Shift from reactive to proactive engagement with customers

In last week’s blog, we focused on finding the right balance between leveraging AI to drive insights and keeping it simple and transparent so that people trust the recommendations and incentives tied to CX analytics. In addition to our usual breakdown of ChatGPT’s answer to our series of questions, we talked about overcoming AI’s complexity gap and the right use cases for incorporating AI into your CX

This week, we’ll take a look at ChatGPT’s answer to the next question in our series: “What are the best practices for designing AI-powered CX interactions that are natural, intuitive, and effective?” We’ll outline our own thoughts on striking this balance, after quickly sharing some thoughts on how well ChatGPT did answering the question.

Chat GPT’s Answer (repeated here, verbatim from ChatGPT): 

Designing AI-powered CX interactions that are natural, intuitive, and effective requires careful planning and attention to detail. Here are some best practices to consider when designing AI-powered CX interactions:

1. Start with user research: Before designing any AI-powered CX interactions, start by conducting user research to understand customer needs and preferences. This can help ensure that the interactions are designed with the user in mind, making them more natural and intuitive.

2. Use conversational language: When designing AI-powered chatbots or voice assistants, use conversational language that mimics human speech. This can help make the interactions feel more natural and intuitive, and can improve the overall customer experience.

3. Incorporate context: AI-powered CX interactions should take into account the context of the conversation, including the customer's previous interactions with the business. This can help make the interactions more personalized and effective, and can improve the likelihood of a positive outcome.

4. Design for fallbacks: AI-powered CX interactions should be designed with fallbacks in mind, in case the AI is unable to understand the customer's request or provide an appropriate response. This can help ensure that the customer is not left frustrated or confused by the interaction.

5. Test and iterate: Like any CX interaction, AI-powered interactions should be tested and iterated upon to ensure that they are effective and meet customer needs. This can include A/B testing different versions of the interaction, or conducting user testing to get feedback from customers.

6. Provide human support: Even with the best AI-powered CX interactions, there may be times when a customer needs human support. Make sure to provide clear pathways for customers to get in touch with a human representative, and ensure that the handoff between the AI and human support is seamless.

By following these best practices, businesses can design AI-powered CX interactions that are natural, intuitive, and effective, improving the overall customer experience and building trust in the AI-powered interactions.

How Did ChatGPT Do?

Once again, Chat GPT does a pretty good job at answering the question with some relevant and pragmatic recommendations. While ChatGPT doesn’t specifically provide a set of use cases to apply AI for natural, intuitive, and effective interactions with customers, its answer is particularly useful for automating interactions with customer service agents. It emphasizes the importance of using an iterative approach to integrate AI, as well as leveraging a customer’s past experiences as a means of personalizing interactions. ChatGPT’s recommended approach, while helpful to lower cost-to-serve, doesn’t fully address the breadth of use cases where AI can be applied. In particular, ChatGPT is silent on ways to shift from a more reactive approach to customer service to a more proactive one that seeks to increase the frequency and quality of interactions along the customer journey. We think there are two opportunities ChatGPT doesn’t address, both of which help create natural, effective AI-powered CX interactions.

Opportunity 1: Shift from reactive to proactive engagement

Why wait for the customer to call with a problem? AI makes it more cost effective and feasible to engage customers throughout the journey without the need for human beings to take the time to spot the opportunity for more proactive engagement, prepare the content, engage in the conversation, analyze the results, and recommend the next best action to take, if required, afterwards. There are a growing set of platforms that support messaging (e.g., texting, in-app), and these can be applied to take a more proactive approach to engage customers at key moments that matter across industries, such as after a customer gets a delivery from a retailer, during an outage for a utility, after a guest checks into a hotel, or after a BtoB customer gets a custom price quote from a supplier, to name just a few.

By shifting to proactive engagement, you not only make it more likely that you will surface the customer's needs earlier, but you also gather valuable information that helps you to tailor the experience. This can be used to route an action to a human being, or to immediately provide a recommendation to the customer. Netflix pioneered the use of collaborative filtering, which leverages what others like you have enjoyed to make personalized recommendations. Similarly, in ecommerce, companies use this approach to look for natural “attachment rates” where products go well together. The same principles can be applied in the customer journey to get beyond addressing immediate problems, creating new peaks in the customer journey, rather than just reducing valleys.

Opportunity 2: Leverage AI to drive ongoing insights

ChatGPT’s emphasis on conducting user research before implementing AI-driven initiatives is commendable, but it doesn’t call out the opportunities that arise from mining the chatbot interactions themselves. As you prioritize and implement use cases for more proactive customer engagement, you will generate a significant amount of unstructured data beyond the voice-to-text records from customer calls or chatbots to handle problems. This presents an opportunity to shift towards a continuous listening approach, which is key in being able to proactively engage customers.

Analysis of your unstructured data can be used for more than just trend analysis and scoring the calls, as useful as that is. It can also be used to deepen your insights into where to focus in your experience design efforts. Just as your 3 star reviews tell you the most about what it takes to get to a 5 star review, this broader set of unstructured data can be leveraged to build ongoing insights from how to deliver a better overall experience. By iteratively testing improvements to how you apply AI at specific moments along the journey, you can strengthen your muscle at developing and applying these insights to leverage AI for a more proactive approach to customer experience. You can also share these insights with your frontline employees, spurring them to be more “clue conscious” for their own interactions with customers.

For more on these opportunities, check out my blog series with Lou Carbone on Reimagining Insights. You may also enjoy my video interview with Farzad Aref from Farlinium, with whom I’m partnering to implement these opportunities in Qualtrics. It’s included here, along with other relevant videos with other business partners for JourneySpark Consulting.

Up Next Week

In next week’s blog, we’ll be writing about the next best frontier for AI to create value in CX beyond reducing call volumes and automating work for call center agents. I’ll continue posting weekly for this AI series every Wednesday covering each of the ten questions. I’ll also continue to post one of my book reviews for my favorite ten business books every Friday. If you’d like to jump ahead to see all ten reviews now, click here to check them out.

We’re looking forward to continuing the conversation!