Designing the [24]7.ai Conversations gave me the opportunity to dive deep into the world of Conversation Design. I had to design a tool that enabled conversation designers to design AI chat experiences: conversation flows, bot responses, and all bot logic and rich functions to support those experiences. Conversation design, similar to UX design, is very sophisticating, tedious, and complex. In this article, I will discuss some key principles that we ensured to maintain in our product experience for our users.
Preface
When I took on this project, I knew I had to become an expert on understanding conversation designers and their thought processes and workflows for creating a robust AI chatbot experience. I attended industry remote conferences and webinars, like: Conversational Conference in San Jose (in person), VoiceGlobal (remote), Level 3 AI Assistant Conference (remote), and other small-tier talks that hosted domain-specific topics and AMA's (Ask Me Anything) with industry conversation design leaders. I had recurring 1:1 interactions with our internal conversation designers at [24]7.ai and also examined their behaviors and needs through concept/beta testing and usability studies. I examined competitor products and similar applications that resembled our product goals.
This article will NOT focus on how to do conversation design, rather it will focus on the necessary product experience principles that enable and support conversation designers to optimally create their conversational AI chatbots.
Mission
To ensure conversation designers build chatbots efficiently, sustain these key product experience principles in the conversation builder platform.
Principle 1
In the moment
Keep the designer in their desired position, then provide contextual functionality to them. Allow them to create functional needs as they are designing conversational flows and bot responses. Design a system supporting a mindset that allows for the now as opposed to a mindset of 'build-your-parts-first' outside of your flow, then connect later.
Enables:
Convenience, speed, and ease that reduces the time to design, optimize, and build their bot, so they can get their bot to their end-users faster.
Focus, flow, and persistence that reduces design-time by not requiring them to navigate somewhere else to perform a task or break their focus.
Ideation: Ideas leading to other ideas. The process is organic, evolving, iterative, free, and ever-flowing.
Context: Determining and declaring the appropriate requirements and needs as they design responses and flows: identifying use/edge cases, interactive patterns, and rich content.
Principle 2
Relationships and connections
Provide a graphical interface that allows designers to visually see how their conversational design is outlined, displayed, and is being experienced by the end user. Using flow diagrams, graphs, and enabling WYSIWYG (What You See Is What You Get) mechanisms allows content to be edited in a form that resembles the product.
Enables:
Identifying the appropriate flows and bot responses needed to further optimize their conversation design: intents/sub-journeys, rich content, and edge cases.
Scanning, finding, and managing at a high-level, so users can optimize sequence, move, connect, and relate nodes in a visual diagram.
Previewing of the chatbot and experiencing it as the end-user would, helps designers to make further optimizations.
The reduction of cognitive load required to imagine, assume, and attempt to correlate the experience on their own.
Principle 3
Rapid validations
Allow designers to quickly validate their design-time changes through run-time experiences. Designers instantly want to see how the change will be experienced by their users one at a time, how the change impacts the flow, and how the change adapts to the channel's display. Bot building will need to be fast. Designers want to test after each incremental change.
Enables:
Quick reviewing of change effects and outcomes leading to faster optimizations and faster deployment times.
Easy management and tracking of bot updates or changes: keeps the designer focused on one particular change as opposed to multiple changes, which can be a burden to remember.
Efficiency and accuracy attributed to single and focused changes instead of bulk changes.
Principle 4
Independence
Make bot building accessible to conversation designers. Strive to optimize for designers only, ensuring that they shouldn't require dependencies on other associates to build a bot. Find opportunities to automate processes that allows designers to move forward on their own time, and quickly.
Enables:
Reduced time to build, test, and deploy a bot.
No dependencies: can work in isolation, work faster, and ability to move forward on their own.
Better management and organization of work and parts, which also attributes to single source of truth.
Principle 5
Channel adaptation
Traditionally, each channel and modal would require separate apps with different logic and content adaption. This increases the time to design and deploy. This was also the effect of modal expertise. The user behavior is changing to generalists who will acquire all responsibilities of creating AI chatbot experiences. Design once, deploy everywhere.
Enables:
Reduced times to design, test, and deploy to other channels and modals.
Seamless and automated translation and content adaption to other channels and modals that will not require manual work to optimize any longer.
Only one designer to design conversations, which can be a generalist that does not require modal or channel expertise.
Principle 6
Clear separation
If bot app development can't be fully automated and still requires multiple practitioners to build parts, then ensure there are clear and distinct locations specific for each expertise. Don't mix functions relative to each practitioner: model building, QA testing, data analysts, linguist experts, etc.
Enables:
Clarity: reduces confusion about which practitioner does what, what their responsibilities are, when and where to do what, and the technical requirements needed.
Focused: streamlined workflows, functions, and job-to-be-done per practitioner.
Clean user interfaces tailored to each practitioner or expertise with key objectives and goals.
Summary
These product experience principles will always serve as our baseline and quality assurance criteria to measure, prioritize, and validate new features in our product. These principles will also guide the design direction and integration within our product. The [24]7.ai Conversations is a simple, intuitive, and usable experience, proven through user testing, and accredited by the sensibility of our product experience principles.
About the author
Sean Lazo is a Principal UX Designer at [24]7.ai who leads the inception, assembly, and design of [24]7.ai Conversations, an industry-leading omni-channel AI chatbot SAAS platform. His passions are DesignOps, detailed design, research, and human relations.
Kommentarer