The Harris Poll/Interactions study shows AI may be making consumers fall In love with voice-based customer care

FRANKLIN, Mass.—Today, companies want to take advantage of every channel available to reach their customers: text, phone, webchat, email, social messenger apps, and more are in the mix. With this array of options, companies have more opportunity than ever to interact with their customers and build loyalty and trust. Yet with so many choices, getting it right—supporting customer conversations in the manner people want, while also being efficient—is of critical importance. To enable the millions of customer conversations happening each day, companies across industries are turning to artificial intelligence (AI) as the linchpin to their customer experience strategy.

A new study conducted online among over 2,000 U.S. adults by The Harris Poll on behalf of Intelligent Virtual Assistant (IVA) company Interactions, Trust in AI Solutions, is the first to look at what consumers really want AI to bring to their conversations with companies—and what to leave behind. It reveals that two of three American adults who prefer texting/typing when communicating with a company (66 percent) would be likely to switch their preferred channel to voice if a company they interact with has a consistently effective and quick AI-powered voice solution. In fact, a slightly larger proportion of millennials (18-37) would be likely to change their preference to voice compared to older generations (68 percent vs. 66 percent of Gen X ages 38-53, and 64 percent of boomers ages 54-78).

Many Americans (51 percent) may prefer texting/typing over voice channels because they are frustrated with poor voice experiences. Nearly all Americans, 94 percent, have had frustrating experiences using voice systems to communicate with a company. “We believe that consumers are opting out of phone trees and giving up on their home and mobile devices in an attempt to bypass the limited set of self-service voice options that don’t address their questions or meet their needs, and turning to other channels” said Jim Freeze, CMO of Interactions LLC. “The survey results indicate that an effective AI-powered voice solution can solve this dilemma by providing a true conversational customer experience, and could greatly simplify the growing complexity of customer care channel options.”

Freeze continued: “Today’s leading companies are constantly questioning how consumers want to interact with their brand, and how comfortable they are with various types of AI capabilities. We’re often asked which situations are best suited for AI’s support, and how using it might impact brand-customer relationships. This curiosity is what sparked our interest in partnering with The Harris Poll to reveal how people want AI to be applied in the customer experience. The results of this survey fill a gap in the current AI conversation, providing insight into how AI, and voice AI in particular, can ease the burden of thousands of customer conversations each week—without frustrating your customers.”

The results of this poll suggest that regardless of channel, speed and effectiveness are top deciding factors in American adults’ relationship with AI, and are key drivers of customer satisfaction and positive business results.

Key Findings

Text vs. Voice : Battle of the Channels

In general, Americans are split when it comes to their preference for voice versus text-based solutions when communicating with a company, with text edging out voice by a slim margin: 51 percent to 49 percent. However, if a voice AI solution works effectively and quickly, the perceived value of the system skyrockets. In other words, getting answers/results fast is the ultimate driver of consumer preference.

  • 94 percent of consumers have been frustrated when using voice technology to communicate with companies.
  • Yet, in certain circumstances, 94 percent of consumers would prefer voice over texting/typing.
  • 66 percent of Americans who prefer text over voice when communicating with a company say they would be likely to choose AI-powered voice over text if it were consistently effective and quick at answering their questions.
  • 69 percent of American adults, including 75 percent of Millennials, say if a company they interact with has a consistently effective and quick voice AI solution they would be likely to recommend the company to a family member or friend.

Heated Conversations and AI Desirability

While many find value in interacting with human customer service agents, the majority of Americans are comfortable with a human not being involved in various circumstances. Additionally, there are specific situations where Americans prefer to speak with AI based on the nature of the interaction, the customer’s motives and the customer’s mood.

  • 65 percent of Americans think it is easier to negotiate/get away with more when speaking to a human agent rather than with a voice assistant, and about 3 in 5 (61 percent) think they have a better chance of succeeding in bypassing a company’s rules with a human.
  • 76 percent of Americans would be comfortable with a human not being involved when interacting with a company, with Millennials most likely to be comfortable with this lack of human interaction (88 percent of Millennials vs. 74 percent of Gen Xers and 65 percent of Boomers).
  • Half of Americans (50 percent) would prefer to interact with a virtual assistant over a human agent if they are dealing with an embarrassing customer service situation, and more than 2 in 5 (44 percent) have that preference if they are upset or in a bad mood.

For further detail on the study’s findings, including a detailed breakdown of generational differences in preferences and an in-depth analysis of preference for voice-based AI solutions in enterprise customer care, see here. To review the full Trust in AI Solutions report, please contact

Survey Method

This survey was conducted online within the United States by The Harris Poll on behalf of Interactions from August 14-16, 2018 among 2,022 U.S. adults ages 18 and older. This online survey is not based on a probability sample and therefore no estimate of theoretical sampling error can be calculated. For complete survey methodology, including weighting variables and subgroup sample sizes, please contact