Conversational interface for chatbot & voicebot: the French touch
An interview with Amina Esselimani, top French specialist UX designer.
Now a new form of expression is available to brands: they can express their image through a specific and more friendly tone of voice, so users can feel more included and emotionally linked.
Can you explain to us what a conversational experience is?
It’s a basic service: a virtual assistant that helps people with daily tasks, or to quickly obtain information.
A conversational interface enables users to interact with this service, or with a brand, in an easier and a more natural way. Customers only have to write a few lines or only have to speak: “What’s the best deal for a trip to Australia?”, “I want to book an Uber”, “How much is an electric bike?”. This conversation can be accessed from a website or a mobile app, and also via typical daily channels (e.g. Messenger, Google Assistant, WhatsApp, Slack, etc.).
These days, more and more devices are dedicated to a user vocal experience (smartwatch, speakers, GPS, etc.), fitted out with microphones and AI, like Google Assistant for Google Home or Android devices. A lot of brands choose IoT so they can offer their clients a more natural and direct service with vocal input.
These kinds of products offer customers new options which are evolving very fast. As designers, we have the responsibility of fully exploring their potential and we need to make sure we design experiences adapted to user’s needs.
How can we explain the popularity and fast development of these tools?
With a 100 % live vocal interaction non-digital people can use a service like anyone else.
- A more human-like interaction: Users can express their needs, or ask their questions in any way they like, in any order they like. The bot will adapt and respond.
- Accessibility: It’s a great opportunity for non-digital people, especially the visually impaired, to access a digital service! Currently their only option is a mechanical synthesised voice reading out all screen content. With a 100 % live vocal interaction they can use a service like anyone else.
- More efficiency, fewer constraints: The conversational experience helps to decrease the cognitive load and will ease the user journey. On a typical website, the user has to carry out a certain number of steps including: interface analyse, comprehension of the website, product identification and selection. Through a conversational experience, the user’s cognitive load is significantly reduced thanks to AI.
- Time saving: Long waiting times on the phone, endless redirections, and repetitive tasks, can now be avoided!
I guess there are business advantages too?
- Brand identity: Now a new form of expression is available to brands: they can express their image through a specific and more friendly tone of voice, so users can feel more included and emotionally linked.
- Increase key knowledge: Brands can combine these new tech tools with their user base to create a better customer experience. It can save valuable time for sales people and customer services, and free them up to focus more on a good quality service.
Users shouldn’t have to look for information by themselves anymore.
- A bigger connexion: Today, I think e-commerce companies should be more connected to their clients via the daily channels they use (Messenger, Slack, etc.). In my opinion, users shouldn’t have to look for information by themselves anymore. That’s why, with their authorization, the OUI.sncf bot sends personalized train times directly to the passengers, when it knows they need it at a specific time, so the users don’t have to request the information at all.
What are you currently working on? What are your challenges?
I currently have the opportunity to explore different sectors, like transportation with OUI.sncf, the medical field with the DeepOP start-up and service delivery with the SMARTLY start-up. I have designed customized vocal experiences and bots for platforms like Messenger, Google Assistant, Google Home and Alexa.
These projects enabled me to approach vocal use in several completely different contexts. For example, a nurse, who gives medical data to a virtual assistant, will not have the same conversation as a customer buying a train ticket. Also security and legal challenges aren’t the same. A conversational experience for family connected speakers shouldn’t be designed in the same way as a personal device.
Do you see differences between B2B and B2C?
A crucial part of the job is to work on the bot persona, to define its personality, tone of voice and specific goals.
Yes! For B2C applications, the main challenge is convincing users to adopt this new tool; to be successful a tool must be relevant, especially if you want users to change their habits. The key is the trust users are willing to give a brand, and the emotional attachment between them. A crucial part of the job is to work on the bot persona, to define its personality, tone of voice and specific goals.
For B2B, if you are designing an HR or a legal conversational tool for example, the challenge is more about efficiency.
I know you worked on mobile apps, interactive kiosks and websites. What do screen content design and conversational design have in common?
The design approach is just the same! You have to know end-users very well. My work is also based on personas and on the study of customer needs and pain points.
However, we have a lot more knowledge about website or app design. For conversational design, especially voice interaction, there is no design pattern and no best practice exists. I had to start from scratch.
After a lot of testing, I have been able to identify some best practices to make information processing easier. In a vocal experience, humans have a very short-term memory, so if a voice bot gives three or four precise pieces of information in the same sentence like train times, users can usually only remember half of them.
How do you work?
After the user research phase, I design the “happy path”, i.e. the conversation when everything is going well, then the other cases (e.g. the user says “I don’t know” or “no”, etc.). Finally, I have to focus on “repair conversations” which is equivalent to the error messages and redirections on a website.
The more you anticipate specific cases, the better the experience will be.
A huge part of the work is to analyse and identify contexts in order to adapt the design to typical conversations in a given setting. I use macrographs to represent the whole user experience, which gives an immediate overview of the project and also helps to define product scope and goals. The other main specific deliverables are Context map, Intelligence rules and Dialog flows. The more you anticipate specific cases, the better the experience will be. In particular, we pay great attention to non-response scenarios.
Finally, I also include a plan to measure success (KPIs), including giving users the opportunity to say if they are pleased with their experience via pre-set answer suggestions (“Thanks”, “Great!”).
What advice would you give to UX designers, who want enter this field?
First of all, do not change the classic design approach! Like in any project, do a lot of co-design workshops with your client to define your persona bot, identify and prioritize your goals, and choose your wording very carefully.
Above all, test and test again, I will never stop repeating that. Your creation is worthless if it hasn’t been tested on real end-users. Feedback from user testing is essential and irreplaceable: “I never have a problem buying my ticket, the problem is finding the train”, “I had to interrupt the bot to give it more medical information”, “I feel I am forced to choose from only a small range of products. I want to have more choice”. You never know in advance if you are going to have to change your project strategy at any given time.
Don’t hesitate to invite your client to assist in test sessions. In my experience, a “Wizard of Oz” method of VUX (Voice User Experience) testing thrilled a lot of my clients because they could see how their product could be used by the users once developed.
Finally, if you are not good at writing, never hesitate to work with a copywriter!
What do you see in the future for conversational experience?
I believe in the automation of recurrent tasks to relieve humans of boring work. Bots can help humans in a non-intrusive way (notification) to remember or become aware of myriad of things.
Most of my clients choose these new platforms for business reasons. My goal is to make them realize that without a positive experience it’s hard to guarantee the adoption of a conversational tool. Given that GAFA offer conversation templates adapted to e-commerce, client service, etc., we can focus on the value and quality of the service we want to deliver.
How do you rate France in relation to conversational tool design and development? Are we leaders, good students or only beginners?
I’m optimistic so I would say we are good students. Tech is here to stay and companies are aware of the UX approach. There is no doubt, in a near future conversational experiences are going to be more and more popular!
Amina Esselimani is a conversational UX designer. She graduated from L’École de Design de Nantes Atlantique in 2011. She’s passionate about user research and testing applied to chatbots and voicebots (B2B and B2C). Putting AI at the the service of the user experience is her daily challenge.
