Guide · Skills
The skills to learn for an AI future
A practical list of the skills that matter as AI changes work, from using AI well to the human skills it makes more valuable, and how to start learning them.
30 October 2026 · 3 min read
“Learn to code” was the default career advice for a decade. Now AI writes a lot of the code. So what should you actually learn?
The honest answer is a mix: skills for working with AI, and skills that become more valuable because of AI. Here is a practical list, roughly in the order most people should tackle it.
1. Using AI tools on your real work
The single most useful skill right now is simply getting good at using AI for the tasks you already do. Not in theory, but on your actual emails, reports, spreadsheets, code or lesson plans.
That means learning to:
- Brief the tool clearly: context, audience, format, examples of what good looks like.
- Iterate rather than accepting the first answer.
- Check the output, because AI can be confidently wrong.
- Know when not to use it, for sensitive data or decisions that need a human.
This is less about clever prompts and more about habit. People who use AI daily on real work learn its strengths and limits fast.
How to start: pick the most repetitive task in your week and try doing it with an AI assistant every time for a month. Our 30-day plan for beginners walks through it.
2. Judgement and verification
As AI produces more first drafts, the scarce skill becomes judging them. Is this accurate? Is it right for this audience? What is missing? Being the person who catches the error before it ships is valuable in every field.
How to build it: go deeper into your own domain. The better you understand your subject, the better you can supervise AI working in it.
3. Problem framing
AI is good at answering questions and weaker at knowing which question to ask. Defining the real problem, breaking it into pieces and deciding what “done” looks like are skills that transfer across every job.
How to build it: before starting any task, write one sentence on what problem it solves and how you will know it worked.
4. Communication and persuasion
When information is cheap, getting people to act on it is not. Explaining a recommendation clearly, writing a crisp summary, presenting to decision makers and handling objections all grow in value.
5. Working with people
Coaching, collaborating, negotiating and resolving conflict are among the hardest things to automate. They also matter more as teams reorganise around new tools. If you have never managed anyone, mentoring a junior colleague is a low-risk way to start.
6. Data literacy
You don’t need to become a data scientist. You do need to read a chart critically, ask where a number came from, and spot when a statistic is being misused. AI tools will happily produce analysis; data literacy is how you tell good analysis from bad.
7. Automation basics
Simple no-code automation (connecting your email, calendar, spreadsheets and forms so routine steps happen on their own) is becoming a basic office skill. People who can automate their own busywork free hours every week and often become the go-to person on their team.
8. A deep specialism
General knowledge is what AI is best at. Deep, specific expertise in a domain, a regulation, a customer group or a craft is harder to replicate. Combining a specialism with AI fluency is one of the strongest positions you can be in.
How much time does it take?
Less than you might think. A few focused hours a month is enough to pull ahead of most people, because most people are not doing it at all. In the AI Survival Score, learning habits are one of the five factors, and moving from “no time for learning” to a few hours a month is one of the quickest ways to raise your score.
Make it personal
The right skills depend on the tasks that fill your week. Take the test to get a personalised “what to learn next” list based on your own most exposed tasks, or browse the skills for each task.
Keep reading
What's your AI Survival Score?
12 questions, about 2 minutes, no sign-up.
Take the test