Accelerated Answers: ANALYSIS

There are many common misconceptions about how Large Language Models (LLMs/AI) can be used and utilised effectively to streamline business processes. One of the core applications for me, in the area of in-depth market research, is at the analysis stage.

So, for clarity, the source data or “dataset” is original, primary insight captured from 1-1 depth interviews or focus groups (NOT AI generated synthetic responses). So, just like any normal research output.

(1) The traditional, manual process requires the researcher to meticulously analyse and make sense of recorded interviews, often transcribed into written scripts. This necessitates ‘coding’ many thousands of words, sentences and phrases into themes. It is a very time-consuming and intensive process. However, totally necessary, if any genuine insight is to be gleaned. It is a skill in itself.

(2) The skills of the researcher are not replaced by AI. In fact they are put to much better use with the help of AI. LLMs do the laborious heavy-lifting, sorting, retrieving and interrogating the dataset. The researcher is in control, engaging directly with the dataset and exploring its corners by skilled interrogation of the source data via LLMs. 

(3) Less time is spent on the mechanics and much more time on seeing patterns, interpretation and deeper understandings. There means MORE and BETTER direct engagement with the source data.

EXAMPLE:

A recent international survey generated 107 in-depth interviews over two related but separate interviews. This was over 5 countries in 5 languages. Normally, this would have taken months to complete, transcribe, analyse and report on. With Accelerated Answers fieldwork and analysis were completed within a month.

VS.

In contrast a previous study I completed of similar complexity, done “manually”, with approximately half the sample size (57 in-person depth interviews), took 5 months to complete. It was over twice the cost.

#AIForResearch #AcceleratedAnswers #StewartRedpath