Customers: MegaFon Moscow; Telecommunications and Communications Product: Artificial intelligence (AI, Artificial intelligence, AI)Project date: 2021/01 - 2021/07
|
2021: Launch voice assistant in capital branch
On July 27, 2021, MegaFon announced the launch of Eva's voice assistant, which will allow subscribers not to miss important calls, while saving from voice spam. "Eve" can accept an incoming call, independently determine the source and purpose of the call, support the conversation and send the decryption of the missed call to the subscriber in the messenger.
Eva knows how to maintain a full-fledged conversation, for this, the company's specialists have developed several dozen dialogue scenarios, with the help of which the voice assistant can equally effectively process a call from a bank or from a loved one. For example, when calling from a bank with a mortgage proposal, Eva will independently clarify all the conditions with specialists, after that it will transfer the conversation to a text format and send the subscriber audio and a printed version of the recording to Telegram.
As part of the pilot project, several thousand subscribers of MegaFon's capital branch can test the voice assistant for free, it is planned that in the fall the solution will be launched in all regions.
"On average, subscribers miss about 20% of all incoming calls - when the phone is off, or it is inconvenient to answer the call. "Eve" is an effective virtual assistant that will help not to miss important information. All control of the assistant is carried out through a special chat in Telegram - for installation you do not need to download third-party applications. At the pilot stage, we continue to improve the solution, complementing it with new functions and capabilities. In the future, Eva will be able not only to accept, but also to make calls for a subscriber, "said MegaFon Commercial Director Vlad Wolfson. |
The voice assistant was developed at the MegaFon Product Factory, where the company creates technological solutions together with startups. The project team used a database of several thousand calls to train the assistant, involved professional script designers, announcers and producers in the development in order to achieve a high quality of dialogue.