After reading yet another LinkedIn post about jobs that Artificial Intelligence is making obsolete, I found myself thinking about the paradox of how many existing jobs (particularly the ones I’ve held) can actually foster the development of AI-first professional skills that might be more resilient and perhaps indispensable in the future.
I find this a more constructive approach to the conversation, which is why I decided to write this post. I hope it can serve as a practical guide for anyone who’s skeptical, doubtful, or discouraged about the impact of generative AI on the workplace.
Professional experience shapes AI collaboration
The first time I started using ChatGPT in 2022 (ChatGPT was my gateway drug), I treated it like software and I got stuck immediately. There wasn’t a single task it could perform, it seemed, that I didn’t feel I could do better, faster, and more effectively. To be fair, it was a pretty basic model compared to what we see today, in June 2025.
The first time I took a prompt engineering course in 2023, I began to sense I’d gotten the approach wrong, but I didn’t really have the mental bandwidth to find a place for the tool in my workflow. It wasn’t until 2024, when I took an updated version of the course, that I finally started to internalize this reality: the skills that make Large Language Model applications effective aren’t technical—they’re professional and human.
The moment I started treating ChatGPT, Claude, and Perplexity (the trio of tools I pay for and have integrated into my workflow) as assistants, co-pilots, and quasi-human extensions, my life actually changed. Not because I learned to code, use APIs, or build complex agents. Rather, because I put to work the soft skills I’d developed throughout my career to get the results I needed from generative AI applications.
Here are two concrete examples of the results I mean:
- getting back to writing regularly, including long-form pieces, because I no longer start from a blank page but from I have mirroring conversations that the machine transforms into notes and outlines
- tackling complex projects even when I’m exhausted and can’t concentrate, because the machine helps me break them down into smaller, manageable tasks
A strategic approach, here too
Before detailing the jobs and soft skills, a note about this post’s title. When I write “using strategically,” I mean employing generative AI not as a tech gadget, but programmatically, in daily operations, to amplify my existing skills and support achieving my goals while respecting my standards (quality, values, brand).
This broad-vision approach is reflected in the two practical principles I use to integrate generative AI into my work:
- I only use it to automate and deepen activities that I’d be capable of doing well on my own
- I always review and revise outputs (in fact, I never use them as-is, even when they contain correct information)
But let’s get to the heart of this article (essay?).
10 roles and how they taught me to use generative AI
What follows are some jobs I’ve held or roles I’ve played since I started working consistently in 1998. They’re not organized chronologically but by application areas.
For each, I’ve created a brief profile of what it taught me (that’s relevant for AI-first professional skills), the skills transferred to my use of generative AI, and how I apply them in practice.
1. Content creator and blogger
The first personal brand I developed was born alongside a blog that allowed me to give voice and form to that brand. It took time, lots of practice, trying to repeat the concepts important to me until I found the form that effectively communicated my values and vision. From there, it was mostly editing work, selection, “editing out.”
What it taught me: that brand voice is built through consistency over time via continuous iterations. The first draft is rarely published, but each version can reveal new possibilities.
Skills transferred: recognizing effective communication patterns, being comfortable with the iterative process.
How I apply them with AI: I’ve created projects and styles trained on my writing and thought processes, and I’m not afraid to course-correct and try multiple times to work through reasoning.
2. Mentor to interns
Since 2022, this is a role I often play with people starting internships at Intersezione. In many cases, I guide them through their first agency experience and always try to maintain a healthy balance between training, meaningful experiences, and feedback.
What it taught me: to reverse-engineer processes for activities that now come naturally to me, transforming general titles into clear, simple instructions that can be put into practice. Then to delegate to people who don’t have the full context but still need to produce decent quality results.
Skills transferred: reverse engineering my work to derive instructions, balancing control and autonomy in delegation.
How I apply them with AI: I try not to take anything for granted, providing precise briefs, practical guidance, and sometimes sensory nuances. But I try not to be rigid, leaving space for unexpected solutions that I explore and then validate.
3. Telephone interviewer
If between 1999 and 2000 a woman asked you dozens of questions about your relationship with whiskey over the phone, that might have been me! Market research through interviews (later I also did them in person, sampling or in focus groups) is a tool I find extremely useful, in its improbable balance between statistical vocation and personal sensitivity.
What it taught me: to construct questions that are as neutral as possible, yet natural, to gather precise and reliable information without influencing the subject.
Skills transferred: precision questioning and bias control in responses.
How I apply them with AI: the more uncertain I am about the output I need, the more I formulate prompts like structured interviews. I try to establish neutral context, clean my writing as much as possible of styles and mannerisms, offer clear parameters, and ask specific questions. Then I ask for counterarguments.
4. Marketing consultant
Since I started focusing on what comes before promotion and communication at work, helping companies (re)build the marketing foundations to develop everything else, I’ve consulted for freelancers and large companies.
What it taught me: that solid, lasting results are only achieved by considering goals, targets, resources, and constraints.
Skill transferred: systemic strategic thinking.
How I apply it with AI: I feed every prompt with its strategic context. I always explain what my objective is, how I expect the LLM to behave, and who I am to it in that moment. Just for starters.
5. Photo producer
Over the years, I’ve produced events and photo shoots. In some ways, the goal (sometimes explicit) is always for images to capture an identity. But around that is all the complexity of real life, people, weather conditions, and so on.
What it taught me: to break down complex projects into manageable, coordinated phases in service of a vision.
Skill transferred: ability to sequence complex tasks.
How I apply it with AI: it’s rare that my activity starts and ends within the same program. More often I work in phases and use outputs to feed other models or other phases.
6. Operations manager
Being an organized person who functions best in organized contexts means that wherever I work, I end up working on processes. I’ve managed quality plans and procedures for a finance help desk and organized an integrated editorial team of over sixty creators, among other things.
What it taught me: to identify bottlenecks, optimize workflows, and systematically measure improvements.
Skill transferred: the mindset of systemic efficiency (not continuous improvement, because that’s not efficient 😉).
How I apply it with AI: I treat AI integration into my workflow like any other process. I monitor which prompts work best (I save them), create reusable templates, automate repetitive parts of my workflow.
7. Italian as L2 teacher
I’ve taught Italian to 20-year-olds and 60-year-olds. I haven’t always left a mark, but I’ve always sought to bring language closer to the person, rather than the reverse.
What it taught me: to adapt language methods and structures to the receiver’s level, provide concrete examples and constructive feedback.
Skill transferred: adaptive communication.
How I apply it with AI: when ChatGPT makes up statistics or Claude tells me it updated a document even though it didn’t, I don’t get angry, don’t waste prompts starting useless philosophical conversations about trust and reliability. Instead, I ask for verification, guide reconstruction, and provide examples of the level of complexity and accuracy I expect.
8. Business strategist
In some cases, working backward with clients reaches all the way to the business model, and I find myself supporting companies in existential reflections that are new to them. It’s a phase I love because I firmly believe in the connection between identity and purpose, and that values, mission, and vision must find expression first and foremost in company processes.
What it taught me: to ask big questions about “why” and “to what end?”, find answers, and then use them as a guide for everything else.
Skill transferred: clarity about the centrality of purpose.
How I apply it with AI: I break down ChatGPT’s sycophantic responses and always try to ground context in high-level objectives and purposes.
9. Seasonal farm worker
I started working in the fields at 15, picking apples and pears during summer. I’d take the 6 AM bus, eat lunch (sandwiches I brought from home) alone sitting in a concrete shed, splash my head under the fountain, and climb onto the cart at 2:30 PM with a soaked bandana on my head, hoping it would ease the heat. At the end of the day, my skin smelled sweet and ripe like fruit. But the next summer I’d take a trip with the money I’d earned.
What it taught me: that results typically require methodical processes and patience for natural timing. And that processes sometimes stink or are uncomfortable, even when necessary.
Skills transferred: operational resilience and results orientation.
How I apply them with AI: I don’t give up when a prompt doesn’t work, I go back and recover part of the work myself, then return trying to improve the prompt. Most importantly, I stay consistent in my experiments.
10. Freelancer and serial learner
I’ve been freelancing since 2009, and this condition inherently requires making adaptability an innate virtue. But my self-teaching spirit goes back much further and I don’t think it will ever leave me. The techniques and theories I’ve learned on my own, from books and trial and error, are countless.
What it taught me: that everything constantly changes, but there are tools, processes, theories capable of training us for change. And that you don’t need to wait for other people’s permission to learn them.
Skills transferred: autonomous experimentation and self-training capacity.
How I apply them with AI: I constantly adapt my approach based on results, learn from various sources but then develop my method through practice and experimentation. Most importantly, I still see generative AI as a territory to explore with curiosity and caution, not something to fear.
The meta-lessons that change everything
I could have stopped here. I think just the list above, with the formula traditional profession > soft skill > AI management competency, is eloquent and useful in helping you modify your approach to the subject and maybe implement an “AI-first” work style.
But as I was writing, I realized the enumeration was also a timely and clear demonstration of some fundamental cross-cutting considerations:
- generative AI amplifies human skills, it doesn’t replace them. This changes the approach to the tool
- value always lies in human validation, not in AI-generated output. How you decide what to keep, what to modify, what to discard. The role of professional taste
- the ability to collaborate with an LLM is a professional soft skill, not a technical skill. There are tech-savvy people who can’t maximize the potential. Those with more varied experiences often use them better
What now?
How do you turn all these considerations into practice? I’ve taken the liberty of starting the distillation myself, but I expect you’ll find many other insights!
If you have less than 3 years of work experience: experiment a lot. Focus especially on comparing the results you get by working on prompts, iteration by iteration, and practice fact-checking.
If you have 3-10 years of experience: you can rely on decent domain experience, so start listing the tasks you perform in your work and identify those you can delegate at least partially to AI. Use your skills to produce high-quality prompts.
If you have 10+ years of experience: your superpower is your “expert intuition”, the ability to instantly identify inconsistencies even in complex patterns because you’ve internalized them through decades of work. Use it to provide complex contexts to your AI assistant and then validate its outputs.
The point I wanted to make is that the substantial changes to how we’ll live and work, caused by the spread of technology like Artificial Intelligence, don’t make professional skills obsolete. Some specific jobs might disappear (like coachmen when cars arrived), but the professional skills you’ve developed make you perfectly capable of finding space and purpose in a world where AI is everywhere.
Cover image by Lindsay Henwood/Unsplash.
