It’s hard to avoid the hype that ChatGPT and similar generative AI tools will change everything — including customer experience (CX). But scratch beyond the surface of the click-bait headlines about AI and CX, and you’ll find that writers have conflated CX with customer service or marketing. From chatbots to predictive product recommendations, most of the “CX” use cases that these articles discuss are related to delivering experiences. But these articles don’t address critical CX functions such as creating a CX vision or improving CX measurement.

So what does generative AI offer CX professionals like us? Not much — yet. Simply put, generative AI is not a mind-reader or an oracle. ChatGPT and its ilk only know the large language models (LLMs) that they were trained on — which don’t include your company’s customer feedback data. As a result, generative AI alone will not tell you how your specific customers feel about your specific products or services or replace the human efforts required to run a successful CX program.

That being said, there are some emerging use cases that CX pros should keep an eye on, including:

  • Increasing efficiency. Customer feedback vendors such as QuestionPro are releasing tools that use AI to build surveys based on simple prompts that could help teams create surveys much faster. As we know, however, survey design is both art and science. While automating survey design may save some time, it doesn’t remove the need for a human to ensure that the survey doesn’t contain “coherent nonsense” and to pre-test the survey with potential respondents. In fact, given that only about a quarter of voice-of-the-customer and CX measurement leaders are pre-testing their surveys today, organizations leveraging AI-generated surveys may find that they need to invest even more in human oversight and survey QA processes — or else risk prioritizing efficiency over quality. Generative AI may also help CX teams improve internal communications by creating presentations, summaries, and other reports based on specific queries of existing feedback. Indeed, customer feedback management (CFM) tools already offer some automated reporting and alerts. That’s good news, since sharing customer insights across an organization is critical to driving systemic CX improvement and CX teams spend too much time building bespoke dashboards and decks for different audiences on different reporting schedules.
  • Collecting unsolicited feedback. This recent Q&A that we had with Bing Chat shows some of the capabilities for summarizing ratings and reviews. Question: Are consumers living in Braintree, MA, likely to recommend Macy’s? Use only data from reputable consumer information sources.
    Answer: I found some reviews of Macy’s in Braintree, MA, on Indeed and Yelp. … On Yelp, there are mixed reviews about Macy’s in Braintree, MA. One reviewer mentioned that “Macy’s in Braintree, MA, South Shore Plaza mall is awful! They don’t answer any of the phones! No one has returned my voicemails!” I hope this helps!It’s easy to imagine similar queries summarizing social media sentiment or other public source data. While not enough to act on, these types of queries may help CX teams create hypotheses about the experience and quickly pull verbatims to use in reporting.
  • Understanding unstructured data. For CX leaders using a CFM or text analytics solution already, ask your solution provider how they plan to use LLM models to enhance their existing text analytics models to reduce tuning and training time and potentially derive intent more effectively. Separately, generative AI also has the potential to help CX pros search and summarize their existing customer feedback more effectively.
  • Closing the loop with customers. CFM providers (e.g., Customer Alliance and Data Appeal) are already experimenting with generative AI to respond to ratings and reviews, and it is easy to imagine leveraging generative AI to draft responses to a variety of feedback types. But we continue to urge caution before using these tools alone for customer-facing communication.

These emerging use cases for generative AI in CX are intriguing, but the top challenges facing CX teams include a lack of a clear CX vision or strategy and the lack of collaboration and buy-in from other departments. These are human problems that will continue to require human solutions.

Want to chat more about ChatGPT and its ilk? Forrester clients can request an inquiry or check out some of our other blogs, such as ChatGPT: Cybersecurity Ramifications Beyond Malware and How To Navigate The Generative AI Hype In Customer Service.