In January 2025 the World Economic Forum published a report on the future of the labour market. By 2030, 170 million new jobs will appear worldwide, 92 million existing ones will disappear, and the net gain will be 78 million. AI and big data were named the fastest-growing skill, and 77% of employers surveyed plan to retrain staff to work alongside AI.
In the spring of the same year Microsoft reported in its annual Work Trend Index study that 78% of leaders are considering hiring for AI-related roles. The authors also proposed a new concept, the “agent boss”: a person who builds AI agents, sets them tasks and manages their work.
Does this resemble the profession of “AI operator”? At OM-Digital we work this way every day: everyday tasks are done by AI agents, while people set tasks and accept the result. Below are different points of view on the question and what anyone can learn from our experience.
Short answer
Most likely, both. Separate AI specialisations have already appeared in job listings, and the ability to set tasks for AI and check the result is gradually becoming part of most professions. You can learn it without programming and in Russian, and now is a good time: the market is only taking shape.
What this work is called in job listings
“AI operator” is a colloquial expression; hh.ru analytics has no such title. In January 2026 hh.ru experts identified seven in-demand AI-related specialisations: ML engineer, AI engineer, AI solutions architect, AI trainer, prompt engineer, neuro-creator and AI facilitator. The last four are closest to an “operator”: these people do not build models but work with them — train them, set tasks, create texts and images with their help and bring AI into teams’ work.
A separate profession or a general skill
Job listings speak for the first point of view. The new specialisations on hh.ru and the intentions of leaders from the Microsoft study show that companies need people for whom working with AI is their main occupation.
The second point of view: it is a skill that almost everyone needs. Indeed economist Allison Shrivastava put it well in an interview with Fortune in May 2025: the ability to write prompts for AI is “definitely useful to have, but it is not a separate job”. hh.ru data for 2026 show that the market is only starting to move: AI skills are so far required in about 2% of job listings, while the number of CVs with such skills has roughly tripled since 2023. For a job seeker this is an opportunity: the skill is rare and demand is growing.
The state is moving in the same direction. On 26 August 2026 the Ministry of Education announced the professional development programme “Artificial intelligence in a teacher’s work”: by the new school year 250,000 teachers, almost a quarter of all teachers in the country, will complete it. This is an example of how working with AI enters an ordinary profession.
Those choosing a direction will find the Directory of Professions useful. Under Government Resolution No. 138 of 11 February 2025 it became part of the “Work of Russia” digital platform, it is coordinated by the Ministry of Labour, and the directory collects descriptions of professions, expected salaries, universities and demand.
Both points of view agree on the main thing: the ability to work with AI will be useful to almost everyone, and some people will make it their main job.
What vibe coding is
Vibe coding is a way of creating programs by describing the task to AI in plain words instead of writing the code yourself. The term was proposed on 2 February 2025 by researcher Andrej Karpathy, one of the founders of OpenAI. In November of the same year the Collins dictionary named “vibe coding” its word of the year.
In practice it looks like a conversation with a colleague: a person explains what they need, looks at the result and says what to change. The code is written by AI.
Can you work with AI without English and without code
In our experience, yes. All our work with AI is in Russian: of 71 working instructions for AI agents, 65 are written in Russian, and all the tasks in the company queue as of 18 October 2026 are also worded in Russian. Tasks are set in a messenger in ordinary phrases, without special syntax: AI understands an imprecise wording by its meaning, and if there are several options, it asks a clarifying question.
English appears in the names of technologies and in programs’ system messages. Here AI translates by itself and explains what happened and what to do next.
You do not need to write code, but it helps to understand what came out. The person who sets tasks for AI is responsible for four things:
- setting the task: the goal, the constraints and an example of what the finished result looks like;
- acceptance: open, tap, check against the task before showing it to others;
- security: access rights, keys and personal data are stored on the server, not in the app — this is a basic rule in our projects;
- support: after launch the product has a person who is responsible for it.
Complex systems with a high cost of error, for example in banking, still need engineers. Vibe coding shows itself best on prototypes, internal tools and small products.
What to learn to get started
Your original education hardly matters here. Four habits matter more:
- describe the task in words: what is needed, for whom and how to tell it is done;
- work in iterations: first version, feedback, fix;
- check the facts and figures that AI produces;
- keep successful instructions in files so that AI does not start from scratch every time.
Our whole system of work grew out of such instructions.
Key takeaways
- According to WEF, the world will gain 78 million jobs by 2030, and AI and big data are the fastest-growing skill.
- Job listings already include separate AI specialisations, and “AI operator” is their common colloquial name.
- Working with AI is becoming a profession for some and a required skill for most: the Ministry of Education is teaching it to 250,000 teachers.
- Vibe coding lets you create programs by explaining the task in plain words, including in Russian.
- Setting the task, acceptance, security and support stay with the person.