2025 Work Wrapped: AI is Changing Knowledge Work Fast, and Skilled Trades Are Becoming Even More Essential
It is New Year's Eve, and if you are doing a mental "2025 wrapped" on work, one trend keeps showing up: generative AI moved from novelty to daily tool. A recent Microsoft Research paper, Working with AI: Measuring the Occupational Implications of Generative AI, put hard numbers behind that shift by analyzing how people actually used Bing Copilot for work tasks and then mapping those activities to occupations.
Their metric, an "AI applicability score," blends (1) how much an occupation's work activities show up in Copilot usage ("coverage"), (2) how often Copilot successfully completes the task ("completion"), and (3) whether the AI's help is meaningful in scope ("scope"). In other words, it measures where AI is already showing up in real work, not a crystal ball that claims which jobs will disappear.
Still, the lists in the paper's appendix read like a year-end recap of where AI is accelerating and where the human, physical world still dominates. That is exactly why trade professions are not just "safe"—they are strategically important for 2026 and beyond.
📄 Read the Microsoft Research publication and view the arXiv preprint (or PDF).
In the paper's "top 40" occupations with the highest AI applicability scores (Appendix Table S3), you see roles where the core output is language, knowledge, or structured information. Examples include:
- Interpreters and translators
- Historians
- Writers and authors
- Customer service representatives
- Telemarketers
- Technical writers, editors, proofreaders
- Public relations specialists
- Management analysts
- Data scientists and web developers
That pattern makes intuitive sense. If your day is heavy on reading, writing, summarizing, explaining, researching, drafting, and communicating, generative AI is already a strong copilot. The research shows that people are using AI for those activities and the systems are often successful at completing them at meaningful scope.
Source: Appendix Table S3 and methodology in the paper. arXiv PDF
The Overlooked Story: Many Hands-On Jobs Sit at the Bottom of AI Applicability Today
In the paper's "bottom 40" occupations with the lowest AI applicability scores (Appendix Table S4), you see jobs where the work is anchored in the physical world, safety-critical judgment, real-time coordination, and hands-on execution. Examples include:
- Helpers in extraction work
- Hazardous materials removal workers
- Highway maintenance workers
- Cement masons and concrete finishers
- Roofers
- Paving, surfacing, and tamping equipment operators
- Logging equipment operators
- Pile driver operators
- Water and wastewater treatment plant and system operators
- Bridge and lock tenders, dredge operators
This does not mean these roles will never use AI. It means that, based on observed Copilot usage and what AI can currently complete at meaningful scope, the center of gravity for these occupations is still physical execution, not text-based production.
Source: Appendix Table S4. arXiv PDF
Why Trade Professions Matter More in a World with More AI, Not Less
If 2025 was the year AI got invited into the office, 2026 will be the year the real-world reminder hits: you cannot prompt your way into a finished hospital wing, a repaired power line, a fire-hardened roof, or a functioning water system.
Here is what makes trades a long-term cornerstone in an AI-shaped economy:
1. The Physical World Is the Bottleneck
Demand is rising for energy infrastructure, housing, transit, climate resilience, and industrial buildout. Even if AI speeds up planning and paperwork, someone still has to install, repair, inspect, and certify real assets under real constraints.
2. Safety and Accountability Are Non-Negotiable
A good electrician, lineman, HVAC tech, or heavy equipment operator is not just doing tasks. They are managing risk. They translate code, physics, and safety into execution. That judgment is hard to automate and expensive to get wrong.
3. Trades Are Becoming More Technical—and That Is a Feature
The modern tradesperson is already working with advanced systems: smart panels, heat pumps, energy storage, EV charging, building automation, sensors, and diagnostics. That technical complexity is a durable advantage.
4. AI Will Amplify Trades, Not Replace Them
The highest-value future is not "AI versus trades." It is "AI plus trades." Estimating, scheduling, parts ordering, documentation, training, troubleshooting, and compliance workflows are ripe for AI assistance, while the hands-on work remains human-led.
What This Means for Anyone Choosing a Career Path in 2026
If you are considering a trade, or advising someone who is, the opportunity is bigger than "job security." It is leverage.
The Best Bet Is Becoming a Modern Craft Professional
A modern craft professional pairs hands-on mastery with digital and business fluency:
1. Core Craft Skill
Apprenticeship, hours, code knowledge, and real job-site reps.
2. Tool-Assisted Productivity
Comfort using AI for estimating drafts, customer communication, summarizing code requirements, creating checklists, writing job notes, generating training plans, and learning faster.
3. Business Readiness
Many trades lead to ownership: running a crew, starting a shop, building recurring service revenue, or specializing in high-value niches.
Practical Moves You Can Make This Quarter
Pick one trade lane that has strong local demand (electrical, HVAC, plumbing, lineman, heavy equipment, welding, fire safety, low voltage).
Choose a credential path with clear milestones (pre-apprenticeship, apprenticeship, licensing requirements).
Start building a simple portfolio: photos of projects, safety certifications, tool competencies, and references.
Learn the "AI admin" side of the trade: estimating templates, job planning, documentation, customer messaging, and invoice clarity.
A Note on Interpreting the Research Responsibly
The Microsoft Research paper is not a "jobs lost" scoreboard. It measures how applicable generative AI is to work activities in each occupation based on observed usage and AI performance, then maps that to occupations. It is a valuable lens because it is grounded in real-world interaction data, but it is not destiny. The real takeaway is this:
AI is accelerating information work quickly, while the physical world still runs on skilled humans—and the economy will need more of them as infrastructure demands grow.
Source: paper methodology and findings. Microsoft Research publication and arXiv preprint
Where CraftPath Fits in This 2025 Wrapped Moment
CraftPath exists for the part of the future that still has to be built by real people. Our goal is to make it easier to discover trade paths, compare programs, and choose training that leads to real, local opportunity.
If 2025 taught us anything, it is that the most important jobs of the next decade will blend human skill with smarter tools—and the trades are at the center of that story.
Happy New Year. If you are betting on building, fixing, and protecting the real world in 2026, you are not behind—you are early.
Ready to explore career paths that offer real security and opportunity? Discover your ideal trade career with CraftPath and join a field where human expertise remains irreplaceable.
Sources: Microsoft Research - Working with AI | arXiv Preprint | PDF