AWS, Google release contact center GenAI tools

AWS releases contact center workforce management features and post-call summaries powered by GenAI; Google previews its own for Contact Center AI Platform.

Both Google and AWS released generative AI tools for contact centers last week at Enterprise Connect.

AWS, whose Amazon Connect contact platform is based on its homegrown software used in Amazon's own 100,000-agent customer service operation, launched several new GenAI features. The first is deeper third-party application support, which passes data among the 10-plus applications that agents typically use.

GenAI post-contact summaries are based on transcription and keyboard data, which agents typically must write themselves. Not only does that take time away from agent productivity, but GenAI can also capture more information for the contact center's knowledge base more accurately, said Pasquale DeMaio, vice president at AWS and general manager at Amazon Connect. The feature can not only potentially save time but also train AI models to be more helpful to both human and virtual agents.

Another feature AWS released was in the workforce management (WFM) vein: automated agent evaluations with Amazon Connect Contact Lens, its speech analytics tool set. Such a concept might worry agents who have read stories about Amazon's history of Flex replacing employees with bots as well as monitoring employees with cameras in vans and warehouses.

DeMaio said agents shouldn't be wary of monitoring features in Amazon Connect because historically, managers have overseen a given agent's calls randomly. If the one call a manager chose to examine didn't go well, that agent might unfairly be singled out as a poor performer.

"And if you happen to get a good call, the agent seems like they're the best agent ever," DeMaio said. "Do you really know how that agent's performing, or do you see trends on how your contact center's performing? Are you learning from all this everything you could? The answer has been no."

Automated agent evaluations run analytics on the whole of an agent's performance. That, in turn, will surface opportunities to update training, identify the agents who need help and give managers a clearer understanding of what's happening with every customer contact.

The GenAI features AWS released for Amazon Connect match what many other vendors are doing right now, said Shelly Kramer, an analyst at TheCube. AWS isn't alone in its automated agent performance evaluation. Some vendors such as Cisco are also monitoring agent wellness with GenAI, which Kramer said is a use case all contact center software vendors should focus on next.

"The job of a contact center agent is incredibly stressful. You've got workers who are pretty much always dealing with [customers] that are not having their best days, [and they're] working for low pay," Kramer said.

"High churn rates are an issue across the industry and have been forever. Developing functionality that helps monitor agent stress levels and 'wellness' based on their call performance is infinitely more attractive as a feature than 'we are excited by the ability for managers to have performance evaluations done automatically.'"

An example of an AWS Amazon Connect contact center agent post-call evaluation.
Generative AI-powered post-call evaluations for Amazon Connect show managers how each call an agent takes compares with everyone else for several common metrics.

Google previews its own WFM

AWS isn't the only vendor focusing on workforce management with its contact center generative AI. Google previewed new features for its Contact Center AI Platform (CCAIP), which pairs with Ujet's agent desktop.

Developing functionality that helps monitor agent stress levels and 'wellness' based on their call performance is infinitely more attractive as a feature than 'we are excited by the ability for managers to have performance evaluations done automatically.'
Shelly KramerAnalyst, TheCube

Previewed last week were tools for AI-powered agent scoring similar to AWS's release; actionable agent insights, which scan customer contact content for emerging topics and measure how agents are performing; and new dashboards to visualize contact center metrics.

Google also added multi-region failover systems to its contact center infrastructure for business continuity and disaster recovery to bolster uptime in the event of a regional outage.

Next week's Google Cloud Next conference might bring more CCAIP previews and releases.

AI tools such as these from AWS and Google are difficult to budget, said Dan Miller, founder of Opus Communications. Contact centers he has talked to are still trying to determine setup, staffing, training and possible LLM costs over and above the consumption-model pricing their chosen GenAI chatbots might incur. While contact centers have measured agent performance and customer satisfaction for decades, GenAI is mostly an unknown and continues to evolve.

"Contact centers are looking for ROI metrics. That's why they have homed in on summary for GenAI, because every second they shave off average handling time saves a 'bajillion' dollars," Miller said. "What's missing is that nobody knows what this costs."

Don Fluckinger is a senior news writer for TechTarget Editorial. He covers customer experience, digital experience management and end-user computing. Got a tip? Email him.

Dig Deeper on Customer service and contact center