Blog | Nexcom

AI in CX

Written by Rolf Gordon Adamson | Nov 30, 2021 1:29:00 PM

AI in CX? What’s the difference?

At Nexcom, we work hard on making the best solutions available to support the needs of our customers and end customers. The way we do that, is by delivering an “intelligent” machine that is able to give the right answer or provide advice to the customer when needed.

Let us tell you about AI and why it’s efficiency.

AI, or more precisely machine learning, is a process that can facilitate both the lifecycle of customers, and organizations with its ability to focus on difficult processes simultaneously. It uses data points, lots of datapoints, to construct its potential in accuracy and usefulness, but it is selective in the data it needs. It’s not as much about the quantity, but how the data can be reproduced so that the machine can comprehend it.

To be a little technical, patterns need to be identified, and parameters set. Imagine leaving a box of LEGO blocks, without the assembly leaflet, in the hands of a 4-year-old. What happens if you are not helping and guiding the child? Well, the mess will be similar if you don’t give AI the guidance and support it needs.

AI needs to be fed directions constantly, in a constant reproduction of the data, over and over again. Fail to do that or feed it different data, it will not be able to recognize the patterns, which it depends upon to give you the results you are expecting. It’s mathematical algorithms which rely on “good” data to be effective.

AI needs to be told what’s considered “right” or “wrong”, and how to interpret the data we feed to it. Just like our 4-year-old, it cannot identify difference all by itself. The difference with our 4-year-old, is that once the AI has understood what you want it to, it will leave your human team behind on the starting blocks while it is crossing the finish line. It’s that fast!

iAgent

Now, let’s not panic! AI is not there to replace your human team! We build AI to support the team to the best of its ability, and limit the human error, to ensure end customers have the right information or support at the end of their call or email request.

Contrary to humans, AI is only able to tell you if something happened, but not why something did or did not happen. It is able to recognize whether the right steps were taken, which, in terms of quality monitoring, is invaluable to evaluate the success of interactions with customers.

As it recognizes keywords as well as sentiment, it will also be able to give hints or suggestions to agents, to help them avoid mistakes or find the right information to support this particular request or issue the customer has. And this is what everyone wants in CX and customer service! The right answer given to the customer, at the right time.

But AI monitoring should not replace your side-by-side monitoring or all monitoring done by employees, as agents still need training and coaching, as well as attention from their manager. AI will facilitate the training and work of the agents, and show if the training has had an effect on improving performance. If it does not know whether the agent knows the solution, it most likely knows what the answer should be and then the AI will leave the why to the organization to deal with. The AI can basically tell us that things happened and how we have dealt with them, but it cannot tell us why things happened. This still requires human analysis. But that still leaves organizations with lots of opportunities to avoid errors and ensure customers get their issues resolved and finish their interaction satisfied with the company.

The Good, the Bad, and the AI

There is a lot of hype about AI. Let’s be realistic, AI cannot do everything we think it can right now, it is only able to do what it has been taught to do. It has limitations and has to be controlled, guided. It is meant for specific purposes and needs to know its limits. As cool as it sounds, AI is not intelligent, it only looks at patterns, limiting it to how you deploy it and re-tweak it.

What you can rely on with AI, however, is the fact that it’s not human, and it’s not subjected to what influences human life. It won’t oversleep, it won’t get drunk, it won’t be dumped and come to work depressed or be sick, go to prison… it’s very dependable, reliable, and consistent.

AI will also beat humans on its ability to learn. Yes, it needs to be fed a lot of what it learns, but it will learn it in a matter of seconds. And the difference with us, mere humans, is that it will not forget what it has learned, as long as it is used the right way. Yes or no, right, or wrong.

It’s all in the math! A good algorithm is above 70% accurate. The first dataset we ever produced at Nexcom was 78% accurate – now we can produce up to 83% accuracy. At Nexcom, we aim for 85 or 90 to 95% accuracy and do not limit ourselves to 70%. 70% accuracy means 30% inaccuracy, and we do not like the idea that customers are left with 30% inaccuracy. We are aiming for very high levels because there are (actual) people receiving our solutions and AI should be used to enhance/boost the experience. If organizations today can train staff to deliver 85% issue resolution then we need to make an AI that can move beyond that and improve upon it. By using and combining datasets we are able to push our accuracy higher. This is where our experience and knowledge from the industry plays a significant role as we know where to look for valid data!

Magnum, A.I.

To put our ability with AI in more detail, with our Nextgen AI component you can:

  • Predict customer satisfaction and dissatisfaction
  • Predict possible complaints before they happen
  • Find particular words and patterns for business-critical analysis
  • Address issues before they turn into trouble
  • Detect specific elements within conversations and keep track of:
    • Mentions of competitors or their products
    • What is not working right now
    • What type of inquiries are there more of than usual
    • What words and combinations of words trigger negative reactions from customers

Nextgen AI within RevealCX helps you manage agents’ performance, customer satisfaction and conversion rates. But that’s not all…

Nextgen AI is crucial for its intuitive sentiment analysis which uses the tone of voice to determine whether a customer is positive, negative, or neutral, based on how they speak. To identify the sentiment, it looks at what’s hidden behind the words, such as pauses and breaks, pace, agitation, and intonation. 

So, there it is! AI is a man-made intelligence fed to a machine, that is meant to monitor agents and offer them support, in suggesting the right answers and service when requested by the customer.

It is not here to compete with humans, it can’t, but it is a very strong tool to support the CX in a company that relies heavily on customer support and its call center to be competitive.

Here you can find out more about Nextgen AI and how we can provide you with the capabilities to drive customer experience!