Loading...
Loading...
Loading...
In bots, intents describe a goal or task that a user wants to do, such as tracking their package or booking a cab. A user interacts with a bot to resolve their intent.
Intent miner searches a historical set of recording data, specifically chat or voice transcripts, between agents and customers for a specified duration of time. Next, it extracts the top set of intents and the analytics that surround those intents. Finally, it returns a list of utterances that the system associates with that intent to allow for import into a bot configuration.
When you create a new intent miner, you select the date range for conversations that you want the miner to include in the search. The miner fetches all conversations that occur within the date range and deduces the dominant intents that are present. Then the miner anonymizes customer utterances, computes analytics, and associates the utterances with the intents before returning the results to you.
When you select to mine intents for a new mining session, a progress bar appears in the tile to display the current mining status. The number of conversations mined can impact the length of the mining process. Depending on the data source and the media time, mining the maximum number of conversations can take up to 24 hours. Most this time is spent fetching the conversation transcripts. When complete, the miner contains a list of mined intents. Now, you can choose utterances for the mined intents.
You can access the intent miner directly from a Genesys Dialog Engine Bot Flow or Genesys Digital Bot Flow in Architect. For more information, see Access the intent miner. You can also create a new intent miner by importing a .json file from an external source.
To create a new intent miner from a Genesys Dialog Engine Bot Flow or a Genesys Digital Bot Flow, perform the following steps:
The mining process begins and the Miner Status column informs you of the progress. If an error occurs, you can open the session, review and correct the error, and then restart the session.
Get user feedback about articles.