I asked ChatGPT to plan my career switch. It gave me these 5 options

I gave ChatGPT a detailed profile of my professional experience, skills, interests and career preferences. I asked it to first identify my transferable skills, then suggest career paths that could make use of them and guide me on what I would need to learn before making such a transition.

I first asked AI to look beyond my job title

I did not want ChatGPT to treat my experience simply as journalism or editing. I asked it to break down the work I do into transferable skills and identify where those abilities could apply in completely different professions.

It identified skills such as editing and quality control, research, information synthesis, AI tool usage, multi-tasking, prioritisation, deadline management, critical reading, questioning, communication, workflow management and adaptability.

It then grouped these skills into five broader areas:

• Turning information into understanding
• Turning that understanding into an output
• Managing workload and execution
• Solving problems
• Working with people and information

ChatGPT suggested me these 5 career paths

I then asked ChatGPT to suggest five career paths that were genuinely different from journalism but where I would not have to start completely from zero.

It suggested:

• Business Analyst: Research, information synthesis and the ability to work with incomplete information could transfer to business analysis. The biggest gaps would be advanced Excel, SQL, dashboards, Power BI, requirements gathering and process mapping. ChatGPT estimated a four-to-eight-month transition and a Rs 7–12 lakh annual target for a switcher.
• UX Researcher/User Researcher: Questioning, research, listening and pattern recognition could be useful for user interviews and qualitative research. I would need to learn UX research methods, usability testing, research planning and product thinking. It estimated five to nine months and a Rs 7–14 lakh target.
• Operations Manager/Business Operations: Multi-tasking, prioritisation, deadline management and workflow experience could transfer into business operations. The gaps included process mapping, KPIs, Excel or Sheets, workflow tools and stakeholder management. The estimated transition was three to six months, with a Rs 6–12 lakh target.
• Customer Success Manager: Communication, organisation, research and problem-solving could transfer to customer-facing product roles. I would need SaaS and product knowledge, CRM tools, onboarding and account-management skills. ChatGPT estimated two to five months and a Rs 7–14 lakh target.
• AI Trainer/AI Evaluation Specialist: Editing, research, fact-checking and attention to detail could help assess whether AI-generated responses are accurate, useful, well-written and logically sound. The gaps included AI evaluation frameworks, annotation, model behaviour, evaluation metrics and basic AI/ML concepts. ChatGPT estimated two to six months for entry-level roles and a broad Rs 5–10 lakh annual target.

ChatGPT also gave me a warning about those salary numbers. Professional experience does not automatically translate into equivalent seniority in a new function. Someone with several years of work experience can still be considered relatively new to business analysis, UX research or AI evaluation if they cannot demonstrate the required skills.

Of the five, AI evaluation looked like the most unusual option. It also seemed to have the strongest connection with my existing AI-tool experience and editorial judgment. So I decided to test that recommendation more aggressively.

Could my existing skills really help me make the switch?

I asked ChatGPT to stop encouraging me and instead act as a devil’s advocate. Its first warning was that “using AI is not the same as evaluating AI”.

An AI evaluation job could involve judging model responses against defined criteria, identifying subtle factual or logical errors, following detailed annotation guidelines, comparing multiple outputs and writing precise reasons for a rating. It could also involve recognising hallucinations, instruction-following failures and reasoning problems.

My editorial judgment could transfer, but ChatGPT said an employer could still ask where the evidence was that I could evaluate AI systems systematically.

The second problem was the job market itself. Titles such as AI Trainer, AI Evaluator, AI Data Trainer, LLM Evaluator, AI Quality Analyst, Prompt Evaluator and Model Quality Analyst do not necessarily describe the same kind of work. Some could be specialised positions inside AI companies, while others could be project-based annotation or evaluation jobs.

That created a question I had not considered. Was I entering an emerging professional field or a collection of relatively low-barrier contract jobs?

ChatGPT’s third warning was about the high technicality. A person could potentially enter AI evaluation without becoming a machine-learning engineer. But moving into higher-value roles could require knowledge of evaluation frameworks, annotation schemas, benchmark design, model metrics, error taxonomies, data quality, experimentation and eventually Python or SQL.

Its assessment was that my journalism background could get me interested in AI evaluation, but it would not automatically make me employable in the higher-value parts of the field.

What would I need to learn?

Instead of telling me to collect AI certificates, ChatGPT suggested building evidence that I could actually evaluate AI systems. It proposed a six-month transition, with the first 90 days focused on learning the profession, building a portfolio and developing basic technical skills.

During the first month, I would learn practical LLM fundamentals, evaluation criteria, annotation principles and common model failures. I would create an evaluation rubric and test it on 30–50 AI responses.

The second phase would add advanced prompting, evaluation experiments and a technical layer. ChatGPT suggested basic Python, data analysis and more structured testing of AI outputs.

By Day 90, the goal would be to have three portfolio projects, foundational technical skills, a visible professional presence and initial conversations with people already working in the field.

It also gave me a scorecard:

• Day 30: Evaluate 30–50 AI responses and complete the first portfolio project.
• Day 60: Reach 75–100 evaluated responses and have two portfolio projects.
• Day 90: Reach 150–250+ evaluated responses and complete three portfolio projects.
• By Day 90: Publish at least five useful pieces of AI-related work.
• Networking: Have 15–20 relevant professional conversations.
• Job-market research: Analyse at least 40 relevant job descriptions.

The numbers were not meant to be rigid quotas. The idea was to measure whether I was actually becoming employable rather than simply spending hours watching AI courses. 

What would happen after 90 days?

ChatGPT suggested using the remaining three months of the six-month transition for market testing. Here’s what it suggested:

Days 91-120: Test the market

• Start applying selectively to relevant AI Trainer, AI Evaluator, LLM Evaluator and AI Quality roles.
• Analyse 20–30 additional job descriptions.
• Track rejections and identify recurring skill gaps.
• Start preparing for interviews.

Days 121–150: Close the gaps

• Focus on the skills that repeatedly appear in job descriptions—such as Python, SQL, annotation tools or LLM evaluation frameworks.
• Continue applying while learning.
• Improve your portfolio based on employer and recruiter feedback.

Days 151–180: Push for the transition

• Shift most of the time towards applications, interviews and networking.
• Keep improving your portfolio rather than starting completely new projects.
• Reduce introductory AI learning and focus only on skills directly relevant to your target roles.

What would I say in an interview?

I also asked ChatGPT to help me explain a career transition without presenting my existing experience as irrelevant.

Its suggested approach was to connect editorial work with the new field. Research, critical thinking, quality assessment, identifying inaccuracies and evaluating whether an answer actually addresses a question are all skills that can have relevance in AI evaluation.

The important part was not to present myself as an AI specialist already. The suggested positioning was that I was building technical and evaluation skills on top of an existing professional foundation.

That distinction matters in a career switch. The new role may be different, but the experience accumulated in the old one does not have to become irrelevant.

The warning I did not expect

Finally, I asked ChatGPT to step back from the entire exercise and tell me the one thing a career counsellor with 20 years of experience would warn me about.

Its answer was: “Don’t confuse a fascinating field with a good career fit.”

AI evaluation appeared attractive because it connected my interest in AI with skills I already had. But the actual job could involve repetitive response grading, rigid annotation guidelines, reviewing hundreds of similar outputs, spreadsheet and data work, repetitive testing and dealing with ambiguous edge cases.

That could mean moving from a career with considerable variety and autonomy into a more process-driven role, possibly with lower seniority and salary at the beginning.

ChatGPT therefore suggested a practical test before making any major decision: simulate the job.

It suggested that I:

• Evaluate 100–200 AI responses.
• Create a rigorous evaluation rubric.
• Document the errors I find.
• Categorise recurring patterns.
• Repeat the exercise over several weeks.
• Then ask myself whether I would want to perform the same work for six hours a day, five days a week.

What worked, what didn’t?

ChatGPT was useful in turning a possible career switch into a clear 90-day roadmap. It gave me a structured way to learn, build a portfolio and test whether a new role could actually suit me. It was also useful in rewriting my LinkedIn profile. 

However, I found some of the career options less convincing. For instance, research is one of my strengths, but that does not necessarily mean I want research to become my career. Business analysis also did not feel like a natural fit for me. I felt ChatGPT sometimes looked at the kind of company and platform I currently work in rather than focusing only on my skills and interests. 

For a career switch, I would want AI to look beyond my current industry and suggest options based on what I can do and what I may enjoy, not where I currently work. Overall, the planning was useful, but the career recommendations still need to become more personalised.

Disclaimer: This is a first-person experiment conducted by the author for informational purposes. Financial Express does not encourage or discourage career changes or the use of AI tools for career planning and is not responsible for any decisions, financial outcomes or professional consequences arising from the use of AI-generated recommendations discussed in this article. The career options, salary estimates, transition timelines, skill assessments and other suggestions generated by AI do not guarantee the success results and should be independently evaluated before being acted upon. Readers should conduct their own research, assess their skills and interests, and seek appropriate professional advice where required.