Bupa’s Journey from a Contact Center to an AI-powered Intelligence Center

Sravani Gade

May 27, 20244 min read

Share this Article

Picture a regular day in your contact center. It is abuzz with a million customer interactions. Agents are busy on calls and other channels addressing inquiries. Other channels are also ticking with activity as customers try to resolve their issues in self-serve mode.  

Now, what happens to these customer interactions once they end? Most contact centers do one of two things. One, agents log these interactions, summarize the call and its highlights and move on to the next inquiry in queue. Two, the agents or supervisors have a process where they analyze each interaction to understand WHY the customer reached out – the customer’s intent, the sentiment and how they feel about interacting with the brand. While the former approach is the norm, the latter is slowly gaining traction. Why? Because all too often, contact centers are not harnessing the true power of all the data these interactions generate. It's like having a map to buried treasure but never digging beyond the surface.  

Not tapping into this treasure means your contact center has a lopsided view of the customer. Each time a customer reaches out on any channel, your agents lack the insight and context for the inquiry which results in fragmented customer experience (CX). 

By remaining content with just addressing inquiries, contact centers miss out on the golden opportunity to transform everyday interactions into intelligent insights that can fuel delightful customer experiences. Imagine if each customer interaction were a puzzle piece, offering clues to understanding preferences, predicting needs, and foreseeing trends. That's the promise harnessing customer data holds – it's about turning raw data into actionable insights, unlocking a wealth of possibilities to not just meet but exceed customer expectations with every interaction, it’s about listening to the Voice of the Customer (VOC). 

In our recent webinar with CX Network, Darren Randall, Product Owner at Bupa Health Insurance and our in-house expert, Raghavendra Rao, Senior Director, Product Management talk us through the intricacies of listening to the VOC and how doing this can lead to transforming one’s contact center into an intelligence center using AI. 

Read on as Darren shares real-life examples of the pitfalls and course corrections they did in this journey. Raghavendra Rao or Raghs as he is fondly known, helps you decode the functional steps behind achieving the transformation into an intelligence center.  

Table of Contents

Why is VOC important?  

Darren’s Insights: I’d like to share three learnings that proved to us why VOC is important. We learnt them the hard way when implementing chatbots at Bupa. I hope I can help you avoid making the same mistakes.   

  1. Never assume what your customers want. We implemented some basic FAQ chatbots assuming the most commonly asked questions were the only use cases our customers were searching for. Within a few days of launching the bots, we had to turn them off. We realized that our customers were looking for more assistance regarding their policies that we did not anticipate or plan for. This hurt our overall CX metrics, and we had to double our efforts to ensure our customers remain open to choosing chatbots in any channel going forward. 

  2. Do not put business needs ahead of your customers’. We assumed we knew about customers’ needs better than them. Contact centers usually tend to get in a race where they want to be the first one to launch a channel to beat competition or deflect customers from voice or SMS to digital channels to cut costs, but these moves can be quite counterproductive. We had launched chatbots with an aim to contain customers’ interactions within the channel which they did initially. Later, customers chose to directly call up our contact center and preferred to speak to the agents directly as the chatbot interaction frustrated them, defeating the very purpose.  

  3. Never rely on incomplete and out-of-date data. Before we implemented conversational AI, we manually sampled the data of about 1000 interactions based on which we created a two-year roadmap for our contact center. This data soon became obsolete as the nature of inquiries had changed and the amount of data we analyzed was just not enough to give us the right picture of want our customers wanted from us. For example, by the time we could analyze the data of how COVID was impacting our customers and launched our chatbots, the peak of COVID had abated, the lockdowns were gone and were now scrambling to cater to customers as their needs has changed. 

Raghs’ Direction:  Keeping customers at the center of all contact center operations seems so obvious when we say it. It is a lot tougher to implement, though. Why? Because contact centers have traditionally been wired to be reactive with a feedback survey being the only way for contact centers to gauge how customers felt about the whole experience.    

Here are three areas contact centers can focus to turn around their customer experience (CX) and go from being a reactive cost center to a proactive experience hub. 

  1. Begin with the “Why”. The first step to understanding VOC is to identify why customers are having to reach out to your business. What is the context of their query? Also, how do they feel about having to reach out to your contact center for support? Do they feel angry, frustrated, or simply curious? Understanding the context and emotion behind your customers’ inquiries – the contact drivers is the first step to align with VOC. 

  2. Have a set process of where the inquiry needs to go. Once you know the why behind a customer’s issue, know where to direct them to find a resolution. Should it be to a self-serve channel like an FAQ chatbot or a community page? Should it be routed to an agent with the necessary skills to address the inquiry? Having this process set up and ready ensures a seamless experience for the customer and the agent. 

  3. Analyze 100% of your interactions. Act in real-time. With contact center agents working in fully remote or hybrid environments, their supervisors should have the visibility into the nature of interactions, monitor the quality of responses, assess in real-time how the customer’s experience is unfolding. This visibility can fuel further improvements and personalization in customer interactions. Supervisors and agents should be able to analyze 100% of the conversations to truly understand what works for the customer; to truly listen to the VOC. 

What does ‘transform into an intelligence center’ actually mean?

Darren’s Insights: So, when we talk about transforming our contact center into an intelligence engine, here are the steps we’ve taken towards it:  

  1. Implement real-time predicted CSAT metrics: What we mean by this is that our agents can gauge customer sentiment with each interaction. Agents can see a figure that goes up or down based on the tone or content of the message. By color-coding these indicators—green for positive, orange for neutral, and red for negative, our agents adjust their approach, accordingly, ensuring a tailored experience for each customer. 

  2. Shift from NPS surveys to real-time sentiment analysis. We've moved away from relying only on NPS surveys. Instead, we now analyze the sentiment of every conversation in real-time. This shift has empowered us to make data-driven decisions at both the program and agent levels. By understanding the sentiment of each interaction as it happens, we can quickly identify areas for improvement and take proactive steps to enhance the customer experience. 

  3. Perform granular analysis for personalized feedback and development. With AI, we're now able to drill down to the granular details, providing insights at both macro and agent levels. This level of granularity allows us to offer personalized feedback and support individual agent development. The key is to ensure continuous improvement and refinement, driven by real-time data and insights. 

Raghs’ Direction: Here are three foolproof ways of staying on track in the transformative journey of becoming an intelligence center: 

  1. Leverage interaction data for proactive enhancements. The first step to this transformation is recognizing the wealth of data available within our contact centers. This data holds valuable insights into customer behavior, preferences, and pain points. By leveraging this data effectively, we can proactively enhance the customer experience and drive meaningful outcomes. 

  2. Embrace AI-First approaches. AI plays a pivotal role in this transformation. Having an AI-first approach means to identify the best areas where AI can add value. In this case, AI can at scale, accelerate the process of analyzing, identifying patterns and now with Generative AI, recommending the next best actions to take based on huge volumes of customer data. 

  3. Adapt and improve continuously. An intelligence center is in essence a hub for continuous data-based improvement and adaptation. By harnessing the power of AI and data analytics, we can constantly refine, optimize and personalize our customer experiences. This iterative process allows us to stay ahead of evolving customer needs, predict and pre-empt radical shifts in customer expectations to a great extent and deliver exceptional service at every touchpoint. 

What value does conversational AI add?  

Darren's Insights: Conversational AI has admittedly revolutionized the way our contact center operates. Here’s how conversational AI added value to our operations:  

  1. The switch to intent-based routing: Previously, we relied on outdated menu-based systems, but with NLU routing, we've embraced personalized experiences aligned with varied customer intents. This shift has been pivotal in enhancing engagement and satisfaction. 

  2. Personalization and automation. Conversational AI empowers us to personalize customer journeys based on intent, automating what we call pre-chat capture to streamline interactions. By gathering essential information from the customer beforehand, we reduce the handling time and ensure a smoother experience for both customers and agents. 

  3. Ensuring quality at scale. Leveraging conversational AI, we've strengthened our quality assurance processes, ensuring compliance in real time. As an insurance company, we have a mandate to state a standard set of compliance statements. AI-powered automation flags instances of missed compliance, allowing supervisors to intervene promptly. This approach enhances operational efficiency while upholding regulatory standards. 

Raghs' Direction: From my perspective, a unified customer experience across all channels is paramount. Conversational AI facilitates consistency in tone and assistance, regardless of the engagement platform. This cohesive approach enhances customer satisfaction and loyalty. Here are a few other ways Conversational AI can add value to your contact center:  

  1. Deflection and Complexity Handling: If a contact center were to categorize their inbound inquiries as low, medium and high complexity, conversational AI empowers contact centers to handle low to high complexity issues efficiently. By deflecting routine, low complexity inquiries, we optimize resource allocation, allowing human agents to focus on the high complexity tasks. This balance between automation and human touch is key to delivering exceptional experiences. 

  2. Balancing Cost and Quality: It's crucial to strike a balance between cost reduction and quality service. User-friendly conversational AI systems make it easy for your employees to design, launch and modify customer journeys without needing extensive IT intervention. Conversational AI led journeys should be treated similar to agents. The journeys should be reviewed periodically to ensure they are delivering value and terminate the journeys that are redundant. This ensures ongoing improvements while prioritizing the quality of customer interactions.

Closing Remarks  

Darren’s advice is to start with small, foundational steps in gathering analyzing data and then scale rapidly. For organizations new to digital messaging, let your customer reach out to your first, to allow for valuable data collection. When transitioning vendors or adopting conversational AI, avoid the lift-and-shift approach and instead start fresh to maximize benefits. Lastly, embrace experimentation and creativity to unlock the full potential of these technologies.   

Raghs’ suggestion is about the importance of breaking data silos and unifying your data. This foundational step enables the implementation of various strategies discussed today, including VOC, conversational AI, and agent improvement initiatives. Without addressing data fragmentation that can occur due to channel-based silos, incompatible tech or rigid legacy systems, organizations risk stagnation in their traditional approaches and miss out on truly becoming the intelligence center that can deliver personalized experiences. 

Share this Article

Related Topics

Domain-Adaptive Instruction Generation (DAIG)Close the Experience Gap - How Sprinklr can help US Federal agencies enhance Citizen & Employee ExperienceThe Lufthansa Story: Unlocking Contact Center Efficiency with Generative AI