Over 25 years of serving tech enthusiasts.[19459025]

Context: GenAI is being used in the workplace at a rapid pace. Two new reports, one from Gallup, and the other from Salesforce AI Research, paint a picture that reflects both the growing enthusiasm for this technology and its challenges. These studies reveal a common theme: AI is more widely accepted than ever before, but the gap between its promises and practical performance remains an obstacle for both businesses and employees.

Gallup’s latest Research shows that AI usage among US employees nearly doubled in the last two years. In 2023, only 21 percent of employees reported using AI at least a few time a year in their jobs. By 2025, this figure had risen to 40 percent. Frequent usage – defined as using AI at least a few days a week – has nearly doubled from 11 to 19 percent. Daily AI usage, which is still a minority practice, has nearly doubled over the last year, reaching 8 per cent. This increase is most noticeable among professionals in the white-collar sector. Twenty-seven per cent of white-collar workers now report frequent AI usage, up 12 points since last year. Technology, professional services, as well as finance lead the way. Frontline and production workers, however, have seen little change. Frequent AI use has actually decreased from 11 percent to nine percent during the same time period.

Leadership roles are also more likely to embrace AI. One in three managers of managers use AI several times a week or more, compared to just 16 percent of individual contributors.

Yet, despite this rapid adoption, most employees do not feel their jobs are under immediate threat from automation. Only 15 percent believe AI or robots are likely to eliminate their roles within the next five years, a figure unchanged from previous years, though concern is somewhat higher in sectors like technology, retail, and finance.

A striking finding from Gallup is the disconnect between AI integration and organizational guidance. While 44 percent of employees say their workplace is rolling out AI, only 22 percent have received a clear plan or strategy. Just 30 percent report the existence of any guidelines or formal policies, leaving many to navigate this new terrain without clear direction.

The most common challenge cited is the lack of a clear use case or value proposition, and only 16 percent of AI users strongly agree that the tools provided are truly useful for their work.

Gallup data shows that leadership is crucial in determining whether employees are comfortable and prepared to work with AI. Workers who feel their leaders have communicated an AI strategy clearly are three times more likely to be very prepared to use the technology, and twice as likely feel comfortable using it.

This research also shows that perceptions of AI are very different. 68 percent of employees who have used AI to interact with clients believe it improves the interactions, while only 13 percent of non-users agree.

Gallup’s findings emphasize the human side of AI adoption. Salesforce AI Research focuses on the technology, concluding that AI agents are still far from what is required in real-world enterprise scenarios.

According to the Salesforce study, even the best LLM agents only succeed in 58 percent of tasks that require a single step. When tasks become complex and require multiple-turn interactions such as clarifications or follow-up questions, success rates drop to around 35 per cent.

Workflow tasks where agents follow rules are the easiest to execute for current models. Top performers reach over 83 per cent success. Nevertheless, policy compliance, textual reason, and database queries are still significant weaknesses.

One of the most troubling findings is that AI agents are almost completely unaware of confidentiality. These systems almost never refuse to handle sensitive data unless explicitly instructed to do so. Although adding confidentiality prompts to their systems can improve their ability of refusing inappropriate requests, it often comes at a cost to overall task performance. This highlights a difficult tradeoff between safety and utility.

The study also found that models with better reasoning abilities performed better in all tasks and that agents that asked more clarification questions during multi-turn scenarios had higher success rates.

While some models, like Gemini-2.5 pro and Gemini 2.5 flash, are attractive because they offer a good compromise between performance and cost, the gap between AI capabilities, and business needs, remains wide.

Both studies agree that AI is making significant progress in the workplace but the journey is not yet complete. Gallup’s research suggests that employees need to be trained and communicated with by their leaders in order to benefit from AI. Salesforce’s research shows that the technology is still far from being able to handle the complexity and responsibility of real business environments.