Do you remember, the 21st night of September?
You will have to forgive us as it's not quite the 21st of September when we are posting this. We got pretty excited about bringing these new features to life and had to blog about it a little early.
We've updated WatchTower at EdgeTier, here's the latest major feature upgrades that are now live.
Emotion Detection
What would you do if you could see what every customer was feeling when they speak with your brand? Would you evaluate agents? Would you check what product features are delighting customers? Would you track their feelings daily?
Now you can start to do just that. WatchTower now scans every message that we process for customer emotions. We've focussed on key emotions that provide a proxy to customer satisfaction: frustration, gratitude, and praise.
Frustration: "This is all your fault. I am a valued customer of yours and this is how you treat me?"
Praise: "That's super, you've been brilliant, thanks for your help!"
Gratitude: "Thanks!"
The EdgeTier data science team has completed a huge amount of work and research to build multilingual emotion detection models for WatchTower for this feature. We'll be expanding on this function with additional reports and views over time, so watch this space.
From today, EdgeTier users can filter by emotion on the anomaly and explore screens in WatchTower.
Interaction Metrics - Handling times, NPS, and more
"You can only improve what you can measure." To run the best contact centre possible, it's important to have a handle on all of the metrics that affect agent performance and customer satisfaction. For supported systems, WatchTower now ingests key interaction metrics for chat and email interactions, including handling time, first response time, queue time, net promoter score, and customer satisfaction.
As a first step, we've surfaced this data in the interaction tables for anomalies and on explore screens, users can sort by these metrics as required when exploring data.
Improved Interaction Views
WatchTower now surfaces all metric and category values to users when examining the details of any interaction. For all users, getting as much information about any interaction is important to understand the complete customer context. As such, all detected tags are displayed when users expand on interactions from any interaction table in the system.
Enhanced Anomaly Filters
On the primary anomaly page, after user feedback, we've added the ability to filter anomalies by interaction type, model type, and model name. This feature is particularly useful if users are interested in specific anomalies or sub-sections of their contact centre.
Coming Soon
Thank you to everyone who has provided feedback on WatchTower thus far, your knowledge and experience has been invaluable. We're not stopping now and have some very exciting new features planned for the coming months. The key areas you're going to see changes next are:
- Additional search filters based on numerics/metrics
- A dedicated emotion summary and exploration screen
- Analyses of agent performance data
- Manual alerting for known phrases
As always, thank you for working with EdgeTier, and if you have any questions or feedback, please reach get in touch via the button below.
