Juniper Networks · 2022

Mist AI Auto Placement

Designing an automated solution to save time and increase accuracy for location services administrators.

About

Location services administrators spend significant overhead manually setting multiple APs for a floor-plan, which can multiply across thousands of stores in a franchise. The multiplying costs of time, administrative resources, and lower fidelity of services can risk a loss of revenue for businesses, which results in frustration for both administrators and installers.

We designed a new solution to set the placement and rotation of access points (APs) automatically, saving time and increasing accuracy for small businesses to scaling franchises seeking user-centered location services. We also overhauled the existing location services environment based on extensive usability research.

Role

  • UI/UX Design
  • User Research
  • Prototyping
  • Copywriting

Team

  • Senior UX Staff: Jordan Batch
  • SWE Director: Victor Tsai
  • PM: Ryan Adzima
  • QA: Kevin Friday

Duration

2 months

Deliverables

  • Significant feature onboarding and usability redesign
  • Desktop and mobile environment mockups
  • Screen inventory of both desktop and mobile user flows
  • Component-based design system

Business Opportunity

  • Competitive features: Auto-rotation and intuitive UX are a competitive edge to retain existing customers and attract new ones
  • Save time and money: Auto-rotation AI accomplishes self-configuration of APs, which reduces significant installation time and cost spanning thousands of stores for franchises
  • Relevance for customers: Deliver personalized experiences based on completion criteria of a floor-plan and display context-based map detail levels
  • Automation: Auto-rotation AI eliminates AP orientation accuracy issues and user fatigue in manual installation, ensuring quality of location services

Goals

  • Design a clear, intuitive, and highly usable product implementation of Mist AI’s Live View environment
  • Improve customer satisfaction and loyalty with user-centered interactions and design features
  • Empathize with users in the design process by user research, user journey maps, and feature feedback
  • Implement competitive features, including auto-rotation of access points (APs) and intuitive UX, to the current design

Outcome

Predicted (in usability testing phase):
  • XX% shorter time to set the orientation of an AP, reducing significant overhead for larger franchises
  • XX% reduced number of clicks for AP installation
  • XX% improved accuracy across AP orientations

User Journey Map

In the user journey map writing process, we discovered that administrators spend significant overhead manually setting multiple APs for a floor-plan, which can multiply across thousands of stores in a franchise. This is a problem because administrators face lost time and installation fatigue from precisely rotating virtual APs according to physical specifications.

The Current Experience

After conducting the user interview process with product management, we identified primary sources of friction in our existing UI.

  1. Disorganized UI
  2. Actions on the current Live View environment were cluttered with little semantic relation to their placement. Important tasks were nested under menus; features weren’t thought of holistically.

  3. Confusing navigation
  4. While researching users and auditing the “Live View” page, we discovered that users were confused on the navigation and structure of live view.

  5. Repetitive tasks
  6. Administrators will need to go through multiple clicks and manual checking with specifications to install a single AP. For customers with thousands of stores and 10-15 APs per store, accurately placing all the APs takes a lot of time.

The initial Live View environment prior to redesign

Spring Cleaning

With the new settings page, users are able to view important information about a floorplan from a glance. Additionally, this new settings page serves as a single-step solution for the initial setup.

Aligning navigation closer to user's mental model

Before designing the experience for Auto Placement and Orientation, I did some preliminary cleanup of the interface. Based on user feedback from Quality Assurance, I addressed various pain points by reorganizing actions to semantic groups. In this process, I decided to create a new Floorplan Settings page to address a glaring issue: there is no central interface to modify the floorplan.

A new settings page optimizing for sitewide actions

Set Sail

For administrators to use the new Auto Placement and Orientation Features, the floorplan requires three APs to be designated as anchors for triangulation on a virtual floorplan. I introduced a visual concept of anchors (designated by a ship anchor as an accessible mental prototype) so users can see any AP related to anchor designation at a glance.

UX flow for designating anchors

Auto Placement

Administrators can set all their APs on a floorplan in two clicks, two minutes (approximated). They can also view precise value changes with an AP table prior to accepting placement. A preview of the floorplan APs allows administrators to see the changes with a visual guide.

UX flow for Auto Placement

Auto Orientation

Because Mist AI processes Auto Orientation overnight, I designed a distinct UI for interacting with the processing queue. Administrators can easily add and remove floorplans queued for Auto Orientation, receiving a notification the next day when the process is done.

In this process, I also developed a modular component system for modal windows, adapting existing design guidelines for novel use cases.

UX flow for Auto Orientation

Outcome

Through this process, we overhauled the user environment with optimized feature navigation and a new solution for administrators to set the placement and rotation of access points automatically.

Through this project, I learned how to identify and prioritize pain points based on user research, the importance of distinct UI elements, and how to successfully overhaul legacy interfaces.

Once user data is available upon feature testing rollout, we will quantitatively determine the impact of this feature via user metrics (total clicks and time spent on AP adjustment).