This is how I started and I was getting mixed results each time even for the same ticket id:
What are the specific inconsistencies found in xxxx when reviewed against related closed stories and technical documentation.
Understanding the syntax of the English it was using was the hardest challenge for me. There are so many ways to say the same thing, it was hard to know where to start.
Setting the context was key for me, so that it didn’t generate extra results e.g. looking for and not finding test cases.
*- Business analysts, software developers, and software testers are analyzing a new requirement.
- They are are determining if there is enough information documented to write test cases and start development on it.*
I also needed to define what it was looking for, and provide synonyms. This reduced duplication in the results and I got better coverage from the analysis it did.
***Definitions**
- A gap is when one of the following is mentioned in the linked JIRA items or confluence pages but not in the specified JIRA item:
Requirement
Specification
Feature
Deliverable
Rule
Criteria*
I then had to break down the instructions as granular as I could think of (after several iterations) . Sometimes it would not find the expected links until I specified the order I want them analysed:
*- You will list the gaps in the JIRA item compared to the following:
- the parent item
- all sibling items
- linked JIRA items with Resolution of Done
- You will list the gaps in the JIRA item compared to confluence pages related to the JIRA item.
- You will list gaps in the JIRA item when compared to related Sharepoint sites.*
You also need to be explicit with detailing the output you want, so I included this to generate a table:
Recommendations will contain a table with the following headings: Persona, Summary of Gap/Inconsistency, Action required
The whole prompt would fill an A4 page now!
Hope this is of use.