ellipse
pattern
May 26, 2021

[24]7.ai Agent Assist Dramatically Boosts Google CCAI Agent Response Rates

Michelle Gregory
Michelle Gregory

Senior Vice President, Data Science

[24]7.ai Agent Assist Dramatically Boosts Google CCAI Agent Response Rates Breadcrumb

We find that using [24]7.ai™ Agent Assist, a feature of [24]7 Digital Assist™, our AI-enabled agent assist platform, dramatically improves the agent response rate for any third-party AI assistance platform, including Google CCAI, as well as for both products in combination. 

AI agent assistance technologies improve contextual “next best response” recommendations, which is proven by live agents in real-world interactions to save time, reduce AHT, and increase containment. [24]7 Digital Assist employs proprietary training data and ranking models to present our agents with the top three responses.

But how does [24]7 Digital Assist stack up with other agent assistance products such as Google Contact Center AI (Google CCAI)?

Before I dive into the analysis, I’ll just note that it is not a binary choice on our [24]7.ai Engagement Cloud™ platform: You can use [24]7 Digital Assist alongside other similar technologies.

Here we analyze the following:  

  • [24]7 Digital Assist trained on our proprietary data
  • Google CCAI trained on the same data
  • Combined output from [24]7 Digital Assist and Google CCAI put through our ranking model
     
Response Ranking Analysis
Analysis used [24]7.ai chat data and did not involve all Google CCAI features.

 

Better Together: Agent Assist Technologies


As noted, with Agent Assist, the system gives agents a choice of responses.

When analyzed separately, agents choose Google CCAI responses at a rate of .34 percent “recall”—meaning agents select one of the top three responses as the best response about one third of the time.

With Agent Assist, by contrast, agents choose one of the top three responses more than half the time (.56).

And with combined Agent Assist and Google CCAI output, run through our proprietary ranking model, agents choose these response recommendations over 50 percent of the time as well, slightly less than with Agent Assist alone. 

This is a good benchmark of our system—and these results demonstrate that [24]7 Digital Assist is flexible enough to work in combination with other tools and provide value on top of them. This is truly a unique capability that enables the most flexibility for our clients. 

Want to Learn More?


Check out these additional AI-enabled agent assist resources:

Web page: [24]7 Assist™

Web page: [24]7 Digital Assist™

Blog: AI for Contact Centers = Better Customer Experiences

Related Posts

Want Happy Customers? Start with Your Work-from-Home CX Agents
May 28, 2021 Want Happy Customers? Start with Your Work-from-Home CX Agents

I recently moderated a panel…

Business Text Messaging: Measuring Success
May 14, 2021 Business Text Messaging: Measuring Success

Welcome to the third post in a blog…