Enriching Qualitative Analysis with Next-Generation AI

From hackathons to corporate board meetings, artificial intelligence (AI) is the hottest topic in applied technology today. Yet, with any hype cycle, it’s important to sift through the noise and determine the actual value it can offer. AI in and of itself is not a game changer for business. Rather, it’s the application of AI as a tool in a business process that can deliver amazing outcomes. AI enhances what you do, but in and of itself is not sufficient. It’s like seasoning in cooking. Alone, it offers little value, but when added to a recipe, it enhances the overall sensory experience.

As customer experience professionals, we obsess over helping our clients rapidly design and improve upon their products and customer experiences to drive top-line growth. And we all know that delivering the right products and experiences with agility requires incorporating the voice of the customer. So, as we assess the utility of AI in the world of CX,  it is important to consider how AI can enhance the community-driven voice of customer engagements.

Qualitative studies (customer verbatims) stand out as a clear opportunity for enhancement with next-gen AI. The value of qualitative feedback from customers is growing in importance as it often provides a deeper understanding of customer feedback and even uncovers unexpected insights not considered when creating quantitative studies.

However, unlike quantitative studies, which naturally lend themselves to standard segmentation, summarization, and analysis at scale, qualitative studies traditionally require human synthesis of large numbers of individual responses to glean intent, sentiment, classification, and pattern recognition. Humans are great at this, but the time required to do so at scale can make rapid decision-making difficult at best.

AI has been tackling this problem for years with mixed results. However, with the advent of new generative AI systems and large language models, qualitative studies can now achieve human-quality outcomes at scale.

A real-life example helps illustrate this evolution. We were approached by a prospective retail client who captured customer verbatims as part of their customer experience program. They received tens of thousands of these per month and were using early-generation AI to help automate the analysis. Before they could even see their first dashboard, they spent extensive resources training a model and designing a taxonomy for classifying verbatims. When they went live, the disillusionment began. The model was highlighting topics like “things” (not joking) because it was frequently expressed. Given this example, it’s not surprising that auto-classification returned over 125 key topics. AI was supposed to make their lives easier. Instead, they had to spend time retagging verbatims and spend more money retraining their models and rebuilding their taxonomies.

We ran the same sample data set on our new generative AI Text Analytics service. The results amazed the retailer. The auto-classification reduced key topics from 125 to 28. “Things” no longer appeared as a topic. Instead, relevant topics such as “response time,” “technical issues,” and “mobile app” were highlighted. Furthermore, the analysis accurately gleaned the discrete sentiment of unique thoughts within verbatims, e.g. “service was great [positive], prices are too high [negative].” And the value of this new generative AI was on full display where the retailer could filter by sentiment, topic, or drill down to individual responses. All of this was achieved in hours, not months.

This is just one example of how generative AI can help customers enhance their community-driven, qualitative voice of customer programs. Other areas include using AI to prompt customers to improve the quality and richness of their verbatim responses. What’s really exciting is that we’ve only uncovered just a small portion of the AI iceberg. Generative AI and large language models also hold great promise in automating translations of survey questions and verbatim responses so non-multilingual practitioners can quickly engage customers and analyze insights globally.

AI is more than just a hot topic. AI has the power to enrich customer experience programs to the next degree, delivering exceptional products and experiences.

Scott Barkley joined Alida in 2020 as the Senior Vice President of Product Management, responsible for product strategy, product vision and roadmap execution. He has over 20 years of global product leadership experience with such companies as Cisco, Jasper Technologies and Siebel Systems. Previously, he was VP, Head of Products for Cisco’s Internet of Things (IoT) business unit via the $1.4B acquisition of Jasper Technologies. At Jasper, Scott was a key member of the Executive Leadership Team responsible for product management. Over his 13 year tenure, he led the first GA release and over 200 subsequent software releases. His products generated over $200M in ARR with 140M subscribers, 7k+ enterprise customers in over 100 countries.