How Generative AI Is Really Impacting the Job Market — According to Microsoft
- Chris Howell
- Aug 2
- 5 min read
What happens when you analyse 200,000 real conversations between people and an AI assistant? You stop guessing how AI might change the job market—and start seeing how it already is. And the results aren’t just theoretical—they’re based on real data, real tasks, and real work interactions.
That’s exactly what Microsoft just did.
In a major new study, Microsoft Research and LinkedIn analysed how people actually use Bing Copilot (now rebranded as Microsoft Copilot). This isn’t another prediction model or speculative report. It’s a deep dive into what people asked AI to help with, what AI actually did, and what that means for real-world jobs across industries, education levels, and salary bands. The data is real. The patterns are surprising. And the implications are already reshaping the way we work and think about the future of employment.
Here’s what stood out—and why it matters.
AI Is Being Used for Real Work — And It's Good at It
Most people use Copilot for tasks that fall into three broad categories:
Gathering information (e.g. researching a topic, summarising an article, checking facts, finding sources)
Writing (e.g. drafting emails, blog posts, internal reports, CVs, or business proposals)
Explaining (e.g. helping someone understand a concept, breaking down jargon, describing how something works)
In those conversations, the AI most often acts as a:
Teacher (breaking down information clearly and concisely)
Advisor (suggesting next steps, tools, or options to take)
Editor (reviewing, rewriting, and enhancing user-generated content)
Crucially, the Microsoft team found that AI completes these tasks well. Users gave consistently positive feedback for activities like writing, editing, and researching—and task completion rates (evaluated by another AI model) strongly supported those findings. Copilot isn’t fumbling its way through tasks. It’s delivering meaningful, usable outputs that users value.
And this isn’t niche usage. The scale of real-world adoption makes it clear: AI has entered the productivity mainstream.
Not Just Tech Jobs: Sales, Support, and Admin Roles Are in the Firing Line
It’s tempting to assume AI is only coming for high-skill tech jobs—developers, copywriters, maybe a few analysts. But this study revealed that sales, customer service, and admin roles are among the most impacted occupations when it comes to AI applicability.
Some of the top jobs where AI is already delivering value include:
Interpreters and Translators
Writers and Editors
Sales Representatives
Customer Service Advisors
Data Scientists
Public Relations Specialists
Journalists and News Analysts
Technical Writers
Web Developers
Brokerage Clerks and Travel Agents
Even roles like radio DJs, host/hostess positions, and personal financial advisors made the list.
Why is this happening? Because many of these jobs involve:
High volumes of written or verbal communication
Explaining technical or commercial information
Managing information and producing documents under time pressure
These are precisely the kinds of tasks that Copilot and tools like ChatGPT are built to handle. In short: the more your job involves turning information into value, the more likely it is that AI will be a key part of it going forward.
Meanwhile, Manual Jobs Are (Still) Mostly Safe
At the other end of the spectrum, AI had almost no measurable impact on roles such as:
Nurses and Care Assistants
Roofers
Plumbers
Cleaners
Machine Operators
Construction Workers
Maintenance Staff
Warehouse Pickers
These roles share a common theme: they require hands-on physical work, manual precision, and often direct human presence or empathy. Generative AI, at its current stage, simply can’t replicate those qualities.
A self-employed roofer might use AI to generate quotes
A care worker might write a shift report faster using a chatbot
A cleaner running a small business could use AI to create flyers or social posts
So even if AI isn’t replacing these jobs, it’s still nudging into the edges of how these professionals manage their workload.
AI Is Augmenting Work, Not Always Automating It
One of the most fascinating insights from the study? In 40% of Copilot conversations, the task the user was trying to complete and the action the AI actually performed didn’t match.
That gap matters—and here’s why:
User goal: Fix a technical issue (e.g. printer won’t connect)
AI action: Provide detailed guidance or instructions (effectively a digital coach)
In these cases, AI is not replacing the user. It’s amplifying them. Acting as a support system or second brain, AI helps users move faster, avoid mistakes, and improve quality.
This is what economists call task-level augmentation. It’s not automation in the sense of “a robot replaces the worker.” Instead, it’s a scenario where the worker gets a new kind of tool—one that scales their thinking, sharpens their writing, or accelerates their research.
That’s a game-changer for knowledge work—and it could mean fewer layoffs and more output across creative, analytical, and client-facing roles.
Education and Wages? The Trends Are Surprising
One of the more unexpected findings from the report was the weak link between AI impact and wage level.
You might assume that higher-paid roles are more exposed to AI. But the data doesn’t bear that out:
The correlation between income and AI applicability is surprisingly weak
Jobs requiring a Bachelor’s degree are slightly more affected—but not dramatically so
Many mid-wage or lower-wage roles (especially in customer support and sales) are among the most AI-augmented
That means AI isn’t just about optimising executive workflows or high-end consulting. It’s also transforming how entry-level and junior workers perform everyday tasks—often in large numbers.
In other words, AI could be the biggest force reshaping the middle of the workforce, not just the top.
So What Should You Do?
This isn’t hypothetical. Generative AI is already changing how jobs are done. If you work in a job that involves any of the following:
Writing or creating documents
Interpreting or communicating information
Researching, summarising, or comparing options
Explaining technical or business processes
…then you need to be learning how to use AI now. Not later.
The same applies if you run a business, manage a team, or hire staff. The tools that help you work smarter are already here—and if you’re not using them, you’re likely falling behind.
If you’re a jobseeker? AI has already become part of how hiring works. Companies use it to filter CVs. Recruiters use it to screen candidates. And the smartest applicants are using it to:
Tailor CVs to specific job descriptions
Write targeted, engaging cover letters
Prepare for competency-based interview questions
Roleplay interviews to gain confidence
If you're not using AI in your job search, you're putting yourself at a disadvantage—whether you're in tech, retail, marketing, or management.
Want Help Navigating It All?
Mercia AI offers:
1-to-1 Discovery Calls: Explore how AI fits your role, industry, or career goals
AI Jobhunter Webinars: Learn how to use AI to stand out in your job search
AI-Optimized CV Review: Get AI-powered edits tailored to your next role
AI Strategy Builder: For freelancers, teams, and small businesses
We meet you where you are—whether you're new to AI or already experimenting.
The Microsoft study confirms what we’ve said from day one:
AI won’t take your job. But someone using AI might.
Let’s make sure that someone is you.
