It has been well-documented on this site and others that customer experience, or CX, is the No. 1 brand differentiator.

Every contact center-as-a-service vendor leads with this when talking about reasons why contact centers need to be modernized. My research shows that today, 90% of businesses compete on customer experience compared with only 24% five years ago.

Another interesting data point from my research is that two-thirds of millennials admitted to dropping a brand after a single bad experience, which shows how damaging a bad contact center interaction can be. I understand that not all good or bad experiences happen in the contact center, but that’s often the first touch point for many interactions.

The challenge for most businesses is how to measure CX. Many companies use customer satisfaction or CSAT metrics, but these can be hard to collect. Businesses often send out e-mail surveys or often ask people to stay on the phone to answer a couple of questions. Although this has been the way for decades, it isn’t all that effective, since only a small percentage of people respond.

Dialpad recently came up with a better way to collect CSAT data by using artificial intelligence to automate this data. The company last week launched a product called AI CSAT, which is a real time, predictive engine to calculate customer satisfaction. The new tool is part of Dialpad’s Customer Intelligence platform, analyzes 100% of calls and produces data that companies can use to improve CX.

When gathering CSAT surveys, companies typically run into issues with customers either not responding or not providing their true opinions. Dialpad research quantified the response rate of CSAT surveys to be a woeful 3%. Additionally, the responses tend to be either positive or negative, but not much in between, which can often skew the data.

Dialpad created AI CSAT to infer CSAT scores automatically from every customer interaction using machine learning algorithms. The data is gathered from 100% of calls, so contact centers can improve how they measure customer satisfaction. Even if a customer doesn’t fill out a post-call CSAT survey, their interaction with an agent still produces customer feedback.

Using the tool, agents can better understand customer sentiments from voice conversations, without inconveniencing the callers. Dialpad’s AI model is trained on anonymized call transcripts generated by its proprietary speech transcription technology to determine customer satisfaction. The AI model learns from conversations and tailors the results to a company’s specific industry or vertical.

Dialpad said AI CSAT constantly updates its more than 100 million parameters based on new conversations, much like the human brain. Since the AI model becomes more accurate over time, contact centers can continue to make improvements to their CX. For example, contact centers can use the call data to follow up with customers that haven’t filled out a survey but have low customer satisfaction, or coach agents to improve their scores.

Dialpad Customer Intelligence with AI CSAT is now available to all Dialpad Ai Contact Center customers. According to Dialpad, early-access customers have reported up to a 15% improvement in CSAT scores within the first three weeks of using the tool.

Among the interesting industry aspects of this announcement is that it brings features to cloud contact centers that you cannot replicate with on-premises systems. Although the adoption of CCaaS has been strong, the value proposition has revolved around its being cheaper and faster to deploy. In fact, the CCaaS vendors such as Dialpad took years to reach feature parity. In my experience, this is typical of technology transitions.

The early days of virtualization held great debate as to whether virtual servers were better than physical ones, outside of being lower cost. Then VMware Inc. created VMotion and server administrators could move workloads across a data center. In this industry, people pushed back initially on voice over IP telephony until the vendors enabled customers to do things such as four-digit dialing across the globe. New technologies hit their inflection point when the new thing enables us to do things we could not do with the old thing.

In contact center, the cloud hasn’t yet reached that point, but we appear to be on the precipice. If a customer asks, “Why cloud?” for contact center, the answer should not have anything to do with cost. Instead, it should focus on new capabilities only available on cloud.

The AI CSAT is an excellent example of this. Dialpad is able to do this because they have the compute resources to run the processor-heavy AI algorithms as well as a massive amount of data from which to learn, both of which are not possible with on-premises solutions. Over the next few years, I expect to see the number of AI-based capabilities accelerate to solve some of the CX challenges that were previously unsolvable.

Zeus Kerravala is a principal analyst at ZK Research, a division of Kerravala Consulting. He wrote this article for SiliconANGLE.

Image: geralt/Pixabay