Conversational UX is the Ultimate UX Discipline
User Experience purists wet dreams 💦
I’ve always felt the graphic design legacy in product design, especially when it comes to the User Experience discipline.
A lot of today’s product designers came from a graphic design background, which can explain the influence.
Have you ever saw a UX designer in a movie or a Netflix show?
Plus, UX applied to digital products is still young and poorly known from the general public. Your teammates can be educated about UX, but again, a polished mockup feels better for them — me too btw —
« Ugliness sells badly,» said Raymond Loewy, what about Reddit or Craiglist, then?
Today, we’ve exploded UI with conversational experiences. There’s no more dropdowns, tabs, even no more text to display with Voice applications.
What’s left? No more infinite arbitrage about a radius, a color, or a shadow. Research is a must if you want to achieve adoption, and a scientist approach seems mandatory.
Is Conversational UX the ultimate UX discipline?
Less intuition-fueled decisions
First, we have to keep in mind that we’re still very early. It’s a new medium for designers, so for the end-user.
Speaking to a machine in natural language is nothing like we’ve experienced before. I was raised with a graphical interface; it seems logical to me to interact with a computer, using a mouse, a touchscreen, as long as there’s a UI.
It helps when it comes to demolishing products in design critiques by PM or stakeholders. There are fewer biases from experiences and less subjective decisions, thanks to stripped-down UIs. You don’t need to show low-fidelity mockups; your prototype is not much than prompts.
The tone of voice, which is the equivalent of art direction but for words, is still a debate. But it won’t be a problem when will be generated prompts thanks to NLG. The assistant will create the answers from scratch, using an extensive database as a knowledge base.
Research First
The first wave of chatbots helped us to have surface-level best practices. We do have a decent idea of how to write for conversational experiences.
But we don’t have user adoption at scale to have research that is authoritative in our fields.
The complexity came from the various level of education/expertise from users to this medium. This result of a very varied usage than we need to capture and understand to offer a delightful experience to every user.
Plus, since the medium is language, culture plays an essential part in the overall experience a user can have to interact with a chatbot.
Those constraints lead to the need for User Research.
Scientist approach
After years of designing conversational experiences, I can tell you that focusing on UX writing has a poor ROI on your overall experience. The UX is mainly impacted by the quality of the dataset of your assistant.
Conversational are learning products; you have to track the impact of the data you will add along with the life of your assistant. Approaching it like a scientist seems like a great idea then.
Planning, experimenting, measuring.
Plus, datasets can be polluted by biases. If you’re using data that you’ve created in the training set, the assistant is performing well when you’ll be using it because you will use the same language, but not everybody has to express themself than you.
There’s a lot of biases that can come into action when you’re designing with AI; it will create new roles — reinforcing the idea that the future of designer will take place into a lab, not a meeting room covered with post-its.
Convinced?
