Why that AI result was not necessarily useless
Dr Susanne Friese · · Research practice
Why That AI Result Wasn’t Useless (Even If It Looked Like It)
You probably have either said it yourself or have someone heard saying it or writing about it: You cannot use LLMs for qualitative data analysis because.....
If you take a closer look, often what has happened is: they have thrown their interview transcripts into a general-purpose chatbot, ask it to “analyze,” and then… get vague or even hallucinated results. And then?
They blame the tech and declare: ❌ “AI doesn’t work for qualitative analysis.”
Wait — not so fast. What if, the AI didn’t fail but the approach did. What if you begin to look at it like this:
A bad output often mirrors a vague or overly ambitious prompt. Instead of discarding the result, you can use it as feedback to reflect on: • What exactly am I asking? • Am I expecting too much too soon? • What context did I fail to provide? • In this sense, a “bad” answer becomes a diagnostic tool for your own thinking process.
--> A bad result reflects the prompt you gave — and reveals how you’re thinking.
An AI result might be inconsistent or a bit “off,” but that doesn't mean it’s worthless. It might offer: • A turn of phrase that triggers a new idea • An unexpected interpretation worth considering • A challenge to your assumptions
--> AI can provoke thinking, even when it doesn't deliver a final answer.
The key isn’t to ask for answers — it’s to spark a dialogue. AI can help you think deeper, notice patterns, and explore multiple interpretations — if you use it the right way.
👉 Curious? Frustrated? Skeptical? Great — that means you’re ready to watch my latest video:
You are using AI completely wrong (for qualitative analysis)
You’ll see: • A typical “wrong” way of using AI • Why this approach leads to bad results • What’s missing: collaboration, context, and the right kind of questions • How you as the researcher stay central to the process
🎥 Watch the video
youtu.be/VQgjFc_j_jE now and let me know what resonates.
Kind regards,