Teams
Prompt Engineering Went Into Every Job, Not Away
The prompt engineer title mostly did not survive, but the work did not go anywhere. It moved into roles that already existed, analyst, support, marketing, engineering, as one requirement among several rather than a standalone job. Build the skill inside a role you actually want, rather than training for the title as a destination. What lasts is knowing your field well enough to specify a task and tell when the answer is wrong; a new model release can flatten a phrasing trick overnight.
The takeCourse sellers quoting a peak salary screenshot and creators declaring the job dead are answering the same wrong question from opposite directions: whether the title survives. Neither side wants to say the more useful thing, which is that a title disappearing is not evidence the underlying work disappeared with it. That distinction is worth more to a career-changer weighing a bootcamp fee than either the hype or the backlash, and it costs nothing to check against actual postings before spending money on either story.
Where Olive fits
Open a role and see what the work shows
The dimensions Olive scores, framing a problem before generating, sourcing the evidence that matters, keeping the judgment you should not hand to a model, verifying against something outside the conversation, are a usable study list for exactly this kind of durable skill, whether or not you ever sit the assessment.
Rank your shortlistWhere Did the Prompt Engineer Job Go?
The prompt engineer job moved into other titles' requirements. The count of distinct US job titles carrying an AI-related term rose from 264 in the first quarter of 2022 to 822 in the first quarter of 2026, and most of those titles now sit outside tech occupations entirely 1. That counts title names, not openings. A support role, a marketing role, or an operations role is increasingly the one asking for the skill that used to justify a standalone title.
The sector data tells the same story from a different angle. As of December 2025, roughly 9% of human resources postings and 15% of marketing postings already carried AI-related terms, well below the near-45% seen in data and analytics roles 2. A mention is only a mention, and does not establish that the role tests for the skill. None of these are prompt engineering postings by name. They are ordinary jobs that absorbed a piece of what that title used to mean, usually as a bullet point several lines down the description rather than as the headline requirement.
That is worth saying plainly because it cuts against both stories circulating about this career. It is not evidence the field collapsed, since the underlying demand for the skill kept growing across a widening set of occupations. It is also not evidence the standalone job is coming back, since the growth is happening inside other titles rather than under its own name. A career-changer deciding where to spend the next few months needs the second reading, not the first, because it points at a different set of postings to actually search: not "prompt engineer" as a job title, but the operations, support and analyst postings in a field you already care about that happen to name this skill as one line among several.
Build the Skill Inside a Role, Not as a Destination
Do not train directly for the title itself. Instead, check where the underlying tasks, specifying what you want from a model, judging whether the output is usable, iterating when it is not, show up in postings for work you would want regardless of whether AI was mentioned at all. Build toward that role, and treat this as one competency inside it rather than the whole résumé.
That also settles how to judge a course before paying for it. Employers reading which roles actually need AI skills are told that course completion is an attendance record, and that what shows capability is a short task on the company's own material with a rubric written beforehand. Read that from your side of the table. A course worth the money leaves you with a piece of work in your own field you can point to afterward, and one that leaves you only with a certificate has not done its job, whatever the title on the syllabus.
This is also the more durable read of the labor-market data. The Burning Glass Institute found that skills exposed to automation were 16% more likely than baseline skills to see posting demand decline, while skills exposed to augmentation were 7% more likely to see demand rise, and that the occupations experiencing the most automation are experiencing the most augmentation at the same time 3. Those are relative likelihoods against a baseline group, not the size of any change. The shift they describe is inside roles, not between a role and unemployment, and that distinction should change what a career-changer actually shops for.
A course whose module names all still use the words prompt engineer is describing a shrinking category. A course built around a specific field, healthcare operations, financial analysis, customer support, that happens to teach the same underlying skill is describing a growing one. If the honest answer to what you will have made by the end is a certificate and some practice prompts, keep looking.
What Actually Doesn't Decay
What does not decay is knowing your domain well enough to specify a task precisely and to recognize when an answer is wrong. Phrasing tricks are the part that decays, because the next model release routinely and quietly absorbs whatever workaround made an earlier one behave, while domain knowledge takes as long to build as it always did.
Whether the pay premium some course marketing cites is real gets the same honest answer. Postings that mention AI skills advertise more, roughly 28% higher, about $18,000 a year, and in 2024 over half of the postings requesting AI skills sat outside IT and computer science 4. That is a raw gap between two groups of postings, which skew senior and toward higher-paying industries, so it describes what employers advertise rather than what anyone is paid. Whether that premium would actually apply to you depends far more on your field and your evidence than on whether you hold a credential with "prompt engineer" in the name.
A candidate who can specify a task precisely in, say, contract review or clinical scheduling, and who can catch a confidently wrong answer in that domain, is harder to replace with a newer model than one whose only asset is a general phrasing technique. The domain knowledge is the part a course cannot shortcut, the same way spreadsheet fluency is one line on an analyst's resume rather than the whole job description, and it is the part worth spending the actual months on. The salary screenshot, the countdown timer on the enrollment page and the testimonial from someone whose field is not yours are all arguments about the title. The title is the part that already moved.
Common questions
Should I still list "prompt engineering" as a skill on my resume?
Yes, when a named task sits behind it. As a bare skill line it now reads as dated; as one line in a description of a real piece of work, it still communicates something specific. The word matters less than the evidence behind it.
Is a prompt engineering bootcamp ever worth the money?
Occasionally, if it is attached to a domain you already work in and produces something you can show afterward. A generic bootcamp that teaches phrasing techniques in isolation is the weaker bet, because those techniques are exactly the part that ages fastest as models change.
Which roles are actually absorbing this work?
Support, marketing, operations, analysis and engineering roles are the ones showing up most in the posting data, not a new standalone job category. The pattern is broad rather than concentrated in one destination role, which is itself the reason to build the skill inside your target field.
What should I actually practice if I want to be strong at this?
Practice on real tasks in your own domain: specify what you need, check the output against something outside the conversation, and note what you had to fix. That habit transfers across model releases in a way that a list of prompting formulas does not.
Does this mean the salary figures from 2023 and 2024 were fake?
Some were real for a narrow set of early roles at a small number of companies, and they were never representative of a durable job category at that scale. Quoting an old peak figure today, without naming when and where it applied, is marketing rather than evidence.
References
- 1. AI Is No Longer Just a Tech Occupation Story: It's Spreading Across Job Titles in the US and Europe hiringlab.indeed.com Supports that AI-touched job titles have grown fastest outside tech occupations, evidence the work spread rather than vanished.
- 2. January 2026 US Labor Market Update: Jobs Mentioning AI Are Growing Amid Broader Hiring Weakness hiringlab.indeed.com Supports the sector breakdown showing AI language reaching HR and marketing postings, sized honestly against the technical-sector baseline.
- 3. Beyond the Binary: How Automation and Augmentation Are Combining to Reshape Work burningglassinstitute.org Supports that automation and augmentation demand shift inside the same occupations rather than one destroying a role that the other creates.
- 4. Beyond the Buzz: Developing the AI Skills Employers Actually Need lightcast.io Supports the wage premium figure attached to postings naming AI skills, and that it already reaches roles outside IT and computer science.
4 sources, numbered by first appearance. How Olive sources claims
General guidance for hiring teams. What works at one company and one volume may not transfer to yours.
Olive assesses how a person works with AI. It does not detect AI-written documents, and it never produces a score, a ranking, or a match percentage for a person. Candidates read the same report the employer reads.