AWS: Get Ready for Generative AI to Rule CCaaS

One of the first places channel partners likely noticed the presence of artificial intelligence and generative AI solutions was in the contact center.

AWS’ Michael Wallace

“Based on current technology and a solid understanding of reality, we know that generative AI can do amazing things, but it’s not a replacement for human interactions. It’s more of an augmentation of what an agent, a supervisor, or a contact center administrator is capable of,” Michael Wallace, Americas Solutions Architecture Practice for Customer Experience at AWS, told Channel Futures.

Wallace shared that the channel might gain a lot if it can convey the might of AI and what it can do for those in the CCaaS (contact center as a service) space. To begin with, there is the potential for gains in agent productivity and customer experience by adding it to various workflows.

He does, however, caution that generative AI is not the end-all-be-all. It is more of an augmentation of human abilities. And companies should not add it willy-nilly. It’s not enough to arbitrarily place it into a process and claim you have generative AI. Partners should be able to determine when and if AI is needed to enhance a process.

Some partner organizations have already done as much, deploying generative AI in the partner prospecting process to narrow down which partners are best suited to work with them. But what’s all the hype about? Wallace says that generative AI could prove consequential as it is a merger of data science and speech science.

“The two have traditionally been separate roles when you think about it,” Wallace said, “and applying what we know about speech science at a much greater scale will give us the ability to solve much bigger problems because of the way generative AI works.”

Partners Have to Convey the Value of Generative AI

Let’s say a retailer wants to build a database for customers to search for a size 11 shoe. Here is one instance of a savvy deployment of generative AI to create a conversation with the customer. Wallace believes this is where a partner could leverage generative AI to take that potential sales process to the next level.

“If you want to talk about features and functionality of generative AI, it could offer alternatives or add-ons, creating an upselling opportunity, etc.,” Wallace told Channel Futures.

Most importantly, in this example it’s the merger of customer data and data on inventory paired with generative AI that enables this kind of interaction to become a reality, he added.

Eliminating AI Bias Is a Must

Partners across the entire channel partner spectrum, Wallace maintains, will be responsible for consulting on the “responsible and secure way to connect customer data” and ensuring when generative AI gets used — that it gets done ethically.

There is the matter of bias and outright wrong answers. Data and the source of that data are vastly important when looking to eliminate said bias.

“Don’t give it biased data and you won’t get biased output, as bias in outputs typically occurs because of a broad swath of information from potentially interesting sources, but it could produce misinformation because the source is poor,” Wallace said.

In the end, bias can find its way into generative AI algorithms in various ways. One of the most notable ways is biased human decisions. There are elements such as gender, race and sexual orientation that influence bias as well.

The consequences of AI bias are visible in the real world, with Amazon nixing an AI-fueled hiring tool because it found it presented bias toward female candidates. Though we still have work to do in perfecting generative AI, the early results of what we have/can achieve with the tool are promising, Wallace maintains.