Even though we often use chatbots as consumers, not many understand what goes behind the scene. How does ‘it’ work? What can ‘it’ do? Who designs ‘it’? Why does a business choose to use ‘it’? This article examines these frequently asked questions that shroud conversational user interfaces (CUI) like chatbots and virtual assistants, and how to best approach starting one for your business.
2016 was coined as the year of Conversational Commerce;
2019 the year of Virtual Assistants.
As human beings, we communicate verbally or textually, it’s one or the other. Consumers can get in touch with your business in a myriad of ways, but at the core of it, every interaction is a two-way conversation – between the customer and a representative (perhaps even multiple representatives) of your business. Hence, it is imperative to have “conversation” as the key consideration when building an engagement strategy today.
Conversational interfaces will become the preferred channel consumers use to approach businesses as they become more accustomed to virtual assistants. To put it simply, a CUI is a platform or a channel that mimics a real conversation with an actual person. This could be as simple as a FAQ chatbot greeting you on your bank’s landing page, or as complex as a highly-skilled virtual assistant/agent that is equipped to help you pay your monthly bills and send text messages to your loved ones. So why are they gaining such popularity in recent years?
Traditional search methods such as using keyword search which requires an exact match and navigating a website or an application with buttons are relatively inefficient. As such, even internal users like employees are partial towards CUI which allows them to save time from scouring through complicated internal systems. Businesses using conversational interface helps consumers and employees by eliminating the frustration of manual searches and providing efficient resolution and direct answers.
The end goal of a CUI is to carry out conversations with users from start to end. How do we design a platform to satisfy this goal?
Intent - Rather literally, intents are “the intention of the user’s words”. It examines the user’s words to determine if the user posing a question, setting a context, or answering the machine’s question.
There are endless possibilities of what a user wants. Hence, AI technology is useful in decoding the true intent of the user and provide solutions efficiently and generate replies that are beneficial to the user. The AI technologies that are present in CUIs are natural language processing (NLP) and natural language understanding (NLU). This dependence stems from the inherent complexity of human speech and its difference from what logical and perhaps more regimented computer system comprehends. Besides understanding the user’s intention, the CUI must also provide an answer or resolution; between the user and the machine requires a logical interpretation, which can be developed using a powerful orchestration layer.
Typically, there are 2 approaches to designing CUI:
Businesses can approach this with a micro-perspective by breaking down a conversation and focusing on intents when building the platform. To design a CUI from the micro-view is extremely pragmatic and useful for businesses whose end-users have clear, distinct needs. For example, the Customer Service team of a retail bank identifies the need to lessen the time spent on answering repetitive questions on account opening procedures, so they contract a vendor to set-up a chatbot to handle account opening related enquiries. The downside to this is the chatbot’s limited capability because it can only respond to the scope of questions it has been trained in. Three months later, the bank notices customers have been asking the bot questions on bank loans and overdrafts and decides to expand the bot’s knowledge base by adding new intents. Over time, the business finds itself on a relentless journey of creating new intents to cover all grounds, with no clear direction on the CUI’s purpose or objectives.
A better way to design a conversational platform is to approach it with a macro-perspective and focus on the end goal of making it fully conversational. Instead of focusing on narrow intents, businesses can conceive an overarching problem statement through a panoramic lens and have a long-term plan to map out the evolution of the CUI’s objectives. The high-level design will be able to serve more complex task flows and be expendable to unlimited scenarios. Returning to our retail bank example, a more scalable approach would be to examine the arsenal of products offered by the bank and the similarities between the user journeys and questions that may be asked in each silo, and map them across different phases for implementation:
Managers who have identified strong business cases will seek approval from management for a budget to develop AI chatbots or virtual assistants. AI chatbot is a scalable solution which can be deployed on many platforms to ease the passage of information for internal and external audiences.
Having a dedicated AI chatbot for your business is a great way to build a persona. With a strong understanding of the business logic and AI logic flow, there can be endless possibilities of what your given chatbot can do. Although it can be tempting to want to do it all, businesses should work with experienced project managers to map out the execution and prioritise the most lucrative or feasible use cases and its intents first, as we have outlined in Approach 2 above.
CUIs can be used to serve a multitude of business use cases but fundamentally understanding the best way to design it will help any business better plan out its customer engagement strategy. Tunnel visioning yourself into a narrow problem statement often results in an incohesive CUI strategy across the organisation. The next time you think of implementing an AI chatbot or virtual assistant for your business, take a step back to look at what are the ways a good CUI can help tie all your touchpoints together for greatest ROI on your technology investments.