Putting a value on prompt engineering

The deluge of AI news in the past few months has had businesses trying to assess whether they need to adjust their priorities.

The early mover advantage when it comes to adopting tech is real. But is it worth putting in the effort, when it comes to adopting AI and more specifically LLMs (large language models) like GPT4 and Bard AI?

In the upcoming weeks we’re looking at, among other things:

• AI and the impact of tools like ChatGPT for businesses.

• What it means to dedicate resources (such as prompt engineers) towards utilising AI.

• The long term prospects of AI.

We’re starting with the basics in this issue: How can you know if prompt engineering–a practice that allows you to take advantage of LLMs–adds value to your organisation?

The short answer: It correlates to however much value LLMs are giving you.

For the longer answer: Read on.

Give us your thoughts on this issue:

So what’s the deal with prompt engineering?

A screenshot of a Twitter post i’m offering here with no explanation:

Something to note: this doesn't necessarily mean that organisations have totally figured out how and where they can best use AI. But they are definitely putting in the resources to find out.

Are they making progress? A quick look at some of these job postings shows us that companies are looking to use prompt engineering to:

• Generate content

• Code

• Create images

• Create presentations

• Generate UI

• Analyze sentiment of text

Broad use cases that, really, any business would find useful. (But we know that the real value comes when you can use tech to solve specific problems)

Prompt engineering to solve specific business problems

A summary of prompt engineering use cases for a few industries:

Finance: Personalized investment advice and planning for customers. Named entity recognition, keyword extraction, and financial data extraction from financial texts.

Healthcare: Summarizing patient medical history and symptoms, quickly identifying drug interactions, and creating personalized treatment plans.

Education: Feedback on student assignments, personalized engagement and learning plans, creating interactive learning experiences.

Speaking of industries such as healthcare and finance, which are particularly sensitive about data, there are some valid concerns that need to be ironed out. Read: The potential and thorny ethics of generative AI in healthcare.

Coming back to value: There’s a lot of progress that is still being made, but early signs show that prompt engineering can contribute much, by way of saving time and resources for businesses.

Things of Interest

(No sponsored links here 😉 )

• H2O Driverless AI - Automatic machine learning (AutoML) platform that uses automation to accomplish key machine learning tasks.

• Durable AI - A website builder that claims to be able to build a site in 30 seconds.

• Bearly - AI tool that helps you come up with and produce content ideas.

🎙️ A case for librarians as prompt engineers

A blog post from February by Laura Solomon talks about how librarians like herself may be in a position to use their skills for prompt engineering.

Librarians excel at creating optimal search parameters (think of the information they need to collect from readers who are looking for a document, research paper, or book). This overlaps with the skill sets required of prompt engineers.

She ends with a good question: Will librarians be viewed as savvy enough to write good prompts, or like for some of the more specific use cases from the section above, will businesses need to look for a different mix of skills from their prompt engineers?

See you next week,

Mukundan

Do you have a unique perspective on developing and managing data science and AI talent? We want to hear from you! Reach out to us by replying to this email.