Prompt engineer is no longer a profession. Who are companies implementing AI looking for today?
Companies implementing AI are no longer looking for people to write prompts, but for people responsible for the entire system. Six of the 15 roles identified in an analysis of job postings from 2026 do not require daily programming.
Until recently, a prompt engineer, a specialist in writing instructions for language models, was considered a profession of the future. Today, it is hard to find in job ads. The editors of My Company Polska analyzed job offers published from January 1 to October 8, 2026 and identified 15 professions in which AI is a significant part of the responsibilities. There is no prompt engineer among them (opens in a new tab).
For a manager planning an AI implementation, this is a practical clue. When building a team, one “ChatGPT person” is not enough. You need people who will take responsibility for data, the product, implementation, security, and oversight.
The fad for the profession has passed, the skill has remained
My Company Polska's analysis covered offers from the five most popular Polish recruitment websites according to Similarweb's September ranking (Pracuj.pl, GoWork.pl, Jooble.org, eRecruiter.pl, Praca.pl), as well as LinkedIn Jobs, Indeed, and Wellfound. Only offers in which AI was a significant part of the job were counted, and a mere mention of ChatGPT was not enough.
This is not an isolated observation. Computerworld Polska wrote in May 2025 that the number of profiles with the title Prompt Engineer on LinkedIn had fallen by 40 percent over nine months. At the same time, interest in prompt engineering courses was growing: Vanderbilt University's course on Coursera had attracted more than 448,000 enrollments by May 2025. After analyzing 202 Polish job ads with salary ranges, the blog promptowy.com concluded in September 2026 that the standalone prompt engineer position had practically disappeared, and prompt writing had become part of other roles.
Putting these data points together leads to a simple conclusion: demand for the skill is growing, while demand for a separate full-time role is disappearing. A company that creates a “prompt engineer” position today is buying a competency that every employee using AI should have.
From experiments to implementations: who has the most job offers
The largest number of ads in the list, at least 25, concerns the AI / GenAI engineer. This is a person who builds applications based on language models, for example a company assistant answering questions based on documents, and is responsible for the entire process: from selecting the model, through connecting it to company data, to testing and launch. The authors of the analysis read this as a signal that companies are moving from testing AI to solutions used every day.
A characteristic role of this year is the AI automation and agents engineer. They build programs that independently read a document, retrieve data from other systems, launch subsequent process steps, and hand the decision over to a human when needed. Employers expect Python, connecting systems via APIs (interfaces that allow programs to exchange data), and an understanding of processes. The threshold is lowered by low-code tools, where most of the work is done without writing code, such as Copilot Studio and Power Platform.
Six roles for people from business
The report divides the market into three paths: those who create AI (where advanced programming is essential), those who combine technology with business, and those who apply AI in business. In this third group are the AI product manager, AI adoption leader, AI business analyst, AI project manager, AI director, and AI Governance specialist. Coding takes a back seat there.
This is good news for companies that do not have the budget for a team of engineers. Some of the necessary competencies are already in the organization; they just need to be supplemented with practical knowledge of AI. The analysis points to the following transitions:
The AI manager or product owner is the second role in terms of number of offers, with at least 15 ads. Programming is usually not required here. Employers do, however, want the candidate to understand AI's capabilities, limitations, and costs, and to be able to assess whether the model responds accurately even in unusual situations.
From a manager's perspective, the most interesting role is the AI adoption and implementation leader. Their task: make sure employees actually use the tools and that the company does not stop at buying licenses. They identify processes in which AI delivers results, run workshops, set usage rules, and check outcomes. The analysis found at least 12 such offers, with rates of PLN 140-190 per hour on a B2B basis, that is, in cooperation with a sole proprietorship.
Oversight and security are no longer an add-on
The more AI there is in a company, the more new risks arise. The AI Governance specialist creates AI usage policies, maintains a register of systems, assesses risk, and detects Shadow AI, meaning tools employees use without the company's knowledge. The ads mention the NIST AI RMF and ISO/IEC 27001 standards as well as DLP tools that prevent data leaks. There were at least 10 offers, but companies did not provide salaries in them, which is why the role closes the ranking.
The AI security engineer protects systems, among other things, against prompt injection, meaning deceptive instructions that take control of the model, and against misuse of tools by an AI agent. Companies are looking here for people with experience in cybersecurity and the cloud.
How much these roles cost and what kind of market recruitment takes place in
Salary data should be read as a range, not a single number. My Company Polska gives up to PLN 37,000 per month for an AI architect and PLN 24,000-32,000 on a B2B basis for a senior AI automation specialist, while a regular position in this group is PLN 90-130 per hour. On its sample, Promptowy.com calculated a median of around PLN 26,000: around PLN 21,500 for the mid level, PLN 26,145 for a senior, and PLN 30,745 for an expert. The differences may result from different role definitions and methods of collecting offers.
The backdrop is less rosy. According to Statistics Poland (GUS), average pay in companies employing at least 10 people amounted to PLN 9,266.48 in September, and employment fell by 0.8 percent year on year, to around 6.4 million jobs.
Although investments are gaining momentum in the economy, which brings with it increased demand for employees in some industries, at the same time jobs are being reduced or eliminated (...). Investments in automation, new technologies, or AI are also playing their part.
Juxtaposing these two pictures is our conclusion, not a finding of any of the sources: recruitment for AI roles is taking place in a market that is shrinking overall. However, these phenomena should not be linked causally: analysts at ING Bank Śląski note that the decline in employment in enterprises mainly reflects structural factors, including the fact that people retiring are not being fully replaced.
What this means for a company planning an implementation
Before announcing recruitment, it is necessary to determine who is responsible for tool adoption, who for security, and who for evaluating results. Some of these tasks can be taken over by people from HR, compliance, or product management if they are given time and training. Promptowy.com notes that employers value documented implementations more highly than certificates - this is also a good measure for internal promotions.
However, the safest conclusion does not depend on the exact numbers. Writing prompts for AI is now a basic skill of the whole team, not a reason to create a separate position.