Design systems for conversational AI
Stop reinventing the wheel
In a previous newsletter, I talked about how persona isn’t simply an add-on.
Teams need clear rules, descriptions, and guiding principles to create cohesive, consistent experiences. Personality can be used as a decision-making framework for any design choices you need to make.
My prediction is that as the industry matures, this will ultimately bring about the need for design systems for conversational AI, supercharged with personality.
New attitudes and best practices
Nielsen Norman Group defines a design system as:
“A set of standards to manage design at scale by reducing redundancy while creating a shared language and visual consistency across different pages and channels.”
Well-known design systems are Google’s Material Design System and IBM’s Carbon. Traditionally, these systems focus on the design and development of digital products and services.
Conversational AI has thrown us a bit of a curveball, though. It’s a mix of everything we know about digital design. It encompasses conversational copywriting, NLU design, behavior design, UX design, visual design, sound design, and more.
As a result, people have been forced to develop new attitudes and best practices for designing good experiences.
Do I really need a design system for that?
Large organizations often have several teams running different conversational AI operations. However, they rarely make efforts for knowledge sharing or an overarching strategy on documenting standards and best practices.
Developing a design system might seem redundant when you have only an AI Assistant deployed in one language. Wait until you’re tasked with coordinating efforts for different deployments (IVR, Web, App) and have teams running localized versions of the same AI Assistant in different markets across the globe.
Are you going to let every team make the same mistakes? Reinvent the wheel, time and time again? Right now, the answer seems to be yes.
Personality-driven design systems
Design systems for conversational AI will be personality-driven, meaning: the persona of your AI Assistant will sit at the heart of your design framework. This will inform a clear set of rules and guidelines on the use of language and the tone of voice.
After that, you want to focus on how different modalities impact the experience. Not only do you want to adhere to conversation design best practices (such as acknowledgments, confirmations, discourse markers), it is also important to incorporate knowledge from other disciplines like UX- and interface design, visual design, etc.
Ultimately, this will give rise to a component or pattern library: a collection of reusable components, design principles, and writing guidelines, that functions as a central repository for your conversation designers and AI trainers.
It will serve many purposes: ensure consistency, create a shared language, help with onboarding new team members… but most importantly, it will allow you to jumpstart new conversational projects in a matter of days, instead of months.