Understanding humans deeply without multi-page Ai prompts
Ai can be a wonderful tool for quick hits, but the prompt engineering needed to gain a deep understanding of people and identify ways to motivate them can be extremely labor intensive. Writing just the right prompt that gets to the information (insight?) can require a lot of time and research, and I’d argue that there is a point of diminishing returns at which your time and energy are no longer worth the investment.
Talking to people on the other hand, provides a rich, interactive platform for understanding them, making live conversations with real people an effective “shortcut” to gaining insights into people's thoughts, feelings, and motivations. Fielding a round of qualitative interviews, small sessions or focus groups can add tremendous value to Ai or quantitative studies for just this reason.
Talking live is actually the “shortcut” to understanding a person, providing a LOT more than Ai can provide: you get context, emotion, rationale - and the emotional impact of hearing from a person firsthand can be incredibly motivating to your team.
A few principles from Conversation 101 may help make the point:
Active exchange: a conversation between humans allows for immediate exchange of thoughts, feelings, and experiences. Focus group moderators and interviewers are actively listening, playing back, clarifying and asking for more detail. Ai can’t do that, so it simply cannot model an authentic conversation: there is no spontaneous interrupting to signal an important almost overlooked thought, no facial expression or non-verbal cues lighting up to suggest an idea deeply resonates, no pivoting to an unplanned direction based on a provocative thought. This direct interaction is the “art of conversation” which can clarify misunderstandings, deepen understanding and provide rich insights.
Storytelling: An often-used tool in our qualitative toolkit is storytelling. We rely on stories to help us understand the context of a topic, e.g., “tell me about”… “the last time you went to McDonald’s,” “the last time you drank a Coke,” “the time you feel most valued as a parent,” - etc. Nothing lights up a consumer like telling a highly positive or negative story about using your brand: it’s easy to recall, laden with emotion, personal, and strongly formative. Stories are important because what they tell US in research is also what they tell their friends, family and colleagues about your brand. So the more detail we get, the more we can help to guide consumer perception of our brands. Can an Ai model be trained to tell a story about something that happened to it? Yes BUT - it’ll involve a highly detailed and nuanced prompt that basically imbues the model with a life story including personal and collective experiences - so, if you decide to do this, you will be researching, crafting and updating the model’s persona on a regular basis to ensure you’re informing the model enough to get good information from it (garbage in = garbage out). Conversely, talk to a real live person who is the sum of its experience simply by living as a human and gathering experience authentically.
Context: Talking to someone reveals their social environment and influences, helping to understand their context better. If you’ve ever sat through a qualitative interview you’ll know conversations can take unexpected twists and turns - in fact, we shouldn’t call them unexpected: humans are messy, life is messy - twists are part of life. We can be talking about diapers one minute and the next minute delving into the cooking habits of our neighbor (true story, actually happened - and helped me understand a lot about the parent I was interviewing and her values around diapering!) This helps make qualitative research a tool for not only understanding defined topics but discovering new opportunity areas for brands, by following human logic and what is important to people. The majority of Ai models operate in a linear manner, processing information based on logic and predefined conditions, making it virtually impossible to discover something new and unintended. And yes it’s true that more advanced models can engage in non-linear thinking, but they still need to be fed ideas to work with - they can’t (yet) imagine an entire lives worth of experiences or capture the complexities of real human life - and I find it hard it to believe they ever will.
So - looking to save time and effort in your voice of the customer research? Qualitative will happily oblige :)
