AI is evolving so quickly now, that an idea like this isn’t all that far-fetched. 2023 is being touted as The Year of AI, and ChatGPT has captured the public’s attention in ways few tech trends have done before. Generative AI is the latest wave of AI-based innovation – best exemplified by ChatGPT – and for better or worse, almost anything seems possible with this new form of technology. The contact center has long represented prime use cases for AI, but more education is needed to convey the use cases and benefits.
To that end, NICE has developed a four-part webinar series, titled ChatGPT Created This Webinar, and I was featured as a guest on the second episode, in conversation with John Willcutts, General Manager, Digital Line of Business at NICE. I’ll leave it to you to decide whether the webinar really was us, or just a really good video deepfake, with all the dialog being generated by ChatGPT.
Making that distinction is part of the anything is possible mantra that comes as AI gets increasingly better at creating human-like interactions and experiences. AI still has a lot of improvements to make, but barring some epic fail, there’s no turning back at this point. For contact center leaders, the point of this webinar series is to show how AI can enhance both customer experience (CX) and agent experience (AX), and as John Willcutts noted during our episode, the time is now. To explain why, here are the key themes we discussed during the webinar.
Theme 1 – Understand the building blocks
The starting point for understanding AI is to know what it is and is not. There’s still a lot of hype around AI, and it’s important to clarify that AI itself is not a technology; rather it’s an umbrella term for a family of technologies such as Machine Learning (ML) and Natural Language Understanding (NLU). Each building block plays a role in helping contact centers improve CX, and John explained that further with a visual model outlining three elements that define customer interactions – Semantics, Episodic and Procedural.
These building blocks have been evolving for some time, and we talked about how cloud and Moore’s Law have been key drivers to get AI to the point where contact centers can now use it at scale to automate both customer service and internal operations. The main focus for our webinar was on self-service, where automation helps improve AX by handling routine calls that take up a lot of their time.
John explained at length about the various parameters contact centers must set to ensure chatbots engage customers in the right way, especially around properly supporting the brand and addressing topics that are relevant to the business. Regarding the use of chatbots, I talked about how advances in conversational AI (CAI) have improved today’s chatbots from earlier generations. Not only are they more conversational with customers – i.e., more human-like – but they are much better at understanding intent, which pertains to the first building block John talked about – Semantics.
The net result is that chatbots are now able to handle more customer interactions, and that in turn creates new data sets that Machine Learning uses to fine-tune its algorithms. As these volumes grow, AI can better model conversations and customer interactions that align with the best outcomes.
While AI is probabilistic by nature, the value will become evident when deployed at scale. Not every interaction will have an optimal outcome, but the positive impact on CX will outweigh what the contact center can achieve without using AI. This in turn, builds trust in using chatbots, which will only further shift the workload from live agents to AI-driven automated self-service.
Theme 2 – State of play
For this section, we talked about the dynamics of the market, and how customer expectations have changed. Rather than the contact center being a passive destination that only responds to incoming calls, we talked about how it needs to be more integral with how customers engage with the product/service, as well as the brand itself.
Doing so requires a much richer understanding of each and every customer, and that’s nearly impossible to do with legacy technology. CRM may play a key role here, but there are many other data sets across the organization that are relevant to CX. This is fundamental to how CX is different from customer service, which is only defined by what happens in the contact center.
Today’s customer has multiple touchpoints with both an organization and the brand, and this is a key reason why legacy-based contact centers struggle with CX. They need to meet the customer where they are, and can no longer dictate the terms of engagement. Good CX means providing the same level of customer service regardless of where the customer is physically located, what device is being used, or what the preferred mode/s of communication is/are.
All of this points to the need to access data from across various company silos – it’s really the only way to get a complete picture of the customer and their journey. Every step along the way has distinct data sets, and AI draws from them all, both inside and outside the contact center. Legacy, premises-based platforms cannot do this, and once AI is viewed as being strategic to the business, cloud migration – including contact center – will become a necessity.
As such, John’s main message is that “today is the day to start your AI strategy.” Automation is generally viewed as the main driver, but it’s really about harnessing data, and there’s no better way to do that than with an AI-driven architecture. Silos may never go away, but contact centers need a way to pull relevant data from each silo and get into one place, where it can be used to drive better CX.
Theme 3 - Setting expectations
To conclude, we talked about the need to think differently about AI. The ROI model used for legacy technology isn’t really applicable here. There is no amortization timetable, and being cloud-based, it’s a consumption-based model. The costs will be ongoing, and while the immediate benefits may not be very impactful, it should be clear that as automation scales, the long-tail payoff will be substantial.
Along those line, the iterative nature of AI is another point of difference from legacy technology. AI applications are not fully formed out of the box, as would be the case with a PBX, IVR, ACD, etc. The initial utility of a chatbot will be limited, but as those data sets feed ML models, their capabilities will become more human-like, more trusted, and more effective for CX.
Finally, as CX leaders consider next steps, we talked about what they should need and expect from AI partners. They need to determine first if looking just for a point solution, or more of a platform-based offering that supports the entire contact center. The focus for type of applications is important as well; some will be specific only to operations, especially for supporting agents or supervisors. Others will only address CX, but if taking a holistic view with AI, both AX and CX really need to be considered.
For the bigger picture, we also noted the importance of overall vision. Each contact center will be on its own journey with AI, and you’ll need to make sure the AI partner’s offerings are aligned with where you are on your journey. If just starting out, you won’t need the most advanced offerings, but also if you’re ready to be bold, you’ll need a partner who can take you there.
