Google Cloud survey finds AI delivers on customer service while ROI drops

The artificial intelligence honeymoon period appears to be ending. After years of breathless predictions about AI transforming every aspect of business, a new survey from Google Cloud reveals that executives are taking a more measured approach to evaluating their AI investments—and the results are decidedly mixed.

The survey, which polled nearly 3,500 senior leaders globally from companies with at least 100 employees and $10 million in annual revenue, focused exclusively on organizations that have already deployed generative AI systems. Generative AI refers to artificial intelligence tools like ChatGPT that can create text, images, code, and other content based on prompts, rather than simply analyzing existing data.

While 74 percent of respondents reported achieving return on investment from at least one AI use case within a year—unchanged from 2024—the definition of ROI here extends beyond pure financial gains to include productivity improvements, faster time to market, enhanced customer experience, and better security. When examining specific metrics more closely, a more nuanced picture emerges of where AI is actually delivering value.

Customer experience emerges as the standout winner

Among all the business applications surveyed, AI-powered customer experience stands out as the only area showing clear year-over-year improvement. Nearly 62 percent of executives reported that generative AI solutions had enhanced their customer experience, up from 59 percent in 2024. This makes customer service the one bright spot in an otherwise plateauing landscape.

Of the leaders who reported customer experience improvements, 75 percent noted increased user satisfaction scores. While this figure represents a slight decline from 80 percent in 2024, the overall trend suggests that AI chatbots, automated support systems, and personalized customer interactions are finding their footing in real-world applications.

This success likely stems from AI’s natural fit for customer service tasks. Unlike complex analytical work that requires nuanced judgment, customer support often involves pattern recognition and response generation—core strengths of current generative AI systems. Companies can deploy AI to handle routine inquiries, provide 24/7 availability, and offer consistent responses across multiple languages and channels.

Revenue and productivity gains show concerning declines

Despite continued investment and deployment, AI’s impact on core business metrics appears to be softening. Only 40 percent of respondents reported direct revenue growth from generative AI solutions, down from 44 percent in 2024. This decline suggests that the initial wave of AI implementations may have captured the most obvious opportunities, leaving more challenging applications for subsequent phases.

Productivity improvements, long considered AI’s most promising application, also showed modest erosion. While 70 percent of executives still reported productivity gains—down slightly from 71 percent in 2024—the magnitude of these improvements has decreased notably. Just 39 percent of leaders said employee productivity had at least doubled due to AI solutions, compared to 45 percent who reported such dramatic gains in 2024.

This productivity plateau reflects the reality that integrating AI into complex workflows requires more than simply deploying the technology. Employees need training, processes require redesign, and organizations must develop new management approaches to capture AI’s benefits effectively.

Marketing shows promise while security struggles

AI-powered marketing emerged as a relatively bright spot, with 55 percent of surveyed executives reporting meaningful improvements to their marketing efforts. Media, entertainment, retail, and consumer packaged goods companies showed particular enthusiasm, with 59 percent of leaders in these sectors citing marketing benefits.

This success makes intuitive sense given AI’s capabilities in content generation, audience segmentation, and campaign personalization. Marketing teams can leverage AI to create multiple versions of ad copy, analyze customer behavior patterns, and optimize campaign performance across different channels and demographics.

Security applications, however, showed more troubling trends. Only 49 percent of respondents reported meaningful security improvements from generative AI, down from 56 percent in 2024. Even more concerning, just 53 percent of those who did see security benefits reported reduced security incident tickets, compared to 65 percent who reported such improvements in 2024.

This decline may reflect the growing sophistication of AI-powered cyber threats, which are evolving as quickly as AI-powered defenses. As malicious actors adopt similar technologies, the defensive advantages of AI may be diminishing, creating an arms race dynamic that limits net security improvements.

The spending paradox: more investment despite mixed results

Perhaps most intriguingly, organizational AI spending continues to accelerate even as performance metrics show mixed or declining results. Nearly half of respondents (48 percent) reported reallocating funds from non-AI budgets to support new generative AI investments, up from 44 percent in 2024.

This spending pattern suggests several possible explanations. Companies may be doubling down on AI investments because they view current implementations as learning experiences that will inform more successful future deployments. Alternatively, competitive pressure may be driving continued investment regardless of immediate returns, as organizations fear falling behind rivals who might achieve breakthrough results.

The reallocation of funds from other initiatives also indicates that AI investment is increasingly viewed as essential rather than experimental. Rather than finding new money for AI projects, companies are treating these investments as core business requirements worthy of displacing other priorities.

Implications for business leaders

Oliver Parker, Vice President of Google Cloud’s global generative AI go-to-market team, characterizes these findings as evidence of a “fundamental change in business mindset.” The conversation has shifted from hype-driven adoption to value-focused evaluation, suggesting a more mature approach to AI deployment.

For business leaders, these results offer several key insights. First, customer experience applications appear to offer the most reliable path to AI success, making customer service automation a logical starting point for organizations beginning their AI journey. Second, productivity and revenue gains require more sophisticated implementation strategies than simply deploying AI tools and expecting immediate results.

The data also suggests that successful AI adoption may require longer time horizons and more patient capital than initially anticipated. While 74 percent of organizations achieve some form of ROI within a year, the declining magnitude of benefits indicates that sustained value creation demands ongoing refinement and optimization.

Most importantly, the continued increase in AI spending despite mixed results suggests that business leaders view this technology as strategically essential rather than tactically optional. Organizations that delay AI adoption risk falling behind not just in efficiency, but in their fundamental ability to compete in an increasingly AI-enabled business environment.

The cooling of AI hype doesn’t signal the technology’s failure—rather, it marks the beginning of its maturation into a standard business tool that requires the same careful planning, implementation, and measurement as any other significant operational investment.