Any conversational AI project has to deal with the challenge of handling ambiguous user input.Teneo provides different tools that help Conversational AI Developers implement strategies to disambiguate the user input.
You can find information about the topic here:
If you are using Teneo Web Chat as your frontend, you can also have a look at the UI options that are available out of the box, such as buttons, quick replies and more.
In our Office Hours today we discussed what an ambiguous input was for us. Our definitions used words like:
The general approach was to use a disambiguation strategy in the projects to redirect the user to the relevant content in the solution when possible.
We also discussed pro’s and cons of using UI elements such as buttons vs. more open-ended questions.
How do you disambiguate unclear inputs in your project?
We built our disambiguation flows rather late, when already live with the first iteration of our bot. Would now recommend though to build these flows early on and avoid falling into the Safetynet too often- gives the users a much better experience!
Hi @tomp,
I completely agree. Including disambiguation flows in your project is something you should consider from the beginning. It allows you to catch and redirect the in-scope intents that were phrased in a way that you did not predict from the beginning, as well as identify potential new intents related to the same topic. And the user experience is much better than falling through to the SafetyNet.
