PERSPECTIVE · INSIGHT
When you ask an LLM (ChatGPT, Gemini, Claude) to produce insights from a corpus (report, interview transcripts, etc.), you get well-turned sentences in a few seconds, often of the kind "consumers are looking for authenticity."
An insight must be precise
This kind of learning can indeed make it easier to grasp the content of your corpus or provide a quick overview. But an insight used to steer a strategy, a concept or a campaign must be something other than a plausible generality: it must be precise. Without a rigorous approach to what an insight is, it is not possible to manage this kind of content; you can only manage what you have defined and mastered.
The SMT© approach embedded in our algorithms
An insight is the combination of several elements: the context at stake, the motivation attached to it, and the tension factor engaging with that context (our SMT© method). This triangulation is essential to avoid generalities, and every insight therefore carries a challenge. It is the precise nature of this insight that then makes it possible to steer thinking about an offer or a communication angle.
Since 2012, the Insightquest teams have trained nearly 10,000 people in 17 countries in the formulation of insights and value propositions, on the basis of a long-proven approach. Our insight-generation modules therefore natively embed, in the form of code, expertise, instructions, framing constraints, checks, examples, counter-examples. Every generated insight is the fruit of an analysis of the corpus you provide, of cross-referencing between different kinds of information (context, motivation, tension). We apply the same rigor when you choose the "web" option of our insight search.