Are the days of the prompt engineer already over?
It was only a year or two ago that the role of prompt engineer was being touted as the next great career path. Thousands of job ads appeared. People the world over updated their LinkedIn profiles to highlight their prompt engineering skills. Online courses promised six-figure salaries for mastering the art of writing prompts in ChatGPT.
Fast forward a couple of years and the role is already on its way out.
What’s changed?
In the last few weeks firms like Microsoft and early AI adopter Klarna have highlighted the prompt engineer role’s impending death. Klarna has stopped hiring for the role, with one executive calling it a plaster which was useful in the early days, but not a long-term solution and now no longer necessary.
According to Jared Spataro, chief marketing officer of AI at Work at Microsoft, two years ago, everybody said, ‘Oh, I think prompt engineer is going to be the hot job. It’s not turning out to be true at all.’
In Microsoft’s most recent Work Trend Index survey, 31,000 workers across 31 countries were asked what new roles their companies were considering adding in the next 12 to 18 months. Prompt engineering was second from the bottom.
The rise of AI agents
This story isn’t just about job titles, it’s also about how fast the AI models are advancing in their capabilities and how AI is being used.
The models have become so good now at understanding what the user is trying to achieve that the previous need to use detailed prompt methodologies (e.g. Microsoft Goal, Context, Source, Expectation) are just not required.
Also, rather than asking people now to prompt a chatbot, businesses are starting to deploy AI agents that can carry out tasks, trigger processes and interact with other tools.
These agents can now do things like:
- Update CRM entries
- Generate and send reports
- Flag issues before they happen
- Coordinate between teams
This sort of stuff is already happening as AI is becoming embedded in day-to-day work, not just bolted on.
But agents aren’t all just easy plug-and-play bits of tech, they need ownership, and oversight. Which brings me to roles that matter now.
What businesses need from their people now
Prompt engineering was useful when no one knew what they were doing with large language models. But we’ve moved past that. What’s needed now is the ability to define the outcomes you want with AI and then be able to use the tools available like Copilot Studio to deliver them.
- That means hiring or training people who can:
- Understand your operations and where automation makes sense
- Build and test agent-led processes
- Connect systems, APIs, and data
- Put the right controls and checks in place
- Work across technical and business teams to deliver outcomes
Where to focus
If your current reality of AI in your business is still just focused on prompting, it’s time to take a step back and ask a few questions:
- Where are the repeatable tasks in our business?
- Which parts of our processes could be run by an agent?
- Who’s responsible for building the agents and making sure they deliver value, are safe, efficient and reliable?
The sooner your team moves from prompting to building, the better placed your business will be.
We are iwantmore.ai – an AI consulting firm that helps organisations to move beyond AI experiments and build tools that last. We help clients design, deploy and manage AI agents and train the teams who’ll run them.
If you’re ready to stop playing with prompts and start building with purpose, get in touch.
Interested in more content like this? Sign up to our Newsletter here.
Other AI articles you may be interested in:
Every company seems to want to use AI. According to Cisco’s second annual AI readiness report, 98% of executives felt an increased urgency to deliver on AI. Many businesses are already implementing it in some way, but based on the findings of the RAND report, more than 80% of AI projects fail.
Every few days, a new AI model seems to drop from one of the big AI companies. But as a normal person, how do you work out whether a new AI model is actually better than what you were using before? In this article we look at ways in which you can test out which model is most applicable for your business needs.
From boardrooms to marketing teams, leaders in businesses are being told they need AI to stay competitive. But amid the hype and noise, one crucial question often gets overlooked - what’s the actual return on investment (ROI)? Is it all worth the effort?
We take a look at the various ChatGPT modules and find out which ones best suited for what business tasks. We look at GPT-4o, GPT-4o Mini, GPT-4, OpenAI o1, OpenAI o1 Mini, OpenAI o1 Pro Mode, OpenAI o3 Mini and look at what they are best for, their purpose and how you access these models.
There’s still some confusion as to what an AI agent actually is. In many cases it’s probably still just simple automation with a fancy name. But at the more sophisticated end of the spectrum, these autonomous systems have the potential to change the way businesses operate.
Many organisations stumble when it comes to implementing AI because they lack that clear framework and organisational structure to manage it. This is where an AI council comes in. But what exactly is an AI council?
iwantmore.ai – The AI consulting firm that helps you build a smarter business
Wherever you are with your AI implementation initiatives, we have a range of stand-alone AI quick start services to help you fast track the transformative benefits of AI across your business.
