Machine Learning — The New UX Process
When we talk about UX we area often thinking of wireframes and prototypes. Even as UX practitioners we often fail to think of the users experience beyond the terms of a formatted journey map or set of process evaluations.
Taking the traditional approach to user experience with algorithm based experiences we are unable to directly chart the path a user will take. We find ourselves in a world of looking for the value to add in possible situations and having to no longer curate a process for a specific user, we need versatile processes that adapt and change as the situation adjusts around the users path.
How can we approach this?
This is something that is still being explored but comes down to the process and systems we have in place to enable and empower users. Why would we build one amazing experience when technology will let us build endless possibilities of them?
The key to addressing this type of problem is not inherently the solution; rather, the ability to provide a solution. If we look at the “ideal” process for the user experience practice we [generally] find identifying the problem at the start of the process and a majority of the process is dedicated to truly identifying and validating with users the true problem. Once we know the problem we can then address the problem being guided by user feedback and iterative validation.
If we can quantify this it would look something like:
Identify problem — talk to users to understand possible solutions to problem — work to address the problem — validate and refine with users
When we look into applything with a situationally dynamic problem we see that it changes to be something like:
Receive user based request — evaluate data on hand and match to previous connections with similar historical data — generate predictions based on available data [lack of data] — present option — allow user to provide direct feedback and refine — log options and changes for further refinement of future visitors
One of the initial noted issues with this approach is that this is a learning system and unless you are injecting hard coded connections this will need to be seeded or have a very low experience expectation at the beginning of the process.
The advantage of this approach is it is 100% user centered and follows the users direct requests.
Is this truly UX design?
Looking into this type of process as a new form of experience design we see large challenges and problems that are becoming more prevalent, as UX practitioners we find ourselves in a unique place to address these experiences and pioneer a design system that transcends visual design and truly takes on experience design.
How can we accomplish this?
By taking an augmented approach of the classic process we find that we can quickly apply Learn — Validate — Refine to the process. This also forces us as designers to push the visual design patterns to be more adaptable and work within code to contextually change how we render visual cues an accept that visual interface takes on a new role as travel guide and not entertainer. Content and relevancy becomes key and visual becomes brand and adaptability.
One of the key advantages this takes on is that it supports directly the concepts of a core UI code base that is quickly adapted and rendered to match user needs while communicating brand. Experience becomes truly about a consistent yet personable experience that is guided by users and empowered to evolve and scale.
Is this a users dream?
Image an online shop that leverages your history, keywords searched to help you find items you may be interested in while matching to similar profiles and helping you to discover things you didn’t yet know about. A truly user based experience like this would not only drive consumer satisfaction but, it empowers a new business model that values UCD (User Centered Design) practice to drive customer satisfaction and sales based on user specified value.
If you enjoyed this please follow for more!
Twitter: @jmcgarryiv LinkedIn.com/in/jmcgarryiv McGarryIV.com
