View source for US Patent Application 17828141. APPLICATION ROUTING BASED ON USER EXPERIENCE METRICS LEARNED FROM CALL TRANSCRIPTS simplified abstract

=APPLICATION ROUTING BASED ON USER EXPERIENCE METRICS LEARNED FROM CALL TRANSCRIPTS=


==Organization Name==

[[:Category:Cisco Technology, Inc.|Cisco Technology, Inc.]]


[[Category:Cisco Technology, Inc.]]

==Inventor(s)==

[[:Category:Grégory Mermoud of Venthône (CH)|Grégory Mermoud of Venthône (CH)]][[Category:Grégory Mermoud of Venthône (CH)]]

[[:Category:Jean-Philippe Vasseur of Saint Martin d’Uriage (FR)|Jean-Philippe Vasseur of Saint Martin d’Uriage (FR)]][[Category:Jean-Philippe Vasseur of Saint Martin d’Uriage (FR)]]

==APPLICATION ROUTING BASED ON USER EXPERIENCE METRICS LEARNED FROM CALL TRANSCRIPTS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 17828141 titled 'APPLICATION ROUTING BASED ON USER EXPERIENCE METRICS LEARNED FROM CALL TRANSCRIPTS

==Simplified Explanation==

- This patent application describes a device that analyzes call transcripts from an online application to identify cues indicating a poor user experience.
- The device then uses these cues to create a model that can predict poor user experience based on network path telemetry data for the online application.
- Based on the predictions made by the model, the device can route traffic for the online application along a specific network path.
- The goal of this innovation is to improve user experience by optimizing network paths for the online application.


==Original Abstract Submitted==

In one embodiment, a device obtains call transcripts from an online application. The device detects cues within the call transcripts that are indicative of poor user experience. The device generates, based in part on the cues detected within the call transcripts, a model trained to predict poor user experience from network path telemetry for the online application. The device causes traffic for the online application to be routed along a particular network path, based on a prediction by the model.



[[Category:H04M7/00]]
[[Category:H04M3/42]]
[[Category:H04L41/16]]