Microsoft study shows AI impacts 75% of major occupations in marketing communications

Microsoft researchers have documented widespread synthetic intelligence capabilities throughout main occupation classes, with gross sales representatives, customer support employees, and writers exhibiting the very best potential for AI collaboration. The great evaluation, printed July 22, 2025, examined 200,000 anonymized conversations between customers and Microsoft Bing Copilot to grasp which work actions individuals are utilizing AI to help with or carry out.

In response to the examine titled Working with AI: Measuring the Occupational Implications of Generative AI, interpreters and translators face the very best AI applicability, with 98% of their work actions overlapping with frequent AI utilization patterns. Customer support representatives, using 2.86 million folks nationwide, ranked among the many high occupations for AI applicability alongside gross sales representatives, who signify over 1.14 million employees.

The analysis crew from Microsoft Analysis, led by Kiran Tomlinson, Sonia Jaffe, Will Wang, Scott Counts, and Siddharth Suri, analyzed conversations from January 1 by means of September 30, 2024. Their methodology categorized every dialog into work actions as outlined by the O*NET database, which decomposes occupations into particular duties and actions carried out by employees.

“Given the fast adoption of generative AI and its potential to influence a variety of duties, understanding the consequences of AI on the financial system is one among society’s most vital questions,” the researchers said of their findings.

The examine reveals essential distinctions between AI help and AI efficiency. In 40% of conversations analyzed, customers sought assist with totally totally different actions than what the AI really carried out. Customers primarily sought AI help for info gathering, writing, and communication duties, whereas AI mostly acted in service roles as coaches, advisors, or lecturers.

Data gathering emerged as probably the most profitable AI software, receiving the very best optimistic suggestions from customers at 78% satisfaction charges. Writing actions adopted intently at 76% satisfaction, whereas visible design and knowledge evaluation duties confirmed the bottom success charges at roughly 55%.

Main occupation teams confirmed various ranges of AI applicability. Gross sales and associated occupations scored highest at 0.32 on the researchers’ applicability scale, adopted by pc and mathematical occupations at 0.30, and workplace and administrative assist at 0.29. These three classes signify over 36 million employees mixed.

The evaluation discovered minimal correlation between AI applicability and wage ranges, opposite to predictions that higher-wage occupations would face higher AI influence. The employment-weighted correlation between AI applicability and wages measured solely 0.07, suggesting AI capabilities span throughout earnings ranges slightly than concentrating in high-wage positions.

Academic necessities confirmed stronger patterns. Occupations requiring bachelor’s levels demonstrated greater AI applicability scores averaging 0.27, in comparison with 0.19 for positions with decrease instructional necessities. Nevertheless, substantial overlap exists throughout all schooling ranges.

Bodily occupations confirmed the bottom AI applicability scores. Occupations involving handbook labor, equipment operation, or direct bodily interplay with folks constantly ranked on the backside of the applicability scale. Dredge operators, bridge tenders, and water therapy plant operators confirmed AI applicability scores close to zero.

The analysis methodology integrated three key measures: job completion charges, consumer satisfaction by means of thumbs-up suggestions, and scope of influence. Actions receiving reasonable or greater scope rankings indicated AI may help with substantial parts of the work concerned.

Writing and enhancing actions demonstrated the very best completion charges at over 85%, whereas visible design and scientific knowledge evaluation confirmed completion charges beneath 65%. The examine discovered sturdy correlation between consumer satisfaction and job completion throughout totally different work actions.

When in comparison with predictions from a 2024 examine by Eloundou et al., the Microsoft analysis confirmed outstanding alignment. The correlation between predicted AI influence and measured AI applicability reached 0.73 on the occupation stage and 0.91 when aggregated to main occupation teams.

The researchers emphasised their findings signify solely AI capabilities, not precise office implementation or financial outcomes. “Our knowledge don’t point out that AI is performing the entire work actions of anyone occupation,” they famous. Process completion charges hardly ever reached 100%, and scope of influence usually remained at reasonable ranges.

For the advertising and marketing group, these findings carry important implications. Sales and advertising roles dominate the highest AI applicability categories, suggesting basic adjustments in how advertising and marketing professionals work together with prospects and prospects. The examine’s deal with communication and data provision aligns with rising tendencies the place AI agents may replace human attention as advertising targets.

Customer support representatives, a vital element of selling operations, ranked fourth amongst all occupations for AI applicability with scores of 0.44. This discovering helps current business analysis exhibiting marketers save an average of 114 minutes weekly through AI integration, translating to $3,520 in annual labor value reductions per worker.

The timing coincides with broader business recognition of AI’s transformative potential. Current surveys present 68% of marketers plan to increase social media spending while incorporating AI tools, with automation rising because the fastest-growing funding space.

Writers and authors, essential to content material advertising and marketing methods, confirmed AI applicability scores of 0.45, rating fifth amongst all occupations studied. Technical writers, public relations specialists, and editors all appeared within the high 25 occupations for AI applicability, suggesting content material creation workflows will expertise important transformation.

The analysis recognized particular work actions most fitted for AI collaboration. “Present info to prospects” and “reply to buyer issues or inquiries” ranked among the many highest for each consumer satisfaction and job completion. These actions signify core capabilities in advertising and marketing buyer engagement methods.

Notably, the examine discovered AI performs totally different actions than customers request help with. Whereas customers search assist with analysis and evaluation, AI usually offers clarification and communication providers. This asymmetry suggests profitable AI implementation requires understanding each consumer wants and AI capabilities.

The researchers acknowledged limitations of their evaluation. The examine examined just one AI platform and can’t decide what conversations occurred in work contexts versus private use. Moreover, the decomposition of occupations into work actions, whereas commonplace apply, could not seize the total complexity {of professional} roles.

Future analysis instructions embody monitoring how AI capabilities evolve over time and understanding how occupations restructure their obligations in response to AI development. The researchers famous that totally new occupations could emerge, as has occurred with earlier technological transformations.

The Microsoft examine represents the primary large-scale evaluation of precise AI utilization patterns throughout occupations, offering empirical proof for understanding AI’s office influence. As synthetic intelligence capabilities proceed increasing, these findings provide essential insights for workforce planning {and professional} growth methods.

Timeline

Abstract

Who: Microsoft Analysis crew led by Kiran Tomlinson, Sonia Jaffe, Will Wang, Scott Counts, and Siddharth Suri analyzed 200,000 conversations between customers and Microsoft Bing Copilot, affecting hundreds of thousands of employees throughout gross sales, customer support, writing, and advertising and marketing occupations.

What: Complete examine measuring AI’s occupational implications throughout 844 work actions and 104 professions, revealing that 75% of main occupation teams present potential for AI collaboration, with gross sales representatives, customer support employees, and writers going through the very best automation potential.

When: Analysis analyzed conversations from January 1 by means of September 30, 2024, with the ultimate examine printed July 22, 2025, coinciding with broader business recognition of AI’s transformative influence on advertising and marketing and customer support operations.

The place: United States workforce evaluation utilizing O*NET occupational database and Bureau of Labor Statistics employment knowledge, overlaying 164 million employees throughout computer-compatible occupations from administrative roles to artistic professions.

Why: Understanding AI’s financial influence has turn out to be essential as 40% of Individuals report utilizing generative AI, with advertising and marketing professionals more and more adopting AI instruments that save a mean of 114 minutes weekly and $3,520 yearly per worker, whereas 68% plan to extend AI-enhanced digital promoting spending.

Key Phrases Defined

AI Applicability Rating: The great metric developed by Microsoft researchers to measure potential occupational influence by combining three components: frequency of AI utilization for particular work actions, job completion charges, and scope of influence. This rating ranges from 0 to 1, with greater values indicating higher potential for AI collaboration. The methodology offers a standardized technique to evaluate AI influence throughout totally different occupations, shifting past subjective predictions to empirical measurement based mostly on precise utilization patterns.

Work Actions: The elemental constructing blocks used to research occupational duties, derived from the O*NET database’s hierarchical classification system. These actions signify particular capabilities carried out throughout a number of occupations, similar to “collect info from sources” or “present buyer help.” The examine centered on 332 intermediate work actions (IWAs) slightly than occupation-specific duties, permitting researchers to determine how AI capabilities demonstrated in a single context translate throughout all occupations performing related actions.

Process Completion Charge: A essential success metric measuring how successfully AI programs end user-requested work, decided by means of automated classification by GPT-4o-mini analyzing dialog outcomes. The researchers validated this metric in opposition to consumer thumbs-up suggestions, discovering sturdy correlation (r > 0.75) between completion charges and consumer satisfaction. Writing and data gathering duties achieved the very best completion charges above 85%, whereas visible design and knowledge evaluation confirmed considerably decrease success charges.

Buyer Service Representatives: The occupation class using 2.86 million employees that ranked among the many highest for AI applicability with scores of 0.44. These roles contain core actions extremely appropriate for AI help, together with responding to buyer inquiries, offering info, and resolving issues. The discovering suggests basic adjustments forward for customer support operations, as AI demonstrates sturdy capabilities in communication and data provision duties that represent major job capabilities.

Gross sales Representatives: A broad occupational class encompassing over 1.14 million employees throughout numerous industries, exhibiting excessive AI applicability scores averaging 0.46 for service gross sales roles. The analysis recognized gross sales actions like offering product info, explaining technical particulars, and responding to buyer inquiries as significantly well-suited for AI collaboration. This discovering aligns with business tendencies towards AI-powered buyer relationship administration and automatic lead qualification processes.

Data Gathering: Essentially the most generally requested consumer exercise in AI conversations, encompassing analysis, knowledge assortment, and data acquisition duties. This exercise class acquired the very best consumer satisfaction rankings at 78% and demonstrated sturdy job completion charges above 85%. The prevalence of data gathering displays AI’s core power in processing and synthesizing giant quantities of knowledge shortly, making it significantly useful for research-intensive occupations like journalism, evaluation, and consulting.

O*NET Database: The great occupational info system developed beneath U.S. Division of Labor sponsorship that offered the foundational framework for this analysis. O*NET’s hierarchical construction breaks down 874 occupations into particular duties, which map to detailed work actions, intermediate work actions, and generalized work actions. This standardized taxonomy enabled researchers to systematically analyze AI influence throughout the whole U.S. workforce construction.

Microsoft Bing Copilot: The publicly accessible generative AI system that served as the information supply for this analysis, offering 200,000 anonymized conversations from January by means of September 2024. The selection of Copilot provided insights into mainstream AI utilization patterns, as customers entry this method with out specialised coaching or premium subscriptions. The researchers famous that totally different AI platforms could present various utilization patterns, with Copilot demonstrating much less deal with programming duties in comparison with different AI programs.

Occupational Affect: The broader idea of how synthetic intelligence impacts job roles, encompassing each direct automation of duties and augmentation of human capabilities. The analysis distinguished between AI help (serving to customers full duties) and AI efficiency (immediately executing work actions), discovering these typically contain totally different actions inside the identical dialog. Understanding occupational influence requires analyzing not simply technological capabilities but additionally consumer adoption patterns and job success charges.

Scope of Affect: A qualitative measure ranking how comprehensively AI capabilities deal with the total vary of labor concerned in particular actions, categorized on a six-point scale from “none” to “full.” This metric helps distinguish between slender AI functions that deal with small parts of complicated work actions and broad capabilities that deal with substantial parts of occupational duties. The analysis discovered that AI usually demonstrates reasonable scope influence, suggesting complementary slightly than substitute relationships with human employees.