The core skills
- Prompting with context, constraints, and examples
- Checking AI output for facts, gaps, bias, and tone
- Workflow automation with spreadsheets, forms, docs, and task tools
- Data judgement: knowing what the numbers can and cannot prove
- Domain expertise that lets you spot nonsense quickly
- Communication: turning AI-assisted work into decisions people trust
A practical learning order
- Week 1: use AI for summaries, checklists, and first drafts only.
- Week 2: build one repeatable workflow for your current job.
- Week 3: learn how to verify sources, numbers, and claims.
- Week 4: document one example where AI helped you produce a better outcome.
- Month 2: learn the specialist tool most relevant to your profession.
How to use this
Look at tasks, not just job titles. The same job can be safer or riskier depending on whether the worker owns judgement, customers, compliance, physical delivery, quality, or business outcomes.
A useful career move is rarely "learn AI" in the abstract. It is learning how AI changes your actual workflow, then moving toward the parts of the work where trust, responsibility, and context matter.
Sources and context
- World Economic Forum Future of Jobs 2025
- UK Government: impact of AI on UK jobs and training
- UK Government: assessment of AI capabilities and the labour market
- UK Government: AI skills labour market projections
- Anthropic Economic Index
- Anthropic Economic Index: economic primitives
- Office for National Statistics automation analysis
- OECD human-centred AI adoption in the world of work