This is the second of a three-part series examining how companies are transforming their customer experience strategies through the use of technology. The first article in the series is about the use of all types of AI in the contact center.
In all my decades covering technology, I haven’t seen one catch on with the speed of generative AI. Already, 27.3% of companies are using the technology for their customer-related activities, with another 47.2% planning to do so this year, according to Metrigy’s Customer Experience Optimization:2034-24 global study of 641 companies.
Generative AI is a category of techniques and models that respond to natural language prompts to produce text, images, audio, software code, or other media from data on which they’ve been trained. Examples of generative AI, also known as Large Language Models (LLMs), include Open AI ChatGPT, Google Bard, Microsoft Copilot, and many others.
Potential Uses of Generative AI
The biggest value of generative AI is to automate tasks—keeping in mind that the algorithms are trained on language, not truth, so the ability to put guardrails around data becomes vital to the success of generative AI in business. How can businesses and consumers use the technology?
Current Uses of Generative AI
According to Metrigy’s survey, organizations are using generative AI on a variety of platforms, including:
Overall, generative AI is involved in 43.6% of customer interactions. Of course, some of this involvement is basic, such as a call summary report. But others are starting to explore content creation or even classification of issues on those calls or chats.
Generative AI also has the most profound impact on staffing costs compared to other types of AI. On average, companies using generative AI say they would have to hire 2.4x the number of agents if they didn’t have AI, which equates to $4.4 million per year.
Concerns of Generative AI
Despite the benefits and potential value derived from generative AI, several concerns exist. Most say generative AI can be trusted to be used on a limited basis. In fact, 30.9% of consumers say it cannot be trusted at all, compared to 17.3% of business leaders when used for customer interactions and 20.4% of business leaders when used for any area of the business.
IT, CX, and business leaders say the top way generative AI would be more trustworthy is to limit the data it can use to create content, according to the aforementioned study. But only half as many consumers agree, according to Metrigy’s Customer Experience Consumer Insights 2023-24 research study of 503 consumers. They would rather see human oversight or limitations on its capabilities. In fact, consumers are 4x more likely to say they will never trust generative AI versus IT, CX, or business leaders. The loud message: More education is needed for the consumer market in order for them to feel comfortable with the technology.
The top concern of generative AI among IT, CX, and business leaders is the loss of the human touch in interactions with customers. Additionally, they are concerned about the content quality, including accuracy of responses, data privacy, the ability to select or limit the data source, and Internet bias. Malicious use of the technology, as well as taking jobs from people also make the list of concerns, though not as high as the aforementioned concerns.
Interestingly, cost is not a big issue. Only 17.1% cite the “unknown cost” of generative AI as a concern, indicating that most leaders have accepted that any cost of AI is outweighed by the benefits generated.
For now, nearly half of CX leaders are addressing the concerns for content quality by developing new processes for validating the accuracy of information. Most (52.0%) have a content team reviewing content generative AI creates, while 47.3% have a virtual assistant review it. (Yes, that’s AI fact-checking AI!). They also have supervisors (46.9%) or agents (41.8%) review content. Moving forward, as vendors perfect the ability to put guardrails around the content from which generative AI draws, content teams will become even more vital for making sure the content within company knowledge bases is accurate, timely, and creative.
During interviews with IT and CX leaders, the appetite for generative AI is strong, but the desire to learn more about the technology and roll it out conservatively dominates initial strategies.
In the final post from this three-part series next week, I will cover the good, bad, and ugly about chatbots.
