The Elusive Personality in Chatbots

If Grogu or ‘Baby Yoda’ as he is affectionately called by fans of the Disney Mandalorian Series was a chatbot, I will buy whatever that Grogu throws at me. Chatbots has evolved as a technology from being just an automated answering machine to being a tool that supports the dream of every character writer. Yet most chatbots that I have encountered are boring, lacks personality and feels more proportionate to a dumb robot.

In this article, I summarise my experience of working with both chatbot developers and marketers in Singapore and Malaysia, as well as my evaluation after working with different chatbot engines. As a game developer for most part of my entrepreneurial journey, I can understand the importance of creating relatable characters. However we are still not seeing more bots with personalities and this is why I think this is the case.

As technology has improved, organisations are beginning to understand that bots are not just a trendy display of digitisation, but a representation of your brand. While the bot is helping your customers make a reservation, doing a product recommendation, answering FAQs or engaging your customers in idle chit chat, your customers are considering the bot as an ambassador of your brand. Just as you will spend resources to develop a sales script and train your sales team, the same level of scrutiny should be put into the dialogues created by the chatbot.

Organisations implement chatbots with various business objectives. However, the main business objective given by my clients has been the automated answering of repeated questions such as operation hours and making reservations or appointments. Particularly in Singapore, manpower is expensive and the COVID-19 pandemic has greatly restricted the availability of foreign workers. Businesses will want their employees to spend their time on activities that have more value-add such as building customer relationships or performing the last mile service rather than replying emails or feedback on social media. 

Although this application of chatbot has picked on how AI can be used to help engage customers, technology adoption needs patience. Organisations need to be given time to explore the technology and understand how to integrate chatbots into their business processes.

Chatbot platforms with integrated Natural Language Processing have the ability to understand both the customers’ intent and context. This allows for a more complex personality to be created for the chatbots, one that is either comparable to Bus Uncle who speaks Singlish and helps you with your bus route, or equal to Rose who is a 37 years old female and has a life story to tell.

As fun as such characters seem, creating such a character can be overly technical for a marketer. A marketer’s role usually revolves around creating collaterals to communicate the brand’s message with their customers. They are not necessarily experts with language structures. Similarly, chatbot developers are basically software developers who are familiar with integrating chatbots with an organisations’ backend system. This creates a gap between the expectations of the chatbot developers and the marketers, with each party expecting the other to possess the expertise of developing the bot’s personality. 

The development of chatbot personalities is a creative process, not dissimilar to the process of character development commonly found in video games and movies. I will need to define the character’s personality and attach a believable background story. Specific to chatbot development, I will need to script out how the bot speaks by getting in-character to craft out the bot’s personality. This process inevitably increases the cost of developing the chatbot.

In Singapore, organisations understand chatbots through the bots deployed by government and banking websites. The deployed bots are the end-results of developments which do not provide these organisations with the visibility of the project management process, the budget and the challenges that happened during the development process. As a result, most organisations have little understanding of the factors affecting the budget, the engagement processes with the developer and the conceptualisation process behind the development of a chatbot. This is in contrast to the more widely understood project management processes for developing mobile apps. 

Hence the short answer to the question is that there are not enough examples of chatbots being developed with personalities that are solving an organisation’s pain points. This makes it difficult for organisations who are just starting out, to justify the spending of additional cost to create personalities in the chatbots.

A newly created chatbot is basically akin to an infant, with limited capability to comprehend what you say. Effort is required to train the bot to converse with you. The first level of training will enable the bot to provide proper responses to simple Q&A. It will then later be able to perform extended conversations with the customer. 

For example, a bot can be provided with multiple replies when the customer starts with a simple ‘Hi’. The tone of reply, with or without emojis, the length of the sentence, or the inclusion of Singlish, all helps to bring out the bot’s personality. The process of developing a bot’s personality needs to be done in consultation with the brand owner before the scripts are passed to the bot developer. 

At the next level of intelligence, the bot will be able to remember the context of your conversation. For example, the bot could be asked if ‘it’ is married and have kids, and the bot can respond that ‘she’ is married and has no plans for kids. When asked with a further ‘Why’, the bot would be able to relate that ‘Why’ to the customer’s question about her decision not to have kids. The bot can then answer the customer based on the character’s beliefs that were crafted for the bot. 

To achieve this level of intelligence, it is important for organisations to continue and invest on training the bots beyond its first launch. Organisations can justify the investment by measuring the success of the engagement through its interaction rate and the bot’s ability to bring more customers back to the organisation.

It may be years before we are able to create a bot personality that mimics C3PO, but intelligent bots with personalities are now no longer just sci-fi. Whether organisations begin to leverage bots to help customers place their orders, advise on personal finance or answer your health questions, these are just the tip of what a bot can do. Bots with personalities are able to help with lead generation and customer engagement rates. In the mid term, bots can help organizations gain a competitive advantage by building a shield around their pool of loyal customers. I am definitely excited to see more use cases of how bots with personalities can be used to increase my entertainment value of interacting with the brands I love.