple software rollouts.
Guidelines Correlate With Better Workforce Outcomes
The idealis data provides another reason for companies to pay attention to governance.
Employees at organizations with clear GenAI guidelines reported an engagement rate of 66%, compared with 54% among employees without clear guidelines. The difference was even larger for access to skill development: 80% of workers at organizations with clear guidelines reported access to skill development, compared with 56% at organizations without them.
Those findings show an association rather than proof that AI guidelines directly cause higher engagement or training access. Companies that already invest in employee development and communication may also be more likely to establish formal AI policies.
Still, the relationship is significant for enterprise leaders. AI policy does not have to be viewed solely through a risk-management lens. Clear expectations can also provide employees with permission and confidence to experiment within defined boundaries.
That may become increasingly important as organizations ask employees to incorporate AI into existing workflows.
Industry Adoption Is Uneven
GenAI adoption is also progressing at different speeds across industries.
Idealis reports the highest professional adoption rates among workers in financial services at 86%, followed by education at 85%, information and professional services at 75%, and health services at 69%.
The differences likely reflect a combination of factors, including the availability of digital knowledge work, regulatory environments, existing technology investment and the types of tasks performed by employees.
Financial services, for example, has substantial exposure to AI-driven analytics and automation but also faces stringent requirements around data protection, model risk and regulatory oversight. Healthcare combines high-value information workflows with equally significant privacy and accuracy considerations.
That makes standardized enterprise AI policies increasingly difficult to design. A generic “don’t put sensitive information into AI tools” rule may be insufficient for organizations operating across multiple functions and regulatory environments.
Instead, companies are likely to need layered policies that distinguish between approved tools, data classifications, use cases and levels of human oversight.
AI’s Workforce Divide Is Not Simply About Access
The research also highlights differences in reported AI use across racial and ethnic groups. Idealis found that 76% of Hispanic/Latino workers, 83% of Black workers and 83% of Asian workers reported using AI professionally, compared with 55% of white workers.
These figures should be interpreted carefully. Survey differences do not, by themselves, establish why adoption varies between groups.
They do, however, reinforce the importance of examining AI adoption through a workforce lens rather than treating the employee population as homogeneous.
Companies introducing AI tools need to consider who is receiving training, who has access to approved tools, which roles are being redesigned and whether employees understand how AI affects their responsibilities.
The Enterprise AI Challenge Is Moving Up the Stack
The broader market is already moving in this direction.
Microsoft, Google, Salesforce and other enterprise software providers are embedding generative AI into applications that employees already use. AI is increasingly becoming a feature of the productivity stack rather than a separate destination.
That changes the governance equation.
When AI is embedded directly into email, documents, CRM systems, development environments and analytics platforms, employees may use it without thinking of themselves as “AI users.” The technology becomes part of the workflow.
For IT, HR, security and business leaders, that means AI governance cannot remain isolated within an innovation team. It increasingly intersects with employee training, cybersecurity, data governance, compliance and workforce planning.
The idealis findings capture that transition. The adoption question appears to be resolving faster than the organizational question.
The companies most likely to benefit from workplace AI may not necessarily be those that deploy the largest number of models. They may be the ones that give employees a clear framework for using them.
As De Graaf puts it, AI can amplify existing organizational strengths and weaknesses. For enterprise leaders, the implication is straightforward: building trust and clear expectations may be just as important to scaling AI as choosing the underlying technology.
Market Landscape
The enterprise GenAI market is shifting from access to governance.
Early workplace AI programs focused heavily on experimentation: which models performed best, which departments could benefit and how quickly employees could use them. The next phase is increasingly concerned with operationalizing those systems across the workforce.
That creates a broader enterprise technology stack:
- AI applications: ChatGPT, Microsoft Copilot, Google Gemini, Salesforce AI and specialized enterprise copilots.
- AI governance: policies defining acceptable use, data handling, model oversight and accountability.
- Security and data governance: controls preventing sensitive information from reaching unauthorized AI systems.
- AI literacy: training employees to evaluate outputs, identify hallucinations and understand appropriate use cases.
- Workforce transformation: redesigning roles and workflows as AI becomes embedded into daily operations.
- Measurement: tracking productivity, engagement, adoption and business outcomes rather than AI usage alone.
The competitive advantage is consequently shifting from simply having access to generative AI toward being able to deploy it at organizational scale without losing employee trust or operational control.
Top Insights
- U.S. workplace GenAI adoption reached 62%, but only 40% of workers report clear employer guidelines, highlighting a widening enterprise governance gap.
- AI users report greater job-security anxiety than non-users, suggesting familiarity with generative AI is not eliminating workforce uncertainty about automation.
- Organizations with clear AI guidelines report higher engagement and skill-development access, although the survey shows correlation rather than causation.
- Financial services, education, information and professional services lead workplace GenAI adoption, creating different governance requirements across highly varied enterprise environments.
- Enterprise AI strategy is expanding beyond model selection toward training, data governance, responsible use, workforce planning and measurable business outcomes.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
