Setting Realistic Expectations with Contact Center AI

This year has been widely viewed as the “Year of AI,” and that certainly seems to be the case for contact centers. Every vendor has built artificial intelligence (AI) into their offerings, and every contact center is looking to AI as a solution to many of their challenges. Expectations are high; the hype is in overdrive — and vendors are more than ready to help.

This blog was written by Jon Arnold, Principal of J Arnold & Associates. 

This year has been widely viewed as the “Year of AI,” and that certainly seems to be the case for contact centers. Every vendor has built artificial intelligence (AI) into their offerings, and every contact center is looking to AI as a solution to many of their challenges. Expectations are high; the hype is in overdrive — and vendors are more than ready to help.

With things changing so quickly, contact center leaders are in a difficult position to make important decisions about technologies they know very little about. There are many paths for adopting AI, and this post will provide some guidance for taking those important first steps.

As a starting point, it’s clear that legacy, premises-based deployments aren’t sufficient for bridging the gap between how customer service has typically been provided and what today’s digital-centric customers expect.

The Promise of AI

As a computer science discipline, AI has been evolving since the 1950s. But only with the recent advent of cloud computing has AI become relevant to the contact center. While today’s capabilities are impressive, they are still nascent, and a long way from solving every customer service issue. To get a more realistic grounding, contact center leaders should first consider the basic tenets for what AI is and what it is not.

What AI Is

Many vendors make liberal use of the term “AI.” And because it has high perceived value, buyers might be quick to move toward anything labeled AI. In reality, AI is a very broad term, and it’s not a specific solution along the lines of an IVR or ACD.

Contact center leaders don’t buy “AI;” rather, they invest in a family of “smart” technologies that leverage today’s digital technologies. In this context, AI is more of an umbrella term for a family of technologies that enable smart solutions.

There are numerous technologies to consider, with leading examples including machine learning, natural language understanding, conversational AI and generative AI. For contact center leaders, the important point here is that AI is not itself a solution. Due diligence is needed to understand what the underlying components are for each vendor.

There is a wide range of capabilities among the vendors, whereby some have developed considerable in-house expertise for these components while others rely instead on third-party partnerships.

What AI Is Not

This brings us to the converse scenario, where “AI” is somehow viewed as a solution that can simply be deployed plug-and-play style without further consideration. In cases where AI is being fast-tracked, it behooves buyers to get past this seemingly virtuous “AI” label and better understand what the constituent components are behind it.

Given that AI is largely unregulated, vendors have free reign to apply this label and charge a premium — even though they may only be applying a nominal amount of proprietary AI. To mitigate against this, contact center leaders need to find out what elements of AI are actually being used, and how each element actually brings new capabilities.

Taking this a step further, it’s important to not view AI as a silver bullet to address all your customer service issues. While AI can potentially make everything better, this can only happen with specific use cases where AI is applied to clearly defined problem sets. In other words, contact center AI works best for situation-specific needs rather than a horizontal deployment that supports a range of customer service needs.

Core Contact Center AI Use Cases

This brings us to the crux of what defines value for AI — use cases. The more specific the use case, the easier it is to define value and set expectations. Since AI is really a framework for a variety of underlying technologies, there is a near-limitless set of possible use cases, which means that contact center leaders shouldn’t feel constrained about where AI can be applied.

It’s certainly good to have choice — and AI should be viewed as an opportunity for taking more creative approaches to customer experience (CX) — not just as a way to make incremental improvements to how things have always been done. However, this puts the onus on contact center leaders to think more broadly about adopting new technology.

This can be challenging when they’re not sure where to start with AI. To make this process more manageable, here are three core use cases to consider.

Cost reduction

This is a universal business driver, and a great starting point for AI. Labor is the biggest cost component for contact centers, so this use case will resonate not just with contact center leaders, but also senior management. Important as this objective might be, you must carefully think through the specific applications of AI.

The path of least resistance would be to simply reduce agent headcount, but that will only be effective if AI is also deployed in other ways to keep service levels high with fewer agents. There are many use cases for that, and they generally center on operational efficiencies, such as using chatbots to improve self-service, or workforce management to better align staffing levels with varying levels of call volumes. As such, cost reduction should be a core use case, but not in isolation from everything else needed to provide great CX.


This is truly the North Star for AI, as the focus of these technologies is on managing tasks and processes that have previously only been handled by humans. Not only is AI increasingly capable of doing this, but it does so at a scale and speed that humans simply cannot match. For most contact centers, the initial automation use case would be chatbots, as this is a well-understood pain point.

Legacy-based IVR has long been the standard for self-service. But as customers shift to digital channels, this technology is just too limited for today’s CX requirements. While it’s true that first-generation chatbots haven’t been much better, recent advances in conversational AI have greatly improved their ability to interact with customers.

These chatbots may have a long way to go for handling end-to-end complex situations, but they are being used now to manage meaningful volumes of inquiries and reducing the need for agents to handle simple requests. Beyond chatbots, it’s important to note there are many other use cases for automation, especially around workflows and intelligent routing. So self-service is really just the starting point here.

Agent Experience

For contact centers embracing the CX concept, the move to digital and cloud-based technology is essential, but that’s not the whole story. Technology alone doesn’t translate into better CX. And when it comes to AI, there’s more to consider than cost savings and automation. In essence, customer service is about interactions between people — and this is where agent experience should be another core use case for AI.

Some forms of CX can be fully automated, where no agent engagement is needed, but that’s really the exception, not the rule. No matter how good the tools, CX won’t be good if agents aren’t fully engaged, and for many contact centers, that’s an uncomfortable reality. Many agents are chronically overworked, and often have sub-par tools that make it even harder to provide good CX.

Contact center leaders are increasingly coming to understand the direct connection between agent experience and CX. There is a multitude of AI-driven applications that make for a better agent experience, such as noise suppression, automated call summaries and sentiment analysis. For contact centers that view human interaction as paramount to CX, the agent experience could be the best use case for AI — not just for customer service, but for maintaining a strong team of agents.

Realistic AI Expectations Start with Outcomes

AI-based technologies are still quite new. Being complex by nature, they’re not easy to understand. Contact center leaders aren’t data scientists; rather than focus on the inner workings of AI, they should instead think about the outcomes they’re trying to achieve. In terms of setting expectations for AI, identifying those outcomes should be the starting point. Deploying “AI” is not a checkbox item for modernizing the contact center — and it’s not a point solution that runs on its own once implemented.

The decision to deploy AI should be viewed holistically, as there will be benefits that extend beyond the contact center that will impact the overall organization. Equally important is the fact that AI is constantly evolving, and while it won’t be perfect from the start, the benefits will accrue as usage increases. Data is the oxygen that drives AI, and as the data sets grow, the outcomes will be more accurate and more precise.

These are the capabilities agents need to feel empowered, and for your customers to experience more personalized service. You’ll need patience to get there, but the results will be worth it.

For contact center leaders, this will require different expectations from investing in legacy systems. And it will be doubly important to work with technology partners who understand those expectations — and know how to effectively support them.

For more customer experience insights, visit J Arnold & Associates online.