Looking to Enhance your Employability? Learn the Basics of Prompt Engineering.

It wasn’t long ago that futurists were making a bold prediction about data analytics.

Soon, they said, every employee will need to be a data analyst. The argument went that the growing sophistication and accessibility of data analytics software, self-service portals, and features such as data visualization meant that the average employee could work effectively with data without needing technical skills. Every team member should be able to unlock data’s potential to help make informed decisions, measure performance, identify opportunities, and uncover valuable insights without the need for a data specialist to interpret the numbers.

Today, the buzz around data analytics has been swept aside by the tide of excitement around ChatGPT and the possibilities of AI language models. As business functions start exploring ways to incorporate and benefit from this game-changing technology, we predict that a new “must-have” skill will begin to make its way into procurement and supply chain job descriptions: prompt engineering.

What is prompt engineering?

Prompt engineering involves asking an AI language model (such as ChatGPT) carefully crafted questions to test its limits and uncover mistakes and lurking issues such as AI bias. In doing so, the input prompts will help improve the accuracy and performance of the model. The Washington Post used the term “AI whisperer”; someone who knows how to talk to unruly AI in a way that will get it to do exactly what we want.  

Because the conversations with AI take place in English, prompt engineering doesn’t require computer science skills or knowledge of programming language – a fact that led former Tesla AI chief Andrej Karpathy to tweet “The hottest new programming language is English” earlier this year. Prompt engineering requires language skills, creativity, communication, critical thinking, and – most importantly – subject-matter expertise.

Now, we’re not suggesting that you abandon your procurement career to retrain as a prompt engineer … although the $375,000 salary does sound tempting. The point I’m making is that developing a basic working knowledge of prompt engineering will enhance the employability of any candidate and dramatically boost your effectiveness in leveraging AI.

Prompt engineering and procurement

While most AI models are developed for highly specific use-cases, ChatGPT and its contemporaries such as Google Bard are extremely generalized language models. This has been the driving force behind ChatGPT’s viral adoption (100 million users and counting), but is also its greatest weakness.

It’s a matter of data volume. The model will perform well so long as there is plenty of material online for it draw upon, but the quality of responses begins to dip as inputs become increasingly domain-specific. As with any machine-learning model, this situation will improve with time as ChatGPT is tested, trained, and guided in this journey by human procurement experts.

Prompt engineering is one of the ways we can customize generalized AI models to meet our specific needs in procurement. By understanding how to formulate prompts, sourcing professionals can tailor ChatGPT to address unique requirements and preferences, understand the context behind procurement jargon, and enhance the chatbot’s utility and relevance to make it ever-more valuable in specific applications.

What sort of applications? GEP points out several use cases in procurement including conducting rapid category research, drafting RFPs, SOWs and other contracts, analyzing spend data, automating supplier communication, and more. However, human expertise is still needed to vet AI output and ensure the technology is used for decision support rather than decision-making.  

Start experimenting with ChatGPT today

One of the most exciting things about ChatGPT is its radical accessibility. Adoption is on the individual, with no need to wait for instructions from your CTO or a months-long review by the IT team. In short, there’s nothing stopping you from leveraging the technology today. Start by:

  • Understanding how AI (in general) works.
  • Understanding the basics of AI language models and ChatGPT’s capabilities.
  • Assessing the impact it will have on your career. (Hint: everyone is impacted.)
  • Using the technology as soon as possible. Experiment!
  • Try incorporating ChatGPT into various aspects of your procurement workload. Where is it effective? Where does it struggle? Are you asking it the right questions?
  • Share what you have learned with your team.
  • Disrupt what you just did!

Don’t hesitate to get started, and don’t be scared of the technology. ChatGPT isn’t here to steal your job, but it will undoubtedly give a competitive advantage to those who understand how to best put it to use.