It will be great to provide an API allowing developers to interact with Cortana following this workflow:
1- Get command from Cortana.
2- Do some processes.
3- Get back to Cortana with response data (do not exit from Cortana UI).
4- Cortana responds the user.
Please vote if you're interested by this feature request.
Peter Wone commented
This is what I want, too. What currently exists is essentially sophisticated recognition and parsing followed by pass-through. It's one way. It's a very good start, but it treats Cortana as little more than a thin veneer over a normal app. If Microsoft wants this to be a category killer, this is my advice: treat Cortana as an AI framework. Apps should be able to function as domain knowledge providers. They may also have conventional UI; this is a separate matter. Use a VCD style file to specify weighted keywords to help Cortana assess service provider relevance. The "VCD" file should also specify a state machine definition for scripting conversations. There should also be a way to return an explicit spoken response. Implicit in the above is a need for developer tooling for preparing the state machine and the associated voice files.
Please add the German language to be faster than in a year . For this to feed into the development of new applications .
Mohamed Ouederni commented
Its about empowering Cortana with knowledge from apps without the need to say "Cortana, open app X do Y".
But by saying "Cortana do Y" directly.
I' don't get the idea! Is it a proxy?