How Intelligent Voice Agents Can Replace Costly Contact Centers

Cloud-based call and messaging automation solution provider Inference Solutions recently announced Inference Studio 6.3, a solution that now extends the self-service capabilities of Cisco Unified Communications Manager (UCM), Cisco Unified Contact Center Enterprise, and Cisco Unified Contact Center Express (UCCE/X). The announcement coincided with Cisco's 10th annual Cisco Contact Center Summit, which took place in Hollywood, Florida from Sept. 17-19, 2019. At the Summit, attendees gained insights into the latest contact center industry trends and best practices. By using Inference Solutions' tools within these Cisco solutions, businesses can create their own intelligent voice assistants (IVAs) or chatbot components for conversation-driven contact center applications.

Continue Reading Below

"Inference will enable Cisco [on-premises] users to add artificial intelligence (AI)-enabled self-service without the need to 'rip and replace' their platforms," said Callan Schebella, CEO of Inference Solutions. "They can continue using their existing UCCE/X contact center seats for live agents while deploying cloud-based virtual agents managed by Inference. If you are a [premises-based] Cisco company, you can now make use of this technology and migrate the rest of your co-handling platform to the cloud at a later point—and still use virtual agents in the cloud as well."

Intelligent voice agents (IVAs) created using Inference Studio 6.3 can help companies automate various repetitive conversations handled by live agents. For small to midsize businesses (SMBs), the solution can provide an affordable option to expensive contact center solutions.

Voice Assistants Are Preferable to Humans

We now live in a world where many consumers are increasingly preferring to interact with voice assistants rather than with humans in contact centers, according to the Capgemini Research Institute (CRI), Capgemini's in-house research center. In its April-May 2019 Smart Talk: How organizations and consumers are embracing voice and chat assistants study, the CRI surveyed over 12,000 consumers using voice/chat assistants for products and services in the Consumer Product and Retail, Retail Banking and Insurance, and Automotive sectors. The CRI found that conversational assistants are going mainstream and are widely accepted, with voice assistants as becoming the preferred way for respondents to interact "across the consumer journey."

CRI's research shows that "customers increasingly prefer to use voice assistants." But the uptick in interest isn't just from consumers; the same report cites that "76 percent [of organizations] have realized quantifiable benefits from their voice and chat initiatives" and "58 percent [of organizations] say that these benefits met or exceeded their expectations."

There's still a lot of room for growth, especially since less than 30 percent of the companies surveyed showed "conversational interface maturity," potentially having them miss out on an opportunity to better serve their customers. The emergence of and preference for voice assistants follow Statista's growth estimates for enterprise virtual digital assistants (VDAs) (see chart below).

Self-Service Creation of IVAs

Deploying Inference Studio skips a lot of steps that were traditionally required to create IVAs or text-based chatbots. In the Cisco environment, the traditional methods required various steps. Businesses would first need to run Cisco Unified Customer Voice Portal (CVP). Inside that, they would have needed to build their conversation experience from scratch by using Voice eXtensible Markup Language (VoiceXML), adding speech recognition engines such as Nuance, as well as build an entire library of professional services.

"To build a conversational-style experience from scratch [took] anywhere from six to 18 months," Schebella said. "Now you can build a conversational experience in just a few minutes."

Within the Inference Studio user interface (UI), users can start with a blank canvas, select from ready-made tasks from a library, copy tasks, or import pre-existing tasks. It takes just a few minutes to select the type of voice assistant or chatbot solution you want and then program various processes, all without writing a single line of code. Similar to the way consumer voice assistant technologies, such as Amazon's Alexa, use various "skills" to add new functionality, Inference Studio gives users a range of tasks and solutions from which to choose within its internal UI. The same goes for text-based assistants or chatbot solutions.

"If you pick a task that has been built by someone else, then you get to benefit from the learning that they have programmed into or designed into the interface," Schebella said.

While Inference Solutions makes its solution intuitive for all levels of users, it also offers a brief Udemy course called Inference Solutions Certified Associate as a deeper dive into IVAs and text-based bot creation. The course was created by Roger Venning, Chief Operating Officer (COO) and Vice President of Customer Success at Inference Solutions. In the course, you will learn how to design and implement an interactive voice response app on Inference Studio.

"The vast majority of our traffic is in the voice channel because of the nature of our partners, and so we offer voice channel first," Schebella said. "And then we offer other modalities as a secondary and tertiary channel; these are channels like Whatsapp, SMS texting, or in-website chat pop-ups."

How much would it cost to onboard a virtual agent from Inference Studio? "You can generally expect the retail price of a feature-rich agent to be in the realm of about $400 per agent per month as a single fixed fee," Schebella said. "So, if you had 10 agents, they would be $4,000. These are the retail prices; wholesale prices would be a lot less than that."

Have any questions you need answered about enterprise voice assistants? Join the PCMag@Work business community on LinkedIn, and you can ask vendors, other professionals like yourself, and PCMag's editors.

This article originally appeared on PCMag.com.