Gathering customer experience metrics relies heavily on AI technology, especially for data analysis and collection.
AI-backed data collection can improve personalization for marketing and problem solving in the contact center, both foundational aspects in the complex prism of CX for enterprises.
Customer experience experts at online conferences sponsored by CX giants AWS and SAP this week discussed how generative AI -- the latest and possibly most transformative version of AI technology -- could potentially be both helpful and harmful to the customer.
Preparing for better personalization
One overarching challenge in customer experience is keeping up with the increased levels of personalization that customers expect, said Pasquale DeMaio, AWS Amazon Connect vice president, during AWS Contact Center Day on April 26.
"Customers are now expecting 90% or more levels of personalization," DeMaio said. "Generative AI is clearly going to be a game changer here."
Meanwhile, hybrid shopping, or combining online with in-person purchases, is laced with personalization, from personalized product offerings to targeted promotions, according to Yulia Groza vice president of ecommerce technology at Levi Strauss and Co.
The apparel company uses AI and machine learning technology to create personalized experiences for consumers, she said at the SAP Industries & CX roundtable, also on April 26.
This personalization can occur in various ways, whether customers are shopping online and want personalized product sorting or recommendations or in a store looking for custom clothing personalization with a tailor, Groza said.
Deeper levels of personalization depend on relevant data collection, according to Ritu Bhargava, SAP president and chief product officer for industries and CX/CRM. Marketers must use not only engagement data but also data from business processes to improve customer experience and increase profitability.
This includes tracking customers' buying preferences -- whether preferring to order online or buy in person -- and giving customers who are ordering products the ability to monitor when the product will be delivered.
"This is where the 360 view of connecting the data really lights up every engagement point," Bhargava said at the SAP event. "What you want, ultimately, is brand loyalty."
Omnichannel data collection is essential for improving personalization, Groza agreed. She added that generative AI, with proper safeguards to curb incorrect information, has the potential to help accomplish this goal.
"There is an enormous opportunity for generative AI," Groza said.
Some specific examples of generative AI-supported personalization capabilities would include unique design creations based on consumer preferences based on users' interactions with a brand, Groza said.
AI technology unlocks call analysis
In the contact center, AI technology is unlocking a trove of previously untouched data sitting in audio from customer calls, according to Forrester Research analyst Max Ball.
Max BallAnalyst, Forrester Research
"What's really cool about that is people are using that to improve quality management," Ball said at the AWS event. "Now with AI, it's actually practical to take all of those recordings and listen to them."
With intelligent listening and transcription tools, companies can analyze text from calls to the contact center, and perform sentiment analysis and quality control. By extracting this data, companies can discover trends by agent, group or the entire contact center, Ball said.
Some examples of key items to check for would be agents remembering to state their names and ask customers for their names and recite the required statement that would deem the call legal and appropriate. AI has the capabilities to perform those tasks, Ball said.
"AI is now available to completely redefine how quality management works," Ball said. "It is cost effective."
Looking to the future, Ball urged caution with generative AI, and OpenAI's ChatGPT in particular. He said generative AI should not yet be used in the contact center because it can often give false information, often doesn't reveal its sources, is biased based on the information it's fed, and doesn't (yet) connect to backend systems, such as CRM databases.
"It lies. It takes data from multiple sources, and it puts it together and it guesses how to synthesize that," Ball said. "It's going to make assumptions that are not always right."
Mary Reines is a news writer covering customer experience and unified communications for TechTarget Editorial. Before TechTarget, Reines was arts editor at the Marblehead Reporter.