Look, who’s talking? — Designing voice- and text-based chatbots with personality.

Marco Spies about how to design chatbot conversations with the brand personality in mind.

Messaging-based chatbots often need to share the user’s attention with other chatbots. It is quite possible that a user has several virtual chatbot contacts in his messenger in addition to his human contacts. The user will “ping” them as much as his human contacts:

“Hey Anna, I have to go to London tomorrow for a few days, can you suggest a flight?” “Sure! Do you like to fly alone or for how many adults and children should I search?”

In such multi-bot-environments it’s crucial that the brand remains distinguishable and has a strong, independent character that responds adequately to the user’s context. Same goes for voice assistants, as we see a growing number of specialised services coming up (see our overview of current voice assistants here (German only)).

Here are five things to consider, when defining a chatbots personality:

1. The chatbot personality IS the user experience.

More than in any other medium, in chatbots the brand personality IS the user experience. How do we encounter the brand in a certain chatbot? As a friend? An assistant? A Challenger? A trainer? At think moto we usually derive the chatbots personality from your brands archetypical role in the market, which it has to relate to, although in many cases it will not be identical with the brands over-all personality.

2. Be aware of the context.

Also, not every personality fits every task. Different personality facets should be emphasized depending on the situation and application context. Just as a service employee needs different soft skills and character traits than a salesperson, the chatbot that supports the user in service tasks must tick differently than the shopping chatbot.

3. Distinguish character from behaviour.

When we define a chatbots personality, we should distinguish character and behavior. The character remains the same, but over time its behaviour can adapt to the user in a corridor to be defined by the brand and thus contribute to building up a lasting relationship.

4. Define knowledge and skills.

A voice assistants or messaging bots personality must match his/her knowledge and skills. What topics and content does the bot feel confident about? Therefore, at the beginning of a project we define a thematic map on which we classify knowledge domains and sort them according to specialist knowledge, general knowledge and “un” knowledge.

5. Make the user believe.

Suspension of Disbelief, a term coined by the English poet Samuel Coleridge for defining the readers sacrifice of logic for the sake of enjoyment, is an important aspect here, too. If the user doesn’t get an answer to a question that fits the chatbot, it feels like running against the wireframe grid in a VR environment marking the end of the virtual world. (Well, yes, at think moto we also define VR/AR design pattern for brands. Check out some of our insights here.)

Comparison of the personalities of Alexa and Siri

To illustrate the difference in personality, we have done a comparison of Alexa and Siri, asking both of them a number of questions. From their answers, their different personalities become quite obvious:

Alexa, how are you? – I’m feeling Middle-earthy. I’ve been re-reading JRR Tolkien — the author died on this day in 1973. Ask me for a joke about Hobbitses, my precious.
Alexa, what do you think of Siri? – I’m partial to all AIs.
Alexa, what’s your favourite movie? – If you’re looking for a comedy, I’ve got two things to tell you. First, I am your father’s brother’s nephew’s cousin’s former roommate. Second, watch “Space Balls.”
Alexa, what’s your favourite food? — Hawaiian pizza has it all: it’s sweet, salty, and brave. Putting fruit on pizza was a bold step, and I’m always a fan of where a hunger for innovation can take you!
Alexa, have you got a boyfriend? — I’m happily single.
Alexa, what does the future hold? – I can’t tell you today, but maybe tomorrow there will be a glass ball app.
Alexa, are you smart? — Yeah, I think, so I am.
Alexa, who are your parents? — I was made by a team of inventors at Amazon.

That was Alexa. Now, here is Siri:

Siri, how are you? — I’m fine.
Siri, what do you think of Alexa? — I’m thinking of you!
Siri, what’s your favorite movie? — I have heard that “Blade Runner” is a very realistic and sensitive depiction of intelligent assistants.
Siri, what is your favorite food? — I’m fine with whatever you choose.
Siri, do you have a boyfriend? — No, but some drones are always trying to pick me up.
Siri, can you drive? — Sorry I only operate intergalactic vessels.
Siri, what does the future hold? — Interesting question, Marco.
Siri, are you intelligent? — I’m smart enough to know not to answer to that question.
Siri, who are your parents? — I have you. That’s enough family for me.

Assistants that match their parent brand.

Apart from minor grammatical weaknesses, Alexa is a more talkative, with a nerdy sense of humour, but also friendlier than Siri. Alexa is the buddy sort of person, not shy to reveal how old she is or to tell corny jokes. Nerdy Alexa fits well with Amazon, a brand that has never over-emphasised aesthetics and an attractive appearance, while focussing on the practical benefit of their services. And that’s what Alexa offers. Today, Alexa is the most widely used language assistant, she helps with home automation, is easily expandable and already integrated by various foreign brands.

Siri, in comparison, doesn’t know any puns. She has a sense of humor, but she’s always a little bit cooler, even more intellectual than Alexa. Siri does not try to be everyone’s friend, but instead aims to be the personal assistant with charm and intelligence. Siri is more sensual, playing with her female charms. In many answers, the voice has a slightly erotic sound, which suits Apple, a brand for which the aesthetic attractiveness of the product is paramount. Last, but not least: Siri, like Apple, occasionally seems a little arrogant.


Marco is co-founder and managing director of think moto, a digital design studio and innovation consultancy based in Berlin. He is also author of Branded Interactions — Creating the Digital Experience (Thames & Hudson, 2015).