More often than not, if you read an article about why chatbots still “suck” you’ll be hit with various descriptions of the same issue – that chatbots have little to no “personality” and that conversations often fall flat, or are unsuccessful, because of their lack of understanding of common user phrases.
However, the root cause might not actually lie in chatbots lacking charisma and wit but instead in a lack of developed AI.
What are people actually wanting from chatbots?
When “personality” is discussed during chatbot design sessions, the term usually refers to a bot that has a specific tone of voice, is able to understand and communicate through small talk, has some kind of humour, and uses emojis in creative and playful ways. Yet none of these desired characteristics actually end up mattering if the chatbot isn’t able to provide end users with real utility and solve real problems.
48% of consumers don’t care if a chatbot has a personality, as long as it solves their issues instantaneously (and 19% don’t want any personality at all)!. This data indicates that people prefer fast and efficient conversations, over trying to re-create the human-to-human chat experience. We live in a world that has a growing culture of impatience and instant gratification – this is why companies are capitalising on same-day delivery services and providing the ability to stream complete seasons of TV shows within seconds.
So, if you have a chatbot that’s brimming with personality and wit, but can’t provide instant access to relevant information, or provide a useful service via tight business system integrations, then consumers aren’t going to tolerate another second with your brand’s bubbly, but a basic bot.
Utility and usefulness are the most important factors
Today’s intelligent assistants are filled to the brim with useful skills. They can set alarms, check the weather, play music. If you’re building one for your brand, maybe it can check an order status, or search for a product. But, when it comes to rich conversations that are not purely transactional in nature, things can break down pretty quickly. Often times, this is mistaken as the chatbot lacking personality, when in reality, the user is bumping up against the current limitations of Natural Language Processing AI.
So it may surprise you that there are still plenty of bot building tools out there that lack any sort of AI integration. If the latest in cutting-edge AI still has limitations when it comes to natural conversations, then it’s easy to imagine how a chatbot built without any AI at all can very quickly turn into a brand-damaging disaster. If a company over-focuses on dressing up their bot with personality and wit, this can distract from the hard work needed to implement an effective layer of AI that truly helps customers with their problems.
This is why utility and usefulness always need to come first – most people already trust chatbots with basic requests and are more than willing to interact with a chatbot rather than having to wait for a human agent. However, only 15% of people would interact with a chatbot for the fun of it. This is a clear indication that consumers want performance over personality – they don’t want to interact with “funny” or “interesting” personalities, they want their problems solved instantly – especially if the bot takes on a significant role in an organisation.
This doesn’t mean that people are happy interacting with a lifeless, monotone “thing”; it just means that people are more willing to sacrifice the ease of human interaction for the promise of getting a useful answer, instantly.
Create a persona instead of a personality
So, does this mean that a business can get by with having a chatbot that is lifeless but provides a tonne of value? In essence, yes, but you can still add interesting expressions of your brand values in iterative phases after your initial launch.
An easier alternative to cooking up a complex personality is to develop a simple persona for your bot that aligns with your brand and target audience – it could be as simple as a name and age to keep in mind when writing the language your bot will use. In the beginning stages, designing an intricate personality isn’t necessary – the majority of its persona is going to stem directly from what your company represents: the tone of voice, the language, the values that you’re most likely already utilising on all your other digital channels.
There may come a time where it makes sense to plan out complex personalities for your chatbots, but at the very start, it’s far more important to create an experience that is useful and to aggressively monitor AI accuracy and improve AI performance instead of worrying about what joke to crack.
Key takeaway: When building a chatbot, personality does not equal intelligence. Intelligence comes from your AI solution and it’s wise to invest primarily in an accurate AI model that can handle a broad range of queries. Only once your AI is in place is it wise to spend time developing a simple persona for your chatbot, rather than a complex personality. Consumers are more than willing to tolerate a chatbot that is able to provide them with instant answers as opposed to one that is overflowing with personality and wit, but can’t solve any of their problems.