Welcome to the system testing section!
System testing comprises a set of checks that you as developers of the agent do to identify how well the agent and its components operate when exposed to diverse input. This is done in a continuous manner when developing your bot, by simply engaging in conversations with the agent and analyzing its functionality. You should use your bot a ton (at least 10 different conversations per person), and try and vary the conversations you have. You should also try and trigger each pattern, page, intent, and filter to check if it all works. As you improve it and add things you should perpetually re-test your conversational agent. We recommend that each team member should try and test other sections' parts.
Part of your end report will be about how you did this (i.e. do you know how to trigger each intent, page, etc.).
In the 2023: End Report you will see a section asking how you tested and evaluated your bot.
You should use your bot a ton(at least 10 different conversations per person). Test everything to see if it is working i.e. make sure you trigger all possibilities, patterns, pages, filters, intents etc.
Then, analyze no-match conversations in Training. Go to the Training page in Dialogflow and filter conversations at the top left, this video provides a more thorough explanation of how to do this: Use Dialogflow Analytics & Training to Improve Your Chatbot (2021). Due to us using SIC Analytics part does not apply.
Things we think you should keep in mind during testing and should be included somewhere in your End Report:
Capabilities of your bot
what do you think is most important to test in this phase?
Test set up
What an example conversation should look like
What did you test and how
mismatch conversation analysis
how did your test go: good and bad( focus on the bad and why it went wrong)
how could one fix problems. will you improve it before turn in or is it not feasible?
during use, what kind of extensions do you think could be useful or improve the performance
how could one even further extend
0 Comments