Why Your Chatbot Should Be a Friend, Not a Salesperson.
Part II of the series: Chatbot Personalities That Don’t Suck. (feat. Kév the KoalaBot)
So, you’ve picked the perfect chatbot personality in our first article, and now it’s time to bring that character to life with your content. But tread carefully, the right way may go against your instincts.
If you’re a marketer or business owner, you may find there’s a temptation to turn every chatbot response into a call to action, or worse, an advertisement. You probably don’t realise you’re doing it, so ingrained within you is the ability to sell (and do it well). But that’s not what a chatbot is for. If you envision your chatbot as a customer service agent first, you will always get better results. They are here to help, assist and listen. They may increase sales, yes, but they are not (usually) there to preach or sell directly to the customer!
To put it in perspective, a chatbot that offers a product at the end of every sentence is the equivalent of a salesperson in a shop who can’t answer a question without finding an accessory that “suits the shirt you chose perfectly!” AKA: unwanted, unnecessary and frankly, annoying and superficial over-help.
That’s why I believe a chatbot is most helpful when it interacts like a friend, not an advertisement.
To prove my point, let’s take a look at three versions of Kév the Koala (probably from alternate dimensions) and see which is most successful.
1. Sales-y Kév
2. Gimmicky Kév
3. Helpful Kév
← Here, Kév keeps it fun with personality when appropriate, but makes sure to answer our customer’s question first.
He doesn’t overwhelm them with unnecessary information, but rather, opens up the dialogue for their response and keeps them engaged with a real turn by turn conversation.
This way, he’s only telling the user what they really want to know.
Remember your customer isn’t always having a good day.
Much like any decent Customer Service Agent, one of the essential traits for a Conversation Designer to have is empathy. Until AI is at the level where it can (accurately) sense for us, it’s our job to detect which questions are coming from customers that are having a bad experience and design the answers accordingly. This is integral because if an individual is already unhappy with your organisation, a simple misplaced word or unintentionally flippant tone can push them over the edge.
For example:
4. Insensitive Kév
5. Sensitive Kév
Bonus Tip:
It’s not just the answers. Your whole bot’s personality could be off if you forget to consider the different positions of your users. For example, when designing a bot for an airline, there could be a temptation to continually discuss holidays as if that’s the only reason people fly. Eg. “Pack your beach bag, get ready for cocktails and let’s book that flight!!” But there are a lot of less glamorous or even difficult reasons someone might take a flight, such as a tiring business trip or a personal emergency. Try to be sensitive by considering all the reasons your users might interact with the bot and let that influence every part of the work you do. It might help to write user personas, like in our first article.
In conclusion, It’s important to stick to your brand voice and bot persona, but remember that too much personality can annoy users who simply want their answer.
A simple trick to keep things natural and conversational is to try and be witty in the same way that you would on the phone to a customer or in a Whatsapp group chat.
Try to imagine you’re having a conversation with a real person and write each answer as a direct response to their question. If one solution doesn’t match two similiar queries, then create two replies. Something that will help you a lot in this process is to test, test, test all content in the chat widget to get a real idea of flow. Happy chatting!
Show Kév some appreciation with your 👏. 🐨 He’s a bit of a diva and requires constant validation.
Learn more about chatbot personalities with Kév the KoalaBot here:
