#AI Tip – Ask AI for multiple suggestions, to leverage the human element
Ray Poynter, 18 January 2025
Here is another tip from my courses on how to use Copilot and ChatGPT for market research and insights. Generative AI can be used to ask for suggestions on how to do a specific task. For example, we might ask the AI platform to design a research project. However, I have found that it is a good idea to ask for multiple suggestions.
For example, I might use the following prompt with an AI Platform:”My client wants to conduct a qualitative research study into how modern couples in America organise their purchase of food for eating at home. Suggest three research designs/approaches to meet the client’s needs.”
The result I got was:
1 Ethnographic Study:
This approach involves observing and interacting with couples in their natural settings to understand their behaviors and routines around food purchasing. Researchers could conduct home visits to observe meal planning, shopping, and preparation processes. This method would provide deep insights into the social and cultural contexts that influence food purchasing decisions. Ethnographic studies are particularly useful for capturing rich, contextual data that quantitative methods might overlook.
2 Diary Study:
Couples would be asked to keep detailed logs of their food purchasing activities over a certain period, including decision-making processes, the roles and responsibilities of each partner, and the motivations behind their choices. This could be complemented by periodic in-depth interviews to delve into particular aspects of their diary entries. The longitudinal nature of diary studies can reveal patterns and changes in behavior over time, and the qualitative data obtained can provide a narrative of the shopping experience from the perspective of the participants.
3 In-depth Interviews:
Conducting in-depth, semi-structured interviews with couples can uncover the underlying reasons, beliefs, and attitudes towards food purchasing. The flexible nature of these interviews allows researchers to explore topics that emerge as significant during the conversation. Interviewees can be selected to represent a range of demographics, lifestyles, and dietary preferences to ensure a broad understanding of different experiences. The analysis of these interviews would focus on identifying common themes and unique insights into the food purchasing process.
This is the point the human steps in. I know my client won’t have the time and the budget for ethnography, so I need to think about the diary and the In-depth interview options. I decided to focus on the in-depth interviews and interactively explored switching from in-person to online and from one researcher to three researchers.
The point about asking for three, four, five or whatever options is that it leverages the researcher’s knowledge. We do not ask perfect questions and so we do not share sufficient context with the AI. By gathering several options and then taking some of them and digging deeper we use the machine as a prompt for the human to more fully consider a wider range of options. I think over the next three to five years, using AI to leverage the human will be a key area for success.
Interested in learning more?
You can check out the courses I run at https://newmr.org/training-courses/