Over the last several years, an acronym and a buzzword have permeated many contact center technology discussions – AI and omnichannel. For the contact center, AI promises to add context to customer conversations, improve agent productivity without requiring more staffing, and more. The allure of omnichannel is that it can make managing various customer channels easier with a single screen. While these terms represent different contact center concerns, they’re often paired as essential components of a forward-thinking customer experience (CX) strategy.
But like every technology that came before, IT decision-makers need to do what they do every time they are presented with new technology – they need to take a step back, sift through the marketing messages, and determine if the technology can truly provide business value to their organization. In the case of contact centers as a service (CCaaS) – which often tout both AI and omnichannel features as their competitive differentiators – technology leaders have to assess how each of these technologies will change operations.
In the first part of a multi-part series on migrating your contact from on-prem to a contact center as a service (CCaaS) offering,
I explored several of these key considerations, including understanding what features you need and want, deciding on a public or private cloud, assessing service reliability and availability, and more. In the second part of this series, we focus on how enterprises should factor AI and omnichannel into their CCaaS purchase decision.
Sorting Out the Marketing Hype, Understanding Limits to AI
In addition to being marketed as possessing game-changing CX capabilities, AI and omnichannel tools are often touted for ease of set-up, promising that agents can use the service straight out of the box. But according to
independent contact center consultant Nerys Corfield, contact centers and their agents using AI and omnichannel is seldom as simple and straightforward as vendors would have you believe.
“The hype is predominantly coming from the CCaaS vendors who sell these things within their solution. The vendors do have good reason to try and underline the value – they just have a misconception that you bring the tools, and the people and process will fall into place,” Corfield said.
While AI has provided improved business results in industries like manufacturing, healthcare, and cybersecurity, Corfield said, "the execution of AI within customer service experiences is generally poor." Many enterprises have rushed to put AI in, and they often do so without having a great foundation or the time and effort to improve AI models over time, she says.
"AI is not a plug-and-play tool," Corfield said.
One AI technology that is often deployed poorly is natural language processing (NLP) and natural language understanding (NLU), which is often used as the basis for chatbots, she said. While chatbots can be useful with specific and closed queries like finding out opening and closing times or how to make a payment, they often run into issues when a customer request is more of an open-ended or conditional query. "Chatbots ... are not intelligent," Corfield said.
However, Corfield says "great strides [are] being made" around sentiment analysis, which can help contact centers identify areas where customers are getting frustrated in the customer journey. "Hopefully, the increase in sentiment analytics will mean a decrease in the incessant and generally pointless
customer satisfaction (CSAT) surveys that businesses so heavily rely on for a much-needed customer experience metric to present to the board," Corfield said.
Contact Centers Face Barriers to Support Omnichannel
Similarly, while omnichannel has been marketed for roughly a decade "as the panacea of effective transaction management" and "marketers have been striving for a single customer view ... forever," Corfield noted that few contact centers are set up for voice channel and digital channels to funnel into the "CCaaS automatic call distribution (ACD) and out in a uniformed fashion." While several obstacles exist to fully utilizing omnichannel, Corfield highlighted three crucial ones:
- Contact centers aren't typically staffed for omnichannel operations: Many contact centers recruit, train, and manage agents to either respond to customer requests via voice or digital channels, Corfield explained. Since there are often two teams that tackle different channels, enterprises don't "see the value of having all interactions route through the same funnel," she added.
- Lack of a business case for omnichannel and contact center together: While the "washer-dryer combination makes complete sense [for] most households," the omnichannel platform and contact center service combination isn't as simple, and there might be a harder business case to purchase them together, Corfield noted. Even though there is often a discount for combining the services, omnichannel capabilities can be seen as an unneeded (not business-critical) expenditure.
- Getting that “single” channel can be difficult: Historically, email communications are routed directly into a CRM platform (like Salesforce), and social media channels have been historically managed by digital marketing tools and then kind of became a “bastardized” digital interaction tool (like Hootsuite), Corfield said. This makes it difficult to route all into a single channel that agents can monitor.
Before deciding on any omnichannel capabilities, enterprise IT leaders should first confirm that their contact center organization is set up in a way that can actually utilize the capabilities or else the enterprise runs the risk of spending money needlessly.
Factoring in AI, Omnichannel: Focus on the ROI
While vendors will try to woo prospective customers with capabilities like AI and omnichannel, Corfield echoed a point made in the first part of this article, that enterprises should focus on ensuring that the product has a solid foundation: "First and foremost, make sure that the routing engine, the machine interface tools, the user interface, and the system administration parts meet your current and potential needs."
If those foundational elements are being addressed, then IT decision-makers might want to explore how a CCaaS vendor is choosing to direct their research and development – are they making investments in bringing more advanced AI and omnichannel capabilities to market? While it's helpful to understand how a CCaaS vendor typically invests in R&D, Corfield pointed out that R&D budgets can be spread across multiple product lines, so it's important to ask questions about how the money is being spent.
“Vendors move quicker than most contact centers do in their transformations, which means the gap between functionality [that is] available and functionality [that is] live is probably getting greater and greater,” Corfield said. “Make sure you don’t get swept away with the shiny bits if you have no use case to suggest you will ever use it.”
When factoring AI and omnichannel into a CCaaS purchase decision, enterprise IT leaders need to first evaluate these capabilities and assess if they can bring business value to their organization. They also need to ask themselves two key questions: How can these capabilities address my CX needs in the here and now, and how might the service address them in the future? By answering these questions, enterprises can ensure that their IT dollars are being used wisely and not adding capabilities that’ll never make it into an agent’s workflow.
In the third part of this multi-part series on migrating to CCaaS, we will explore how to prepare contact center agents and IT departments for a migration.