With Generative AI garnering so much attention lately, it’s clear that AI has become the leading technology trend for 2023. That attention, however, reflects a healthy mix of anticipation and apprehension, not just for consumers, but also for businesses. For the latter, AI is mainly viewed as a path to automation to improve operational efficiencies and drive cost savings.

While there are good reasons for businesses to embrace AI, the technology is evolving at a rate that is faster than our ability to keep up. To some extent, IT leaders must move along this path to keep up with competitors, but that also comes with risk. Aside from AI not being well-understood, the track record is short, making for less-than-ideal conditions for adopting this technology.

That said, contact centers have a particularly pressing need to modernize, and AI offers many opportunities. As contact centers migrate to the cloud, AI applications become more feasible, especially those related to conversational AI (CAI). This is the underlying technology behind chatbots and will be a key enabler for improving self-service and easing agents’ workloads.

All contact centers want to be doing this, but they are proceeding with caution. To better understand why, this blog post outlines the dynamics around deploying AI in the contact center and three factors that are holding them back. The next post will outline the key benefits of AI in the contact center.

Unclear Use Cases

The starting point with AI is to recognize that this term encompasses a family of technologies that are all designed to emulate human behavior in an automated fashion. At a high level, automation and cost reduction are great use cases for AI, but that’s too generic to build a business case and deployment plan around.

Most contact center solutions are purpose-built (IVR, ACD, CRM, call recording, call routing, etc.) so the use case is very clear. For IT leaders rooted in the legacy world, this is how most technology was bought and deployed. With the shift from hardware to software to cloud, this process has changed. AI represents an entirely different form of business value.

The technologies associated with AI, such as Machine Learning and Natural Language Processing, are more like a tool set from which purpose-built applications can be developed. Rather than “deploying AI” based on the potential to automate customer service, the real value comes from specifying a problem set that could benefit from automation.

This is a new way of thinking for contact center leaders. They are now able to set the agenda with vendors rather than wait to see what vendors have to offer. Being highly adaptable, AI-based technologies present a new opportunity for contact centers to be proactive with modernizing, but only if they are ready for it.

Distrust in AI

Even when clear use cases for AI have been identified, trust looms as a bigger issue, both for those deploying and using AI-based applications. According to a global study report from KPMG and The University of Queensland Australia, only 39% of people across countries say they are willing to trust AI systems, with 29% report to feeling unwilling to trust AI.

The term “AI” has many connotations, and no matter how promising an AI-driven future looks, there is apprehension amongst even the most tech-savvy of people. Fear of the unknown certainly applies here, not just because AI is just becoming mainstream now, but also because very few of us have any meaningful understanding about AI. As much as we’d like to think technology is benign and judgment-free, AI is evolving to become human-like in ways we’ve never experienced.

This form of trust is existential, but there are also fundamental trust concerns around the efficacy of AI, without which it has little utility. Until very recently, chatbots were poor substitutes for human interaction, making it difficult to trust AI-driven chatbots with your customers. Recent advances in conversational AI (CAI) have greatly improved the chatbot experience, making it much easier to trust AI in the contact center.

First impressions can be difficult to change, and that’s a key challenge holding back AI in the contact center. Initial offerings didn’t have enough accuracy or knowledge base to automate customer service in a meaningful way. Since the operational challenges for contact centers haven’t abated – outdated technology, rising volumes, and a shortage of agents – the promise of AI remains strong.

Sooner or later, contact centers will need to embrace AI. While it won’t be perfect out of the box, it does improve over time, as will its value for improving customer service. As such, a different approach to the trade-offs that come with trusting AI might be needed.

Frustrating User Experiences

Aside from thinking about how AI will impact your operations, you need to consider how the user experience (UX) will drive adoption. No matter how determined contact center leaders are to reduce costs and enhance automation with AI, the results won’t materialize if agents and supervisors struggle to adopt AI-driven applications.

Just as customers get frustrated with poor chatbot experiences, the same applies to agents and supervisors who are using similar applications to augment their performance. Consider when chatbots base routing decisions on incomplete or inaccurate customer information.

There’s a ripple effect on humans when a poorly-trained chatbot makes one bad decision. Agents get frustrated trying to handle calls they’re not equipped for, then supervisors must do more work to properly route the call. All the while, the customer is now having a bad experience and longer wait times.

Part of building trust as AI applications improve involves having a better internal UX. To date, this UX has not been great, and that translates into yet another holdback for AI adoption. However, just as customer-facing chatbots – and other AI-based applications like sentiment analysis – keep improving, so will the UX for agents and supervisors. Again, it won’t be perfect right away, but as contact center leaders build trust with AI, and are ready to adopt new technology, concerns about UX should subside enough to not be a major barrier for deploying AI.