Static to dynamic. Single modality to multimodal. Flip-phone to Smartphone. Every interface and design goes through a critical moment where designs and prototypes must adapt to predict user behaviours, become responsive, and therefore – smarter.
Conversation design, both for chat and voice, is no different. In the early days of conversation design, flowcharts and pre-populated paths dictated the user experience. Now, designs must learn to accommodate and predict user-led variants in conversation, all without ever breaking the user experience.
Introducing Smart Prototyping in Voiceflow — the easiest way to test, train, and teach your conversation design to handle UX variations using our native NLU/NLP training system.
What is NLU/NLP?
NLU stands for natural language understanding, while NLP stands for natural language processing. These mechanisms transform raw inputs from the user (i.e., what the user says) into readable formats for the platform and conversational AI.
Where does NLU/NLP impact your conversation design?
NLU/NLP truly impacts your design when it’s reached a high fidelity version of your experience. Whether you're evaluating your conversations as you design, running user testing sessions, or demoing for stakeholders, the NLP/NLU can help you easily mimic a production-level conversational experience.
What does this mean for the design process?
Although NLU/NLP can impact all stages of your design process – it is the most impactful when handling higher fidelity designs or user validation.
When designers use Figma, they often go through a multi-stage design process. The first stage resembles more of a skeleton/low-fidelity version of their design, whereas later on, their goal will likely shift toward design parity. As the design matures, the prototype’s goal shifts from testing a hypothesis to reaching parity and validation. The closer their design represents a 1:1 experience when live, the better their validation and learnings will be.
When a Conversation Designer begins their design, their focus is nailing down the initial structure and happy path. As the design progresses, their focus soon shifts toward creating a realistic rendition of what potential paths a user may take and how to accommodate them in their design. When a design is ready to be tested, whether in a predictive or customer-focused testing model, NLU/NLP can play a crucial role in understanding the user's intents and reacting dynamically to their responses; this will create a smoother experience for the end-user and a much more flexible experience for the designer to launch.
What does this mean for the prototyping experience in Voiceflow?
At Voiceflow, we define the conversation design workflow in three parts:
Our goal has always been to create a full-stack tool that helps at any stage of the conversation design workflow. With that said, adding a custom NLU/NLP into our native prototyping experience not only improves the prototyping experience of our tool but also bleeds into the overall success and predictability of our build or launch experience.
When we think about prototyping at Voiceflow, our current model works wonders for many low fidelity tests and experiences. However, with the addition of NLU/NLP, designers can now conduct higher fidelity user-testing all from within the tool.
As the industry continues to grow, so are the use-cases for conversational experiences. It's important for designers to have access to tools that enable them design and test for those use cases as realistically as possible. With Smart Prototyping & our native NLU/NLP, we now support various new use cases and testing opportunities, in combination with our General Assistant projects. This will make it easier for designers to test at a higher level of intelligence and teammates/user testers to experience a more representative version of your designs.
Empowering General Assistants & Beyond
Voiceflow projects already benefit from a direct connection to Alexa and Google's native NLU/NLPs. Alexa and Google, however, only represent a fraction of the possible conversational channels.
As the industry progresses, designers will come across various NLU/NLPs - many built for a specific purpose. It will be up to the designer to choose what suits their use-case best (i.e., an NLU built for healthcare vs. customer success or an NLU focused on financial security vs. entertainment).
Ensuring that we offer a smarter, more intelligent experience at par with our Alexa and Google offering is critical. It also represents a big step forward for the future of Voiceflow.
Why is NLU/NLP training important for prototyping?
A Conversation Designer's job extends far beyond how things 'look' or 'sound.' Today, Conversation Designers need to have a strong understanding of user experience design, as well as the underlying tech that supports their designs.
When prototyping and training your NLU/NLP, designers can dive a layer deeper than surface-level happy paths. In fact, when designers train their experiences, they’re searching to understand how their flows will work in practice; they’re looking to validate their user experience.
Understanding how flows work in practice or under pressure helps unearth potential conflicts or overlap on the overall user experience. Ultimately ensuring the transition from 'design to prototype to user ready' is as seamless and all-encompassing as possible.
Designers can now easily access NLU/NLP powered tests more rapidly, unlock increased understanding of their experience, and enable more rapid design and prototyping cycles.
Best practices for training your prototype
Testing is all about learning, and training your prototype is one of the best ways to learn about your user experience. With that said, we recommend training your prototype whenever you change something meaningful.
The more you build, the more you should be training – and training your most recent model will be the most accurate depiction of your real user experience. Training is important not only to understand where you can improve your design but also to navigate:
- the model structure
- the logic and sounds
The beauty of conversation design versus traditional visual design is that tools like Voiceflow allow designers to test and design in tandem. With a visual design or test, designers often need to upload their models to Invision and create a walkthrough. Whereas in Voiceflow, any designs created on-canvas are created with parity in mind — and are therefore ready to be published. This workflow allows designers to stay consolidated and reduce the time to takes to test high fidelity prototypes.
Some features that make it easier for designers to test in Voiceflow include:
- Starting a test from anywhere on your canvas
- manually navigating your tests
- Allowing designers to click “back” if their test is unsuccessful
At Voiceflow, we’re building tools that support every stage of a team’s workflow – from design and user testing to hand off and development. Adding additional NLU/NLP training to our prototyping experience is a dynamic feature that helps teams and professionals do their best work.
The industry is reaching a stage where designers expect runtime, predictability, and smart designs inside the tool they're using. And so we're working hard to ensure that those expectations are met. By investing in a smarter prototyping experience, we're making strides toward supporting these asks and others - such as launching onto new platforms and simplifying complex design tasks.
If you’re interested in testing out Voiceflow, our Smart Prototyping feature, or any of the features available in our canvas — you can sign up for free here.
We can’t wait to see what you build next,
– The Voiceflow Team