How to Pick a Chatbot Personality That Doesn’t Suck.
Part I of the series: Chatbot Personalities That Don’t Suck. (feat. Kév the KoalaBot)
Ok. So. You’re building a chatbot. You know why you need it, what questions it should answer and the organisation it will represent very well. Now, all that’s left is to choose a personality that reflects your brand and then… well… profit! Right?
Wrong.
I’m sorry to say it’s time for a Total Recall-style mind-wipe because your strategy sucks. Your chatbot is never going to speak to your brand; it speaks to your users. (Unless it’s an internal bot and then your employees are people, not corporations.) No, I’m not talking about your target market either. I’m talking real living, breathing and artifice-hating humans.
This is not some distant end-user group who mindlessly buy products based on metrics and GPS location. These are opinionated and intelligent people like yourself, who hate annoying sales speak and irrelevant answers as much as you do. And when we talk to them through a chatbot, we talk to them directly.
That’s why normal marketing and business logic won’t apply here.
The real trick is — if you’d like a bot that your users want to talk to, then you need to design it from the perspective of your user, not your business.
Oh, and while you’re at it, don’t restrict yourself with the mindset of a business owner or a marketer.
I’m going to teach you how to put yourself in the mindset of a writer.
Design your chatbot as a supporting character
If you’ve ever had aspirations of novel-writing (me too 😉), then this is your moment. The best chatbots are built like supporting characters in books, not a talking FAQ page.
Just as a fictional character feels fake when they have no backstory or quirks, your business bot will feel like a mindless machine without a well-planned identity.
But don’t let them be the main character. The user is your protagonist because an ideal chatbot conversation is always centred around them. Put it this way: Would you enjoy a call with a Customer Service Agent who dominates the conversation with topics that don’t interest you? Users love to interact with bots that are tailored to them and which put their needs first.
So, what to do? It’s time to jump on board with our Conversation Design Team’s favourite word and be user-centric.
Step 1. Do the research (and do it right).
First things first — step away from the analytics. We don’t want to learn about our user’s annual earnings, after all. We’re here to find out how they interact, their true needs and what our chatbot can do to fulfil them. To do that, we’ll create ‘user personas’ or semi-fictional characters which represent a user’s typical behaviour/desires. These help us to pinpoint the ideal user experience from our user’s perspective.
I’ll demonstrate this with an imaginary Australian vegan beauty brand called OzLuxe. (Yes, this is where the koala comes in 🐨.)
There’s a couple of ways to go about this. The first, ideal way would be to get the info directly from your users. Perhaps with a poll on your site, or by putting a dummy bot live and gathering data on which questions are asked.
You want to find out:
- What are your users’ backgrounds?
- What will they use your chatbot for?
- What is their main barrier in achieving this goal?
- How should a customer service agent act, in their opinion?
- What don’t they like about chatbots?
If you can’t do that. Let’s stalk a little. Ethically, of course 😉.
Start where your users are and see how they interact there. If it’s a customer-facing bot, have a look at your customers’ tone of voice on your social media, blog or forum (if you don’t have one, ask your customer service team instead). If it’s a service, look at your feedback in reviews, emails or interactions with your clients. Any info helps - What posts do they engage with most, what’s their style of humour, what issues do they encounter where we could be more helpful?
Here, Ozluxe looked at their Facebook feed and noticed one of their most engaged with posts (that is, having the highest likes and comments) was about a frequent complaint — how hard it is to navigate their online store. Ozluxe searched for keywords like “browsing store” and found a high number of posts on the matter. Perhaps our bot needs to help customers locate products? It could also reflect their sense of humour and love of memes.
More of Ozluxe’s customers have expressed their frustration with hard-to-read ingredient lists, especially when trying to determine how animal-friendly a product is. Maybe our bot could assist them with some knowledge on this topic, explained in a simpler way? The chatbot should also be empathetic and relate to our users’ vegan lifestyles.
Investigate your users and answer the questions above. When your detective work is complete, use these qualities to make your User Persona Profile as I have in the image below. It’s normal to have more than one user type if your research reflects that, in which case, just make more profiles!
Here’s an example:
The profile you just made will form the basis of your bot’s personality in our next step!
Step 2. Create your bot character bio.
Eureka! We have our idea, let’s bring it to life. Write a character bio and plan out the aspects of your bot’s backstory (for real). Try to think about each element as if your bot were a real person (or animal, vegetable etc.). If they’re a certain age, how does that make them act, speak or feel? If they’re from a certain city, which part are they from? And how does that influence them today? All these somewhat meta points will help you build a character who has relatable experiences you can pull one-liners from and a tone of voice that appeals to your users.
If this all feels a bit silly, remember that just as we imagine appearances and origins for the invisible characters in novels, your users will attribute a personality to your bot whether you plan it or not. So, it’s best to take control with a little creativity.
Check out Kév’s bio here:
Don’t be afraid to let your bot be itself.
In a world where the tiniest mistake on Twitter can have you trending across the globe, it’s easy to see why so many Conversation Designers build chatbots that are generic and ‘safe’. However, I would argue that it’s almost as risky to be dull.
Many users have a predetermined misunderstanding of bots and feel a sense of resistance to them. By making a chatbot full of life and relatable to our audience, we create something which can build a true connection and therefore has the power to convert reluctant users into satisfied customers.
After all, a bot that’s as unique as your customers and company is the most genuine reflection of your brand anyway. Your business and its employees are not faceless machines, so why should your bot be?
