Brand voice: BotCharacter Design principles

Lessons from linguists and Pixar character designers

Published in

Prototypr

10 min readJul 29, 2016

Pic by Voxable, a design firm that builds custom AI-based conversational and voice interfacesAI.

There has been a lot of discussion around conversational UI and how conversational commerce is a dominant trend of consumer computing apps in 2016.
We’ve seen how messaging apps have eclipsed social networks in monthly activities. After the launch of ‘bots on Messenger’ Facebook opened a brave new world of direct channels between customers and brands.
Chatbots will play a main role as a primary touch point of the users to services.

One of the first integration on Messenger has been Uber, Lyft and maps.
How many of us who have called DMV and been on hold “forever” wouldn’t love to chat with a rep and get our issue handled in a minute.
How many of us would love to chat with an airline and resolve an issue in real time?
How many would like to shop easily saving money and time?

Every community, marketplace, on-demand service, dating app, social game or e-commerce product has or will soon have messaging as part of the experience to drive retention, engagement and transaction volume.

Key reasons for a Conversational Experience

First and foremost this is the right time. An entire generation of mobile-native consumers are comfortable and fluent with messaging as an interaction paradigm. Then the key reasons to explore a Conversational Experience for your business might be:

Bypass incumbent vertically integrated player such as Apple and Google.
Remove expensive humans from customer communication.
Differentiate and extend the distribution channel and branding values. Someone says the end of apps and app stores is coming. (is it?!)
Tailor recommendations based on personal interests and profiles.
Address customer’s needs in real time providing relevant content.
Create opportunities for services integration.
Open a direct and honest channel between brands and customers.

Are bots always an answer?

No. Sometimes the excitement behind a technology makes us forget that humans might be better in offering a service insider knowledge instead of bots. It’s crucial to keep in mind the limitations and affordances of the bot experience.

Airbnb VP of Product, Joe Zadeh, stated that only humans — in this case, Airbnb hosts — can provide authentic hospitality and help foster the emotional relationship to a new place that travelers often seek.

“Bots potential is in doing things that they can do better than humans, like logistical tasks and taking as much work off the host’s plate as possible, so the host can focus on personal hospitality. Technology is not antithetical to humans but should be used to bring out more humanity,” Zadeh said.

So what we can learn from Zadeh experience in running a 60 million users service (valued $25 billion) is that building bots is not always an answer.

Here are just a few of the constraints and affordances to keep in mind:

Discoverability of the interface and working memory
Features/security of the messaging platform
Storing user preferences
Support for financial transactions
Push notifications
Using voice controls

Your bot could fall completely flat if it relies on the novelty of the latest tech trend rather than the needs of your users. Therefore we need always to ask ourselves what a bot can do better then a human.

Lessons from linguists

Linguists have certainly thought a lot about conversations and as the designer role is shifting from designing interactions to designing relationships and conversations I think linguists might teach us something.

Here’s a sentence diagram; something linguists use to map sentences and break them down into their structural components, verb, subject, object, and so on. Think of it like musical notation for words.

Each sentence has its own whole architecture, its own meaning and structure. It could branch and flow in all kinds of directions.
How does a UX designer model the conversation flow?

In all the excitement of diving into this new technology most of us seem to have forgotten about the most important actor in this enterprise: the human being who is expected to talk to the bot.

Conversations usually consist of two kinds of interactions — the structured (forms, menus, choices etc) and the unstructured (freestyle plain text).

Few examples from an offline structured conversation: ordering from a menu at the restaurant, filling a form at the doctor’s office, sending an invoice to your customer. On the other hand, conversations with friends, family, teachers, colleagues and acquaintances tend be unstructured.

The conversational interface designer must avoid force-fitting a naturally structured conversation into unstructured format, and vice versa.
Thankfully, there are tools being developed for programming bots to handle unstructured conversations (e.g.Superscript) while others are taking the “AI” approach (e.g. Api.ai, wit.ai, getmyra).

Lesson from Pixar characters design

*Inside Out: Early production sketches, help the writers visualize the characters they are still developing.*

How designers at Pixar are bringing to life new characters, can teach us so much on how to get birth to a bot character.

The connection of the audience to the character is the primary goal when designing a Pixar character. Imbuing the characters with humanity, and focusing on their humanity is vitally important to creating this connection. When designing a character, a Pixar artist needs to know what to exaggerate and what to play down, what to add to give a hint of background and depth, and what to do to develop personality.

Think about the meaning of the word ‘character’. You’re supposed to breath life into these things, make them appealing and give them the magic that will allow people to imagine what they’re like to meet and how they might move.
Neil McFarland

If every service is entering messaging channels how each chatbot should behave interact and chat with us and ‘be unique’? Which tone shall they have? Should they chat as one of our friends, more as a loyal personal assistant or keep their robotic soul?

Should they behave as an individual movie character with a defined personality or collaborate with each other as teammates in helping achieving our goals, and supporting what we really care and need?

When opening an app we usually follow an established and structured pattern of categories, folders and buckets, but if we have a free flow conversation with a smart chatbot is like an improv theater show compared to a normal movie that has a fixed beginning and end.

Principles of Bot-Character Design

Here what I extrapolated from linguists and Pixar designers lessons on how to design a bot personality character to help brands better connect with their customers through a converstaional experience.

1. Make it relevant

Think of what your bot can provide as unique and relevant type of content. Bots aren’t people, and shouldn’t pretend to be. Bots are great for making interactions efficient and scalable, the goal is to develop a unique and richer connection with your customers.

2. Find the right voice

Craft the right words and tone. Messaging is intimate and shouldn’t sound like marketing or corporate-speak. Every business will sound different, but think about what’s appropriate for your market, your audience and the medium. Would you like to hear entrepreneurial or inspirational quotes from a sock company or jokes from a funeral home? The point is: find whatever voice speaks to your consumers the best and stick with it.

3. Who is it aimed at?

Think about your audience. Characters aimed for a wellness Agent should be more bossy or more sympathetic? An agent for fashion and style should be more friendly or more ‘Devil wears Prada’ way? Well it depends both on your branding core values and on the audience. The Bot-character should react based on the unique audience personality, interests, goals and needs.

4. Structured VS unstructured format

Avoid to force-fit a naturally structured conversation into unstructured format, and vice versa. The key is to assemble the right medium at the right time. When you need a precise answer or when you are providing a set of choices, the structured format is the answer, but when you wanna give freedom of speech to understand the real needs and interests of the audience the unstructured freestyle plain text it’s for you.

5. Make it chat-able (but to the point!)

Messaging is its own medium with its own pros and cons. Conversationalizing something already done well on a website will often make the experience worse, not better. Instead of recreating experiences that are already good, ask yourself: How can I enhance a trustworthy, honest and loyal relationship (channel of communication) with my audience? Focus on that. Keep the interaction simple, efficient, and natural and people will come back for more. So keep in mind the end goal of your customers, chat with them but go straight to the point!

6. Convey Personality: Goals and Dreams

A bot-character personality can be reveled through tone of the voice, UI colours and animations. Your bot-character should express a range of emotions to express itself. Give your character opinions. The driving force behind a character’s personality is what it wants to achieve. Often the incompleteness or flaws in a character are what make it interesting.

7. Staging — Beyond the character

In the same way that you create a personality for you bot-character, you need to create an environment to help further cement believability in your creation. The world in which the character lives and interacts should in some way make sense to who the character is and what it gets up to. So in this case the overall interface experience should behave as the bot-character as colours, animations, textures, patterns and plot twists.

8. Timing etiquette

When we are chatting with a friend, a colleague or a shop assistant our rhythm and way of talking may vary in tone and timing.
You may interrupt a friend while talking, you may be more indulgent to a shop assistant, or vice versa. The same applies to your bot-character. When and how it should interrupt your customers daily routine with relevant information? Timing of conversation and interruption is critical for establishing a character’s mood, emotion, and reaction.

9. Why it should exist?! Could it be done better by a human?

The lesson from Airbnb is key to understand what could be done better by the bot and what from humans. Is the bot-character generating a long lasting, loyal and trustworthy relationship with the audience?
Why it should exist and which problem it should solve?
If you answer to these question, half of the job is done for your Bot-Character Design!

How to measure the effectiveness of a conversational flow?

Just as in traditional UX design, a good place to start is by comparing the effort required by the user (input) to accomplish a certain goal (output). This gives a measure of the “efficiency” of the conversation. For a given goal (such as ordering pizza or a jacket), the more efficient conversational interface is the one that requires the fewest clicks, words, responses or iterations. Maybe the most efficient pizza-ordering conversation is simply “the usual” since the bot knows everything else from prior history.

And how does the designer debug the conversational flow? Drop-offs, abandoned conversations, delayed responses, extra iterations, re-dos, are all indicative of errors in the flow. Tracking these metrics is key.

Near future of conversational technology

Viv.ai ‘the intelligent interface for everything’

Design is about drawing connections and delightful relationships between people. Bots should not over promise and under deliver.
So long as its capability matches the expectation set with the user, the experience will be great.

As this space evolves, conversational interface design will become a lot more analytical than it is today. As more developers build bots and share best practices, we will better understand rules and techniques of conversational interface design.

One of the most promising technologies in this field I feel is Viv born by the creators of Siri.

Viv is the next AI generation assistant, offering the ability to connect with third-party merchants and vendors so that it can execute on requests to purchase goods or book reservations (similar to Amazon’s Alexa and Facebook’s Messenger bots). The company’s tagline — intelligence becomes a utility — nicely sums up its goal of powering the conversational AI inside a multitude of gadgets and digital services.

Bots may soon address a wider range of consumer needs — by anticipating them through service integration and personalization.
But chatting through an interface or by voice are not the only element of the human experience making a comeback.

*Microsoft Emotion API showing facial scoring*

The emotional intent facial recognition, along with hand and body gestures, will become a critical addition to the integration of the conversation and voice first future.
Combining a real-time decoding of the 43 muscles that form an array of expressions, can communicate not only deeper meaning but deeper understanding of volition and intent.

The last 60 years of computing humans were adapting to the computer. The next years the computer will adapt to us. It will be our voices that will lead the way.