4 Tips to Leverage AI for Qualitative Research Screeners

Published On: March 27, 2024By

“Hi! I’m writing a qualitative market research screener…” I type my prompt into the chat with a Large Language Model (LLM), an Artificial Intelligence tool. Twenty-minutes later, I do not have a completed screener ready to launch, but I am armed with category context, copyable lists, and example question formats to make my screener writing process swifter and easier.

Like this article from QRCA details, qualitative researchers are finding new uses for AI tools every day, from reporting to analysis to recruiting. Initially, I was skeptical to dive in, concerned the tools were undermining creativity and analytical thinking and concerned about data privacy. It took being stuck on a B2B screener for me to cautiously wade into the AI waters, and I quickly found how these tools can be extremely helpful if approached with intention.

Here are 4 Tips for using AI to support screener writing in qualitative recruitment:

#1 Ask questions that gather context for the category

Working across multiple verticals is one of the things I love most about my job, but I can’t pretend to know everything about all the categories we do work in. Recently, I’ve used LLMs to gather helpful context for the project’s industry, from understanding brand competitors to regional differences. I’ve asked questions like:

  • What are the top alcohol brands in the state of Texas based on annual revenue?
  • What income range would be considered “high net worth” in South Korea in their currency?
  • What are management-level job titles with similar functions in the industry we’re discussing?

Asking multiple specific questions—instead of “Tell me everything you know about…”—can better get at the background information and data points most helpful to your end goal. As the “discussion” continues, these can ladder up to understanding the behaviors and demographics of the people you’re trying to recruit.

#2 Provide detailed context in prompts

Thoughtful prompt writing is a skillset all its own; LLMs work best with detailed and precise context. This doesn’t mean spelling out all the project details. Oftentimes, it’s just about adding several clarifying words to your prompts. Instead of “I’m writing a screener,” write “I’m writing a qualitative market research screener for a B2B study in X industry…” If working correctly, the tool will frame all its responses throughout the “conversation” around the context you provided in the initial prompt.

#3 Throw it the busy work

AI tools are sophisticated and can handle higher-level problem solving, but they can also be your lowest level assistant. I’ve asked the bot, “Please alphabetize the above list for me,” and it was there a second later. Want something reordered? Randomized? Put in multiple choice form? Asking AI to do your busy work in screener writing can save time and leave brain space for the “higher-level thinking” only you can do.

#4 Always check the responses—even the “simpler” ones

It’s not just written copy and more sophisticated responses from AI that need to be carefully reviewed. They can flub the simpler questions too. I recently asked for a list of the most popular high-end department stores based on revenue and store visits in heavily populated US cities. After seeing a store name I didn’t recognize and revisiting a search engine of auld (Google), I caught that the bot’s list included a store in Canada. I relayed its mistake, and received the response “You’re absolutely right, I apologize…Thank you again for catching my error.” While AI tools are improving their accuracy every day, it’s essential to be vigilant when reviewing even the “simpler” responses. At least they’re apologetic when they mess up!

As I’ve gotten more and more comfortable asking AI for assistance in screener writing, I’ve gotten better at leveraging the tool. I’ve also had some fun, and realized, bot or not, it never hurts to be polite:

  • Emily: Thanks! This is helpful!
  • Bot: You’re welcome! I’m glad I could provide some useful information…Happy to help further and chat more!
  • Emily: Do you have a favorite brand in this category?
  • Bot: As an AI assistant without personal preferences, I do not. However, I’m happy to provide some objective perspectives on exceptional and popular brands!

For more on our tests of KNow vs. AI, make sure to see my colleagues, Sadie Mills and Julia Isaacs, speak about our experiences with AI at The Insights Association Annual Conference on April 8th.