Skip to main content
A Problem in Kontext represents something your users are struggling with. Problems are detected automatically from Feedback and consolidated over time as new perspectives emerge.

​
Detection

When Feedback is analyzed, Kontext identifies specific Problems mentioned in the text. Each detection includes:
  • A Problem statement describing the issue
  • Highlighted quotes from the original Feedback that support the detection
  • The Actor and/or Context associated with this particular mention, when they can be identified
Not every piece of Feedback makes it clear who is speaking or in what situation. A terse bug report, for example, may surface a clear Problem but offer no signal about the Actor or Context. Kontext won’t invent what isn’t there — if the source material doesn’t contain enough information, the Problem is detected without those dimensions.

​
Consolidation

The same Problem often appears across multiple pieces of Feedback. Rather than creating duplicates, Kontext consolidates — recognizing when different Feedback items describe the same underlying issue. Each new piece of Feedback that maps to an existing Problem adds a new facet — a Contextualization against your Context and Actor definitions — rather than just incrementing a counter. The Problem builds up in richness with every new perspective.

​
Problem understanding

Kontext tracks two dimensions for each Problem:
  • Feedback volume — how many pieces of Feedback mention this Problem
  • Contextualization depth — how many distinct Actor/Context combinations have been observed
A Problem with high volume but low Contextualization depth may need more exploration. A Problem with deep Contextualization is ready to act on — you have the full picture.

​
Similarity and merging

Kontext identifies similar Problems that may describe related but distinct issues. This helps you spot patterns across your Product and decide whether similar Problems should be consolidated or kept separate. When two Problems that were initially distinguished turn out to be the same, you can merge them. Merging combines their Feedback, Contextualizations, and quotes into a single Problem. This also feeds back into the analysis pipeline — every merge helps Kontext better detect duplicates in the future, reducing the need for manual correction over time.