$Microsoft (MSFT.US)$ Executives from Microsoft and other enterprise software companies once hailed 2025 as the inaugural year for AI to automate multi-step tasks—such as generating dashboards based on corporate sales data. However, as the year-end approaches, Microsoft has lowered its market penetration expectations for its next-generation AI products, known as 'intelligent agents,' and is no longer pushing customers to pay for these products quickly. Microsoft shares fell slightly in pre-market trading and are currently down nearly 2%.
According to two salespeople from Microsoft's Azure cloud division, several departments failed to meet their AI product sales growth targets in the fiscal year ending June. As a result, multiple business units within Microsoft have now reduced their sales growth forecasts for certain AI products. The two informed sources stated that it is rare for Microsoft to adjust sales targets downward for specific products.
This adjustment reflects Microsoft’s response to corporate clients' resistance to paying for AI. Over the past year, corporate customers have generally complained about two issues: first, it is difficult to quantify the cost savings brought by AI in tasks such as report writing (e.g., customer expenditure reports, lead organization); second, in scenarios with extremely low tolerance for errors—such as financial automation and cybersecurity—even minor mistakes could lead to significant losses, and AI tools’ reliability still falls short of meeting requirements.
Nonetheless, AI remains the core driver of Microsoft’s business growth. This is primarily due to two factors: one is substantial investments from AI companies like OpenAI (it is forecasted that OpenAI will lease cloud servers worth approximately $15 billion from Microsoft this year); the other is strong sales performance of Microsoft's own AI software, such as the 365 Copilot office tool and GitHub Copilot coding assistant. Additionally, Microsoft and other tech giants have significantly enhanced operational efficiency through internal use of AI tools.
However, driving traditional enterprises to increase spending on advanced AI products is not an easy task.
Take the private equity firm Carlyle Group, for example. Last year, the company began using Microsoft’s Copilot Studio—a product that allows users to develop AI tools without coding, enabling automation of tasks such as generating meeting minutes and drafting financial models based on Excel spreadsheets.
However, according to a person directly familiar with the matter and another informed source, after several months of using the tool, representatives from Carlyle provided feedback to Microsoft that its AI tools struggled to reliably access data from other applications, such as the Salesforce customer relationship management system—a key requirement for some of Carlyle’s automation scenarios. These individuals noted that in the fall of this year, Carlyle had scaled back its spending on the tool.
Editor/jayden
