Artificial Intelligence is burdened with the connotation that it is complex, entrusted to technical experts only, a fact that is true. As a designer, I've always been motivated by taking something complex and simplifying it. I want to make creating AI easy. Why does AI have to be complex and only for experts? Many ordinary people think about AI opportunities everyday, but lack understanding and resources. How can I provide an opportunity for people to access and build their own AI?
Preface
I spent quite a deal of time using website builder services like WiX and Squarespace, to build my professional online portfolio and personal blog. These site builders automate functions like screen and layout responsiveness, page design templates, text and component styles, navigation and links, and SEO capabilities, all without requiring code, making it easy and efficient to build a website by anyone for any purpose: professional or personal. Of course this still requires some basic knowledge of site building for nuances like: consistency, spacing, content and terminologies, hierarchy, color theory, etc. For the most part, a normal person can get a website up and running quickly within hours.
This is the level of automation and simplicity I am striving for in AI chat bot development and conversation design. Anyone can potentially utilize a website service to create an AI chatbot efficiently on their own without code. Creating AI chatbots has traditionally been limited to corporate or professional consumer service domains, but why not extend this to people who just want to create their own personal chatbots, which can be used publicly for free or can be productized as an app that serves a business purpose.
Mission
To enable normal people to simply and efficiently create conversational AI chatbots for their personal and professional needs, provide a service that is easy to understand, automates complex processes, and is independently and available online.
A simplified process
The steps would generally mimic that of simply building a website and would look something like this:
1. Name your bot
2. Choose your channel or modal
3. Enter your domain
4. Choose your vertical template
5. Design your flows and bot messages
6. Enhance your model
7. Preview your bot
8. Publish it to your channel
9. Continue to optimize and improve
Simple right? Yes....and (or but?). And! There are opportunities to automate many of these steps that in today's world still require sophisticating and complex processing and manual code support.
Complex AI technology that potentially can be automated
Natural language processing and intent classification can be designed through straightforward model building mechanisms and workflows which are analogous to configuring website properties for SEO optimization, keyword tagging, and topic categorization.
Logic building such as API integration, entity construction, and conditional rules, can all be designed with straightforward mechanisms and workflows, relative to website linking and live site mapping.
Dynamic, interactive, and rich content creation such as applying and customizing info/form cards, quick replies, selection lists, product carousels, and other chat mechanisms, can be designed with straightforward mechanisms and workflows, relative to website components such as buttons, search functions, filter/sort capabilities, images, video, etc.
Conversational journeys and bot responses can be templated and standardized analogous to website layout and style templates that best suit businesses or personal needs and domains. Styles are applied to best represent the business or personal tonality and voice that's most appropriate and aligned with brand.
Styles for rich text display and voice synthesis can also be configured simply, just as website styles give users the capability of customization.
Complex technologies and services that are automated today
Experts were required to do technical complex jobs, but today, tools and technology has made it easy for non-experts to accomplish and create their own products, media, and services. Below are some examples of traditionally complex jobs that have been simplified with tools.
Basic app creation
Website development and design
Digital media production and editing
Mobile Interface layout design
Merchandise and signage design and printing
Achieve the basics
There are many other consumer household and business enterprise services and products that have simplified tasks that required experts. You can argue that these complex and sophisticating technologies still require a level of detail from the user. We can challenge those requirements and offload those burdens from the user, while focusing on the simple tasks and key important necessities. Think in terms of minimal viable product (MVP). Define the basics, user needs, and optimally design for those cases. Let the technology handle the hard stuff. As a UX designer, protect your users from the technicalities, from the edge cases, and from requiring them to solve complicated problems.
Summary
Make creating models easy. Make conversation design easy. Make testing easy. Make reading reports easy. Put all these things into one product experience and you can make creating an AI conversational chatbot easy. There is a lot of work to be done to automate and streamline AI tech processes, but I believe we will get there, as long as we find the appropriate interactions that best represent the technical needs, user goals, and mental models.
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.
Comentários