This article provides more in-depth information on the Coverage Threshold Guardrail.


Introduction


When using "Minimise Coding Gaps" Auto Coding Mode, the codeit AI calculates how much of the verbatim is "covered" by the suggestions.
The Coverage Threshold controls the percentage of the verbatim which should be covered to be considered as completely coded.


The Coverage Threshold guardrail is available at project-level and at task-level. By default, the tasks will use the project default.



Coverage Threshold Explanation


Consider the following verbatim:
"They give good service, but it was expensive and the staff were rude"

Suppose also that the Machine Learning generates the following suggestions, in response:


SuggestionText Suggested
Confidence
Code 3: Good Servicethey give good service
98%
Code 7: Poor Price/Value / Too Expensivebut it was expensive
91%



Given the suggestions, the total text covered for the verbatim is "They give good service, but it was expensive".
The phrase "and the staff were rude" is not covered by any suggestion, which might indicate that a code is missing.

The Coverage is calculated by the AI using the number of characters in the suggested texts. In this case "They give good service, but it was expensive" contains 45 characters.
The entire verbatim contains 68 characters.
The codeit AI will therefore calculate a Coverage of 45 / 68 = 66%


If the Coverage Threshold is higher than 66%, none of the suggestions will be applied when using "Minimise Coding Gaps" for this verbatim.
If the Coverage Threshold is lower than 66%, all the suggestions will be applied even when using "Minimise Coding Gaps" for this verbatim.