FPF for Analysts: Improving the Accuracy of Assignments, Decisions, and Working Texts

Many errors in analysis do not start with wrong conclusions, but much earlier: at the moment when a task is still underspecified, a decision is not yet formalised, yet the text already looks “clear enough.”

This later leads to extra clarifications, repeated discussions of already resolved matters, documents with drifting meaning, and wasted effort — both for people and for LLMs.

In this talk, we will examine FPF as a practical framework for such cases:

• how to unpack tasks and assignments;
  
• how to capture and record decisions;
  
• how to check texts for semantic and ontological accuracy;
  
• how to apply the same techniques when working with non-human assistants.
  

This is not a general overview of AI and not a discussion on good writing style. It is a talk about making working texts and decisions more accurate in situations where an error in wording quickly turns into an error in real work.