A guide to conversational AI for CX use cases

Pragmatic lowered down3 copy.jpg

What product managers should consider in helping to identify the tangible business value of a virtual agent solution.

Customer experience should be top-of-mind for any customer-facing business in the digital age. But as consumers increasingly take their interactions with brands online (and into the chat channel), how is it possible to keep up with their demands while providing a level of satisfaction on-par with what they have come to expect from human interaction.

In a recent report, Gartner predicts that as soon as 2020, 40% of users will primarily interact with new applications that support conversational UIs with AI. By 2022, it is further expected that 20% of all customer service will be handled by conversational agents. These numbers should send a shock up the spine of any product managers who are not already actively engaged in either implementing or, at the very least, considering rolling out a conversational AI strategy within the next six months.

The challenge for product managers lies in the fact that, while chatbots and virtual agents are some of the most widely used applications of artificial intelligence, it can be difficult to justify the investment unless there is a near-guaranteed return via a visible improvement in customer experience.

To help with this process, Gartner has outlined a framework entitled ‘Sense, Think, Do’ designed to help guide product managers in understanding how to apply conversational AI to build improved customer experiences. In this article, we will analyze this framework and how it applies to boost.ai’s conversational AI solution and our own principles of putting customer experience at the center of a virtual agent implementation strategy.

Sense - Deepen your understanding of customers

Conversational AI has the capacity to help businesses gather and parse valuable data about their customers. A virtual agent is not only available 24/7, but it also ‘remembers’ everything - provided the correct customer consent is given, of course.

Technologies such as natural language processing (NLP), natural language understanding (NLU), and boost.ai’s proprietary automatic semantic understanding (ASU) can filter and analyze chat conversations helping to give a more holistic view of every customer. This data can then be leveraged into better understanding and insights for potential future sales situations.

Our analytics dashboard offers a deep-dive into each and every interaction a virtual agent makes. From the frequency of messages and most common topics, down to the feedback customers give on individual conversations. Having access to this granular level of data means you are able to see how certain intents and responses track with customers and make tweaks to the virtual agent as necessary, improving customer experience virtually on-the-fly. 

Think - Explain, guide and recommend 

In providing the best possible experience for customers, it’s not enough that your virtual agent simply acts as a repository for the information already found on your website. This is, of course, useful for customers who need to search for something quickly, but the potential for AI to do so much more would otherwise be squandered on such a simple ‘answer-bot’.

With user authentication, a virtual agent has the power to perform actions on behalf of customers, as well as be proactive in making suggestions and recommendations.

The possibilities are far-reaching: imagine a customer asks to check how much of their monthly data allowance they have used on their phone plan. A virtual agent built on boost.ai’s conversational AI can not only respond with the desired information but, if the customer is close to exceeding their data cap, offer to upgrade the plan for the remainder of the billing cycle. All within the chat window and without ever needing to bring in a human operator. This example can be easily replicated across multiple scenarios and industries. By giving a virtual agent the authority to automate tasks, as well as take initiative when the AI identifies a potential sales opportunity, a business stands to both boost customer satisfaction and save valuable time and human resources all at once.

Do - Solve customer queries, every time

A virtual agent, when properly implemented, can be instrumental in building brand loyalty. However, there are instances where automation might not be an exact fit for solving every customer interaction. Sometimes - it’s true - a helping human hand is still necessary.

At boost.ai, we recognize that in order to provide a solution that delivers the most frictionless experience for the end-user, human support staff still play a pivotal role. Virtual agents built on our conversational AI are advanced enough to automate over 50% of all chat traffic when used as first-line support. Requests are either resolved immediately or seamlessly forwarded to a human operator with a chat log for reference. This frees up support staff from menial, repetitive tasks allowing them to focus on more complex queries and, importantly, means customers always get the answers they are looking for, whether they come from a human or a machine. 

What to do next?

There are a number of ‘next steps’ that product managers can consider in integrating a virtual agent as part of their customer experience strategy:

  • Take cues from the ‘Sense, Think, Do’ framework in evaluating which conversational AI applications can be used to enhance the overall customer experience for your business.

  • Focus on areas that have clear business and financial goals within a reasonable timeframe; i.e. automate 20-30% of all incoming chat traffic within 6-8 months of launch

  • Start proofs of concepts (POCs) and product trials within the next six months. Make sure to track ROI to help lay the groundwork for future AI investments.

  • Remember that CX should be a companywide project and requires that you evangelize outcomes to senior management and product leadership teams so you can engender support for your project.

Lastly, it is important to ignore the hype surrounding artificial intelligence and carry out proper due diligence when researching a conversational AI solution. Is the solution built with the enterprise in mind? How smart is the underlying technology, really? How well does it scale? These are some of the key questions to ask prospective vendors to make sure you end up with a virtual agent that actually enhances the customer experience, rather than tarnishes your brand’s image.

We recommend arming yourself with these nine critical questions to consider when assessing a conversational AI solution for your business.

Originally published in AI Authority.