Imagine the future of your job. Now imagine how, enabled by generative artificial intelligence, you can destroy it.
If this sounds terrifying, it shouldn’t. Despite predictions to the contrary, so far AI hasn’t produced a mass destruction of jobs. PwC’s 2025 Global AI Jobs Barometer, which analyzes nearly one billion jobs ads and thousands of company financial reports, showed that job numbers and wages are growing in nearly every AI-exposed occupation. The wage premium for people with AI skills is also climbing; on average, it stands at 56%.
That doesn’t mean AI won’t fundamentally change your job and your organization. It will. It’s essential to identify what parts of your work the technology will substitute, what it will complement and what it will change.
Begin by considering the different component parts of your work, which likely are many. “Our jobs consist of many tasks. My job as a professor is not just one thing,” says Ethan Mollick, the Wharton professor and bestselling author. “AI doing one or more of these tasks does not replace my entire job, it shifts what I do.”
I call this hacking yourself, but really it’s a radical reimagining of your work. The mandate is to:
- Stop doing things. Think about the parts of your job that artificial intelligence can do as well or better than you can. You can eventually stop doing these things because the technology works as your helpful assistant. What are the pain points during your workday?
- Do things better. Now think of when AI can act as a natural complement to tasks you’ll continue to do – but you’ll do them better thanks to the help of AI. What are the stoppers that prevent you from doing higher-level tasks?
- Do new things. This is the most disruptive part of the exercise. Imagine your job in the future, and imagine the new things that you can do with technology. You need to reformulate the basic hypothesis of your job. What have you never dared to do?
Adopting AI: Go forth alone
The next step is gaining access to an advanced generative AI tool – such as ChatGPT or Claude – and start experimenting with it. Each model has strengths and weaknesses. For many executives, once you take the first steps, imagining additional use cases comes relatively easy, and you quickly have a sense of the types of things that AI should be able to do for you.
The next part of the exercise is more daunting: you’re going to do this without the help of your IT department. AI promises to unleash so many changes in the coming years that you can’t rely on IT for everything. In the initial stages, you have to be able to create your own experiments and see whether they actually work. Many won’t; others will remain as an individual productivity booster. A percentage of those that are successful can be shared with your team. Then, finally, you can bring in the IT department for those applications that may be worth scaling across the organization.
For example, building your own custom GPT is an option available in the premium versions of, say, ChatGPT and no coding is required. You’ll need to think about the data sources and their formats; you’ll be able to upload documents and connect to applications such as email. Next, you’ll want to think of the structure of the activity flow, setting the order of the actions you want your GPT to complete. If your first GPT doesn’t work the way you expect, iterate and improve upon it.
It’s a radical change in mindset for many executives, who are accustomed to thinking that experimentation is the province of entrepreneurs and techies. But such is the nature of this new wave of technology: it changes everything.
Create a culture of AI innovation and measure
Why is it useful to start in this way? Because relying on other people to come up with uses risks producing AI products that people simply won’t use or that don’t perform any better than widely available LLMs. BloombergGPT, a 50-billion parameter large language model, was shown to do no better on financial tasks than ChatGPT – and Bloomberg is a technology and data powerhouse.
Instead, the role of the organization must be to influence, not impose. Organizational culture matters in innovation, and managers play a key role in creating an environment of openness and experimentation. Employees must know that they have the license to experiment, especially since generative AI is a general purpose technology that doesn’t come with a specific use case.
The experimentation should be systematic, however, and include metrics and KPIs. For example, if you use AI to polish off your emails in 50 minutes instead of an hour, make the time savings one of your KPIs. If you begin to use AI for 80% of the translations you require, track the resources saved on translators.
Additionally, it’s important to share experiences with AI at work; in many workplaces, employees are using AI almost furtively. Companies and managers should create incentives to be transparent and communicate what people learn.
Most organizations will have to combine this bottom-up exercise with a top-down approach. Beyond individuals looking to improve themselves, organizations need people thinking at a strategic, company-wide level at how AI can be incorporated into their business models. The two are simultaneous.
Harvard Business Review in April published a follow-up to its viral 2024 article “How People are Really Using Gen AI”. It found that, new this year, the top three AI uses are: “therapy & companionship,” “organize life” and “find purpose.” People are hacking their personal lives. If that seems unappealing, rather than sitting out the AI revolution, start radically rethinking your professional life.