As Artificial Intelligence Advances, Is Human Error Holding Up the Contact Center?
LiveVox’s CEO discusses the challenges and promise of AI for the industry.
August 19, 2022
![Partner Hurdles Adopting AI Partner Hurdles Adopting AI](https://eu-images.contentstack.com/v3/assets/blt10e444bce2d36aa8/blt5a25021e8074093a/65241b9257fabd4182406154/Artificial-Intelligence.jpg?width=700&auto=webp&quality=80&disable=upscale)
Shutterstock
CF: What are three areas AI performs well in a contact center environment?
LiveVox’s Louis Summe: AI does well on things that are standardized and relatively straightforward. AI doesn’t do as well on complex, very broad topics. For example, if it’s a very narrow topic or question such as booking a hotel room or making a credit card payment or changing a schedule, those kinds of things are very short, defined and specific and narrow in scope. That’s really where AI does best. So, when you’re thinking about an AI application, the first thing you must think about is how much effort and time will it take to pull the data together to make the AI application work. That’s No. 1. But then No. 2 is, what are the functions that are best done by AI? And then No. 3 is, how do you seamlessly route the question to a contact center agent when it’s not a function that’s best done by AI?
CF: How does a customer seamlessly transition from AI to a contact center agent?
LS: We primarily do that by building better knowledge into the AI routing engine. The AI has to not only be strong at servicing the customer, but also when it gets to the point where it needs to be escalated to an agent. The AI needs to be strong at picking the right agent to send information to. There also has to be a seamless flow into the agent’s workstation. Those are the things that are very critical to get right for the overall experience to work well.
CF: What factors make AI expensive?
LS: The question for the contact center now is, how much money do I have to spend to get AI into a highly functioning state? The real benefit of AI is it reduces the amount of time that agents have to spend. But if you’re spending a lot of money, then that benefit is not big enough to account for R&D expenses. So it’s critical that AI be easy to deploy and work without a lot of integration. Otherwise, the cost of deploying it just goes up.
CF: Sometimes AI does not address the needs of contact centers. How much of that is due to human error versus gaps in the technology?
LS: People envision that AI is going to be a silver bullet and it’s kind of just magic. However, It’s an application tool. You have to use it well; otherwise, you won’t be successful. And then the second thing where most people trip up is they don’t view AI as part of a holistic journey. They just view it as this self-contained application. It’s got to flow easily. The application has to be designed [and] integrated with what comes next. And that’s the agent workstation. Applications are just modules. They’re not necessarily designed to be part of an overall customer journey. But if they’re not part of the overall customer journey, then they won’t operate in an ideal state.
Think about your own experiences as a consumer, [where] you call into a contact center and you have some sort of an AI application that you’re working with. You give it information and you get to a certain point where that application no longer serves whatever needs you’re trying to get accomplished. And then you transfer to a live agent. Have you ever experienced the frustration when all the information you gave the AI tool you have to repeat all over again? Our tool seamlessly moves from the AI virtual agent to the live agent, and it keeps all that information. So, it’s a continuous flow.
Channel Futures: What differentiates your product, CCaaS 2.0, from other products on the market?
LS: The biggest differentiator is the ease of use and the timeline to deploy. We accomplish those differentiators because we not only have the AI applications embedded and pre-integrated into our platform, but they’re also pre-integrated with our contact center CRM. That’s an application that our customers use for pulling together data from a lot of different places and then put together into one area. Most companies run multiple CRMs. And our contact center CRM allows them to easily pull data in and out of it, putting it together in one spot. Also, that comes pre-integrated with the AI applications. That means when you get started using AI applications, a lot of the integration work that you must do to compile all of the data that the AI tool would use is already done. So that’s a big advantage for getting started and for speed of deployment. Also, the CRM attached or pre-integrated to the AI allows you to do personalization with customers.
CF: According to the studies that you’ve featured on the LiveVox website, over 50% of contact center decision makers say that they are unable to make the changes needed to machine learning models and AI tools, because those depend on vendors. Decision makers have a lack of staff and expertise to deal with it themselves. How is LiveVox addressing these concerns?
LS: We deploy our virtual agents inside the overall contact center platform. Therefore, as a contact center operator, as a manager, you can manage your virtual agents in the same way that you can manage your live agents. If you don’t manage it, you know, you shouldn’t expect it to be optimal in its results. We expect that managers are going to be monitoring and tuning to their virtual agents in a way that’s analogous to what they do with their live agents. And we enable them to do so.
CF: Less than half of the respondents in one of these studies experienced benefits from implementing AI tools. However, 80% still agree that pre-integrated AI solutions can provide greater flexibility, deploy, test and change applications quickly. How are your applications pre-integrated and how do you define that?
LS: We have a higher degree of pre-integration than our competitors do. And that’s why our CRM is so powerful because it pre-integrates the data, enabling that personalization and customer experience. It also makes the AI application smarter. I think when you look at the 50% that are saying the AI application didn’t work, I’m sure that you would find that a lot of them were spending much of their time and energy trying to launch, trying to get the integration to work, trying to get the data to the AI application. Whereas if it’s pre-integrated, you can take those tasks off the table and you can just focus on tuning the application, which gives you a much quicker path to getting up and achieving the benefit.
CF: On your website, LiveVox refers to practical AI and how it should be implemented in contact centers. I wanted to know what it means to be ‘practical’ for AI? And why is it beneficial?
LS: I think practical in this context means that the No. 1 driver is a strong ROI. That really boils down to how much it costs to get started. It’s a big part of the practical. Another part of practical is that it’s easy enough that the contact center operators can not only launch the AI, they can manage it after they’ve launched it. Those two things are important pieces of making it easy to implement so that it can be done by your employees. It can be done cost-effectively, you can get the ROI that you’re looking for.
CF: Is there room for improvement with how AI handles the routine or simplistic tasks? And where do we see that technology headed?
LS: I think that over time, as AI applications get more and more traffic on them, they get smarter and smarter. And as they get smarter and smarter, you can use them in more places more easily. It’s important to tap this kind of capability by having your discrete tasks defined. And then you want to accumulate a library of functions that solve for a wide variety of discrete tasks. So, if your AI application can handle 10 discrete tasks, it’ll be worth 10 times more when it can handle 100 discrete tasks. That’s a really big part of the maturity cycle. As AI applications build up their functional library of discrete tasks, they can be deployed to make it even more purposeful.
CF: When it comes to ROI, you may not see the return initially. However, if you work with the technology over a couple of years, you might see that return. You would agree?
LS: Exactly. And that’s why getting up quickly and easily on an initial project is so important, because you may only be running two or three modules. As you start to add more modules, it really begins to deliver value for you over the long haul. A lot of times organizations deem it a failure when they spend a lot of money to get the first two or three modules. They throw in the towel before they’ve really had a chance to kind of go down the path with AI and develop it over the long haul, which is ultimately what they need to do to be most successful.
CF: What areas of AI do agents have the most difficulties with? And how does LiveVox try to solve for those? Additionally, where are agents having the greatest success with the technology?
LS: The single biggest challenge for agencies is getting easy access to disparate information systems that contain relevant data points to solve the problem that the customer has presented them. That’s a frustration for agents. And that’s also a blocker for using AI. With our contact center CRM ability, we can pull together data out of a wide variety of sources. We can make it a lot easier for agents; they don’t have to go through five or 10 different applications to find what they’re looking for. It also makes it easier to deploy an AI application.
CF: We briefly talked about the financial benefits of AI. As a follow up, does AI result in contact center savings over the long term?
LS: I think that for those that can deploy quickly the ROI is very strong. And as we’ve been discussing, over time, the ROI can continue to go up as centers add more and more functions.
CF: When only half of respondents say that they benefit from AI, how does that shape the products that you create or are creating?
LS: It inspires us to stay focused on making it easier and easier to deploy and to build a richer library of functions that can be pre-built and pre-integrated. This allows the contact centers to get up and going more quickly and get to a point of productivity faster.
CF: AI is changing constantly. I wonder how partners are affected when trying to sell the product?
LS: I think a lot of our partners are very excited about AI and about its potential. We work with them so that they’re aware of how our applications are evolving and what the opportunities there are with them. I think, in general, they’re thrilled about how it’s progressing. And we work together with them so that we’re aligned.
CF: AI is changing constantly. I wonder how partners are affected when trying to sell the product?
LS: I think a lot of our partners are very excited about AI and about its potential. We work with them so that they’re aware of how our applications are evolving and what the opportunities there are with them. I think, in general, they’re thrilled about how it’s progressing. And we work together with them so that we’re aligned.
Artificial Intelligence (AI) is Louis Summe’s forte. As CEO of LiveVox, a company with more than 20 years of cloud contact center expertise, Summe leads a team that develops AI and other technologies to improve the efficiencies of contact centers. Improving workflows is no easy task when contact centers have a revolving door of employees. Agent churn rates – or attrition rates – for contact centers range between 30% and 45% annually. That’s twice the national average for all other jobs in the U.S.
We recently compiled a list of 20 top CCaaS providers offering products and services via channel partners. |
However, Summe views these circumstances as a challenge. With fewer tenured employees at contact centers, employees that do remain can be saddled with additional work. That’s where LiveVox said its software can alleviate such problems. By implementing LiveVox’s AI technology to handle mundane tasks, contact center managers can redirect limited staff to handle more complex concerns, the company said.
Yet, artificial intelligence might not be the solution for everything that ails contact centers. In this interview with Channel Futures (see slideshow above), Summe talks about the promise of AI as well as the difficulties surrounding the technology for organizations. More specifically, he discusses how human failures implementing AI are the primary reasons for poor outcomes and what to do about it.
Want to contact the author directly about this story? Have ideas for a follow-up article? Email Claudia Adrien or connect with her on LinkedIn. |
Read more about:
AgentsAbout the Author(s)
You May Also Like