Comprehensive Approach to Language Model and Content

Energy Company Cuts Costs While Achieving Employee Satisfaction and Productivity

Challenges

The IT service desk team at this financial technology (fintech) company used a public-facing IT channel on Slack as a means for employees to get help. This caused a strain on the service desk since agents had to constantly monitor the channel, negatively impacting MTTR and employee productivity. In addition, top use cases required agents to handle them, taking significant agent bandwidth for highly repetitive tasks. Also, while the company invested in building knowledge content in Confluence, they had no visibility into what content was working.

Solutions

By making Espressive Barista available via direct messages on Slack, the number of questions going to the public IT channel was dramatically reduced. Barista integrates with Slack as a native app, answering employee questions or creating tickets in Jira on their behalf. The company automated their top use cases, including one that required a custom integration between Barista and their CMDB, enabled by the Barista Control Center. Also, because Barista enables feedback at the end of every transaction, the service desk team can identify content gaps and required updates.

Results

With Barista integrated into Slack, employees save time by receiving immediate, personalized answers. And with fewer employees relying on the public IT channel, agents are able to focus on more strategic initiatives. Now Barista deflects over 3,300 tickets per month, including 500+ that are fully automated (e.g., VPN connection issues, password reset), significantly reducing MTTR.

With Barista integrated into Slack, employees save time by receiving immediate, personalized answers. And with fewer employees relying on the public IT channel, agents are able to focus on more strategic initiatives. Now Barista deflects over 3,300 tickets per month, including 500+ that are fully automated (e.g., VPN connection issues, password reset), significantly reducing MTTR.

Comprehensive Approach to Language Model and Content

Virtual support agents (VSA) are designed to provide immediate, personalized answers to employee questions and issues with the goal of deflecting help desk tickets. When done successfully, this can minimize the need for tier 1 support, yielding a significant return on investment. The key to success is a language model that recognizes employee language with a high degree of accuracy along with content that results in high deflection. This white paper discusses how Espressive accomplishes this.

What you will learn in this white paper:

• What a crowd-based architecture is and how it's used to achieve high accuracy
• Using phrase recognition and advanced NLP to understand employee language
• How to quickly deliver content needed to obtain high deflection
• How the Espressive language model continues to learn