There’s a lot of interest and excitement around generative AI in the contact center. It appears that generative AI is poised to remake the contact center in 2024. There’s a lot of things that will change regarding customer service, including roles, reporting, onboarding, and even customer expectations. But I’m not convinced this impending generative AI revolution will actually improve customer satisfaction (CSAT).
Generative AI refers to a broad set of AI capabilities powered by large language models (LLMs). They offer a more human-like conversational experience than prior bots, but this tech is about more than bots. Without doubt it will clearly have a significant impact on the contact center. For example, generative AI will likely reduce the training times necessary for new agents.
But let’s not confuse technologies with results. Generative AI could improve customer satisfaction, but not necessarily — it’s just a tool. First, there’s the issue of how it gets deployed, and then there’s the reality that not every contact center wants to improve CSAT — sacrilege, but it’s true. Then there’s the issue that if everyone implements the technology (and they will), advantages that generative AI brings become table stakes.
We’ve seen this new tech-in-the-contact center picture before. Consider the IVR, and then press one if it improved your customer satisfaction. It certainly may have and probably did, but there are more examples of it being misused and abused. CSAT doesn’t improve with new technology, it improves when new technology is implemented to properly to achieve it.
Now, let’s talk about the uncomfortable fact that many contact centers don’t really want to improve CSAT. The CCaaS providers love to talk about the importance of customer satisfaction, and they all have examples of how great customer service offers generous returns on investment. Yet, great customer service remains elusive, and, generally speaking, contact centers are not often equated with great customer service.
No one believes that call volumes are currently higher than usual. No one believes their call is really important to the vendor. No one needs to be reminded that faster service might be available on the website. The misuse and abuse of deflective practices, limited digital capabilities, and neutered self-service have taught consumers and B2B buyers to be resistant and skeptical of modern customer service.
There are lots of reasons this occurs. To start, excellent customer service is actually harder to deliver than it sounds. Additionally, many businesses can get away without it. Perhaps they have no viable competitors, perhaps they are protected by high switching costs, or perhaps it is bad leadership — whatever the cause, we all know that great customer service is the exception, not the rule.
It’s an uncomfortable reality, but most contact centers are not designed to optimize customer satisfaction. They are more likely designed around costs and efficiencies. A simple example is digital channels such as chat, email, and other apps. In 2023, the benefits and capabilities of using digital channels are still barely tapped.
Many of us prefer to use digital channels rather than making a phone call to obtain service. Digital channels offer customer benefits, such as asynchronous communications for the multi-tasker, inherent note-taking for the documentarians, silence for meeting participants, and easy social amplification for broadcasters.
It gets even better: digital channels are cheaper and more accessible for the enterprise too. Contact centers tend to favor tried and proven technologies, such as the phone. It is taking decades for contact centers to embrace digital channels, and generative AI is still in the cutting edge stage. Generative AI is game-changing, but that doesn’t it mean it will change how contact centers engage with new technologies.
To be clear, the contact center has been embracing incredible new technologies, and arguably can do more with fewer people than it could before. The contact center has been consistently improving for decades. The list of truly impressive tech in the contact center is long, with big hits like skills-based routing, Workforce Engagement Management, speech biometrics, the IVA, and so many more. But that’s not the same as improved customer service.
I do expect that generative AI will change the cost of customer service. It’s not hard to extrapolate where this tech is headed: We will see more calls handled by self-service successfully, agents will get better assistance and coaching, and supervisors will have it easier seeing what’s occurring and implementing changes. Without necessarily improving customer satisfaction, the cost of delivering it will be reduced.
We would like to think that instead of budget cuts, enterprises will use those savings to improve CSAT. That would be nice. It doesn’t usually play out that way. Generative AI does not change the basic equation: Great customer service is expensive, and it’s up to every enterprise to find the right balance of customer service and cost. I expect that most proposals to implement new generative AI technologies will involve cost cuts to pay for it.
We are heading into a new, explosive chapter in AI. We have also seen research and development shrink along with headcounts. I fear that in most cases, generative AI solutions won’t raise the bar of CSAT. Instead, it will be justified to reduce expenses with the goal of maintaining CSAT.
Dave Michels is a contributing editor and Analyst at TalkingPointz.
