1. Start from the role you have
Most AI demand is not for new AI jobs. 95% of AI-related postings are AI-enabled: existing roles where AI is part of the tools, context or skills. 41 of 44 roles are mostly AI-enabled, so the shortest way in is usually the job you already know.
Each role guide shows how AI is changing that role and the direction employers are taking it. Find your role.
Which jobs are becoming AI-enabled? →2. Learn the AI skills every field asks for
The AI skills in the top ten of most roles, whatever the field. They are about using AI well rather than building it.
The gap employers flag most is Prompt Engineering: it is among the top three gaps in 26 of 44 roles.
What AI skills are employers looking for? →3. Add the AI skills of your field
Beyond the common skills, each field asks for its own. The one asked for most in each:
Engineering & architecture
AI Assisted Coding
53% of the field's postings
Data & AI
Machine Learning
16% of the field's postings
Product & design
AI Tools Literacy
59% of the field's postings
Sales & customer success
AI Tools Literacy
35% of the field's postings
Marketing
Prompt Engineering
91% of the field's postings
Management & operations
AI Tools Literacy
63% of the field's postings
Finance, legal & risk
AI Tools Literacy
64% of the field's postings
4. Decide whether you want to code
You do not have to. 27 of 44 roles ask for AI skills but hardly any coding, and they hold 65% of AI-related postings. The 13 roles that do ask for it are engineering and data roles; in Core AI roles such as AI Engineer and Machine Learning Engineer, Python appears in 56% of postings.
5. If you want a Core AI role, follow a path
The way into Core AI is usually a role that already leans that way. These are moves employers are already making, with the skills that separate each pair of roles.
6. Find the jobs, and see where they are heading
Look at open AI roles in your market, and at how the work is shifting before you commit to a direction.
FAQ
How do I start a career in AI?
Start from the job you have. 95% of AI-related postings are AI-enabled versions of existing roles, so for most people the way in is adding AI to their current work: first AI Tools Literacy and Prompt Engineering, which most roles ask for, then the AI skills of their own field. You only need to code for the 13 roles that ask for it, such as Software Engineer, Data Engineer and Data Analyst.
Can I start an AI career without coding?
Yes. 27 of the 44 roles we track ask for AI skills but hardly any coding, and they hold 65% of AI-related postings, led by Project Manager, Account Manager and Business Development Manager.
How do I move into a Core AI role?
Through a role that already leans that way. The moves employers are already making include Software Engineer to AI Engineer (96% of Software Engineer postings already point toward AI engineering) and Data Analyst to Data Scientist (99% of Data Analyst postings already point toward data & machine learning). Each career path lists the skills that separate the two roles.
See the detail for your role
Every role guide shows its essential skills, the AI skills to learn next, the common gaps and where the role is heading.
Browse all 44 role guidesBuilt from 57,486 public job descriptions across 44 roles, published from March to September 2026 and deduplicated so that a job listed on several boards or by several agencies counts once. Every figure here comes from the guide it links to, where the method is set out in full. How we collect and deduplicate postings.