Using speech analytics, we’ve managed to highlight the characteristic features we need just by pressing a couple of buttons. As a result, we no longer require a dedicated employee to monitor calls and we can hand over some of the workload to a senior salesperson. That person is only involved in searching for the calls we need and, in rare cases, call monitoring (which has been reduced to a minimum), and dealing with negative issues, which is important
calls per day:
This tool has significantly reduced the time required for call monitoring. Previously, we had one employee allocated for analysis of how the call center operated and that was their fulltime job. For example, if they wanted to find a call where a client complained about delivery or a courier, they would have to go through the entire history of calls: figure out the time, identify the client, listen to a few calls, find the reason, and resolve the complaint. One case would take half a day. Now, when you need to find a conversation, you can just do a keyword search in the transcript and quickly familiarize yourself with the contents of the dialogue.
Automatic tagging has made it possible to forego manual tagging to understand the reason for a lack of calls, even when there is traffic to the site. Premier Techno created a breakdown of advertising channels using keywords («refrigerators», «vacuum cleaners», and «televisions») and compared it to the number of requests for the same products by phone. It turned out that the sellers were not to blame for the lack of sales. It was the conversion on the site due to uncompetitive prices. After they were adjusted, sales went up.
After setting up autotagging by keyword, the company was able to calculate the demand for goods to build a sales plan.
Autotagging enables you to see the number of goods being rejected, calculate the amount of financial loss, analyze calls, and make calls to get customers back.
Example of a table that can be made based on statistics for demand and rejections. It shows how much money the company did not make.
Since autotags are clearly the work of an algorithm that finds keywords in speech, we can say that SmartTags are what draws attention to the context, in a generalized manner, everything that is difficult to describe in regular expressions. The neural network reveals abstract patterns that cannot be algorithmized due to their having too many features. For Premier Techno, there was an idea to use this component to generate checklists: combine SmartTag data with autotag data, identify trigger words that are common in meaning, and use them to create requirements for employees when communicating with customers.