Elevating customer experience: The rise of generative AI and conversational data analytics

This article is part of a VB special issue. Read the full series here: Building the foundation for customer data quality.

The rapid advancement of artificial intelligence (AI) and machine learning (ML) technologies is pushing the boundaries of what can be achieved in marketing, customer experience and personalization. One important development is the ongoing evolution of generative AI (gen AI), which is bringing open-source platforms to the forefront of sales. As the digital-first business landscape grows increasingly complex and fast-paced, these technologies are becoming indispensable tools.

Across industries, engagement models are undergoing significant transformations, as customers expect to access products and services anytime, anywhere and in every possible way. While customers still value a balanced combination of traditional, remote and self-service channels, there is a noticeable surge in their preference for online ordering and re-ordering in the post-pandemic era.

To address these escalating demands, achieve e-commerce excellence across the entire customer journey, and improve hyper-personalization, Big Tech and SMB players alike are making major investments in generative AI innovations.

Producing fresh, original content

In contrast to traditional AI approaches that depend on predetermined rules and datasets, generative AI can produce fresh and original content. This cutting-edge technology uses intricate neural networks to discern patterns and generate distinct outputs — a whole new way to generate recommendations and offers.

Using conversational data analytics, businesses gain valuable insights into customer preferences, sentiments and pain points. They can use these insights to further refine products, tailor marketing campaigns and provide better customer support.

In today’s highly competitive and fast-paced digital world, personalization is the preferred strategy for brands seeking to stand out amidst the marketing noise. Effective consumer personalization is the secret ingredient, enabling tailored content and experiences that cater to individual tastes and desires. This amplifies customer experience, enhancing loyalty and retention and increasing return on investment (ROI).

By harnessing generative AI businesses can rapidly create highly targeted content that resonates with their audiences. A prime example is Spotify. The platform uses gen AI to analyze user listening patterns and preferences, then generates curated playlists and provides personalized music recommendations, ensuring that users remain engaged.

Availability of dynamic offerings

According to Beerud Sheth, CEO of AI-based conversational engagement platform Gupshup, companies ranging from Amazon to Netflix have long utilized AI in various forms to provide recommendations based on our past purchasing or viewing history. But the advent of gen AI has created a surge in the availability of dynamic offerings.

“Generative AI can be used to create and target marketing campaigns based on a variety of factors, such as customer demographics, interests and purchase interactions,” Sheth said. “This can help businesses to reach the right customers with the right message and increase the chances of a conversion.”

Likewise, Sreekanth Menon, VP and global leader of AI/ML services at Genpact, said that with generative AI, the landscape of hyper-personalized customer experience (CX) is poised to attain new levels of agility.

“The emergence of cloud-led advanced analytics technologies has allowed enterprises to capture insights from omnichannel customer contact points more efficiently,” Menon told VentureBeat. “Capturing, curating and analyzing sentiment with AI/ML across customer conversations amplifies organizational efforts to reach, react and recalibrate their businesses as per the demands of their customer quickly.”

Conversational data analytics for targeted campaigns

Integrating generative AI data with conversational data analysis has emerged as another powerful method for businesses to identify intricate patterns and trends.

For example, when a user engages with a brand’s chatbot powered by a large language model (LLM), conversational data is stored in the cloud. Later, this data can be analyzed using sentiment analysis to gain insights and understand consumer preferences and pain points.

Gupshup’s Sheth added that analyzing conversational AI data enables the identification of common customer questions and concerns. This valuable information can be used to create more comprehensive and informative FAQs or develop chatbots capable of automatically addressing these inquiries.

He highlighted that the data plays a crucial role in tracking customer satisfaction levels and acquiring insights into customer preferences. This process, in turn, enables companies to enhance personalization and create new products that cater to specific customer needs.

Hyper-personalization

Gupshup recently worked with the Dubai Electricity and Water Authority (DEWA), whose gen AI chatbot provides 24/7 customer support and assists customers in finding answers to common questions and requests such as billing inquiries, outage information and service requests, Sheth explained.

Likewise, California-based end-to-end video commerce platform Firework recently introduced its generative AI sales assistant to accompany its core video commerce offering. The patent-pending technology allows customers to use the in-video chat feature on an ongoing, on-demand basis.

“Long after a live stream has concluded, shoppers can ask questions about the products or services featured therein, and our proprietary AI engine will provide accurate, real-time responses based on user input, the content of the video and other associated metadata,” Jerry Luk, Firework cofounder and president, told VentureBeat. “Our AI engine makes use of an LLM that can understand and respond in a wide range of languages and can be customized to reflect each brand’s unique voice.”

Luk said that with the integration of gen AI and conversational data analysis, his company saw a significant boost in interactions with customers online.

Conversational data analysis combined with generative AI “allows us to analyze conversational data in real time, understand customer needs and preferences, and suggest what the human associate could say next,” Luk explained. “This fusion of human and AI capabilities can facilitate highly personalized and engaging customer interactions and allow associates to handle a wider range of queries, which feel more relatable and less transactional.”

Key considerations for adopting generative AI in CX pipelines

Luk emphasized that AI’s responses must reflect a particular brand’s voice and values. “The technology should be able to adapt to your brand’s unique tone and communication style. This consistency helps maintain your brand image and identity in AI-driven interactions.”

Peter van der Putten, director of AI Lab at low-code AI platform Pega, suggested that granting LLM access to internal documents and data can empower the tool to comprehend brand voice based on historical data. This then enables AI to take appropriate actions.

“By providing consumer-facing AI models with documents and information that are not typically included in generic models or accessible to them, companies can empower their chatbots to offer references to specific services or products,” said Putten.

Jonathan Rosenberg, CTO and head of AI at cloud contact center solutions firm Five9, pointed out that chatbots often tend to hallucinate (make up false information).

“Therefore it is important to include a human in the loop so that it compensates for [that] tendency,” he said. “It also creates a personalized experience for the customer. When they call back, the next agent will be able to know what happened previously.”

Mitigating generative AI challenges

Likewise, the emergence of generative AI has added complexity to the discussion surrounding AI risks from hallucination, said Menon. He emphasized that even with the utmost caution, chatbots are susceptible to adversarial attacks, including prompt injections.

Consequently, it becomes crucial to establish responsible AI strategies and architectures to mitigate these challenges.

“The significance of responsible AI in this context cannot be underestimated,” said Menon. “For enterprises leveraging generative AI, it is a strategic imperative.”

Sheth from Gupshup agreed, highlighting that AI models can sometimes result in discriminatory outcomes. Therefore, businesses must exercise caution and be aware of potential bias in their models. Failure to mitigate bias can make it difficult to interpret the operational processes of these models.

“Given that generative AI models are still in their early stages of development, it can lead to concerns about trust,” said Sheth. “Businesses need to build trust with their customers and stakeholders by being transparent about how they use these technologies and by ensuring that they are used responsibly and ethically.”