Microsoft and other enterprise software executives had touted 2025 as the breakthrough year for AI automating multi-step tasks—such as generating dashboards from company sales data. However, as the year-end approaches, Microsoft has scaled back market penetration expectations for its next-gen AI products, dubbed "intelligent agents," and no longer pushes clients to adopt paid versions rapidly.
According to two sales personnel from Microsoft's Azure cloud division, multiple business units have lowered sales growth targets for certain AI products after failing to meet growth goals in the fiscal year ending June. The sources noted that such downward adjustments for specific products are rare for Microsoft.
This shift reflects Microsoft's response to corporate clients' reluctance to pay for AI solutions. Over the past year, enterprises have widely complained about two key issues: the difficulty in quantifying cost savings from AI in tasks like report generation (e.g., client expenditure reports or sales lead organization), and the insufficient reliability of AI tools in high-stakes scenarios like financial automation and cybersecurity—where minor errors could lead to significant losses.
Despite this, AI remains a core growth driver for Microsoft, supported by two major factors: heavy investments from AI firms like OpenAI (which is projected to lease $15 billion worth of cloud servers from Microsoft this year) and strong sales of Microsoft's own AI software, such as the 365 Copilot productivity suite and GitHub Copilot coding assistant. Additionally, Microsoft and other tech giants have significantly improved operational efficiency by deploying AI tools internally.
However, convincing traditional enterprises to increase spending on advanced AI products remains challenging.
For example, private equity firm Carlyle Group began using Microsoft's Copilot Studio last year—a no-code tool for developing AI applications that automates tasks like meeting note generation and financial model drafting in Excel.
Yet, according to sources familiar with the matter, after several months of use, Carlyle informed Microsoft that the AI tool struggled to reliably integrate data from external applications like Salesforce CRM—a critical requirement for some of Carlyle's automation workflows. The sources added that Carlyle reduced its spending on the tool this fall.
