It Looks like you are using the self.speak_dialog() function and not the self.speak(). The speak_dialog() function requires a *.dialog file located in the /dialog/en-us/ directory of the skill. As an example if you had a self.speak_dialog(“test”) then you would require a file called test.dialog in that directory. The test.dialog file could contain…
This is a test for research
This is a research test
Testing for research
The speak_dialog() function would randomly select one of the three entries for a spoken response.
Hope this helps.