As generative AI tools like Microsoft Bing Copilot flood workplaces, businesses and workers grapple with a critical question: which jobs are most susceptible to transformation? A landmark study from Microsoft Research, published in July 2025, provides unprecedented insights by analyzing 200,000 anonymized user conversations with Copilot. This data-driven approach moves beyond theoretical predictions to reveal how AI is actively reshaping daily tasks—and which occupations are at the forefront of this shift.

The Anatomy of AI-Assisted Work

Researchers Kiran Tomlinson, Sonia Jaffe, Will Wang, Scott Counts, and Siddharth Suri classified work activities from the Copilot dataset, uncovering that users most frequently seek AI help for gathering information (e.g., research, data analysis) and writing (e.g., drafting emails, reports). Conversely, AI itself excels at providing information, assistance, teaching, and advising. By measuring task success rates and the scope of impact—such as time saved or output quality—the team computed an AI applicability score for each occupation. This score quantifies how effectively AI augments or automates core functions, offering a nuanced view beyond hype.

Occupations in the AI Crosshairs

High AI applicability scores clustered in knowledge-intensive fields:
- Computer and mathematical roles: Developers and data scientists use AI for coding assistance and problem-solving, accelerating innovation.
- Office and administrative support: Tasks like scheduling, documentation, and communication see efficiency boosts, reducing manual workload.
- Sales positions: AI aids in client research, proposal writing, and personalized outreach, enhancing information-driven interactions.

Surprisingly, the study found weaker correlations between AI applicability and factors like wage or education level. High-paying tech jobs aren't the only ones affected—admin and sales roles, often mid-wage, show significant AI integration. This suggests generative AI's democratizing potential but also raises red flags about job displacement in routine information tasks.

Broader Implications for the Tech Ecosystem

For developers and tech leaders, these findings underscore the urgency of adapting skills. As AI handles more writing and research, human roles must evolve toward complex problem-solving and ethical oversight. Companies investing in AI tools should prioritize training to bridge gaps, ensuring workers harness Copilot-style systems for augmentation rather than replacement. The research also validates concerns about economic inequality, as roles with high AI applicability may face destabilization without proactive policy. Ultimately, this real-world data paints a future where AI's promise hinges not just on capability, but on strategic human-AI collaboration.

Source: Microsoft Research study, "Working with AI: Measuring the Occupational Implications of Generative AI," July 2025.