MVP Design for Enterprise: Event Monitoring Console
Lead Product Designer
~ 2.5 months (design+developement)
We discovered that Umbo’s AI detection has the potential to help alarm centers speed up their monitoring process. Our goal is to launch an easy-to-use MVP that enables them to receive and manage alerts filtered by Umbo.
I played a key role in leading the product design for this project, collaborating with Wendy Wang on UI design.
Achievement
Successful launched the MVP with 3 Alarm Centers onboarded, (including a Security Center in a Fortune 500 company in the US.), lower churn rate among users.
Contribution
- Synthesized research data. Generated and prioritized functional requirements.
- Created prototypes strategically and lead the testing with customers and stakeholders.
- Mentored and collaborated with a junior UI designer to achieve comprehensive design specifications.
Impact
- Implemented workshops to enhance team collaboration and cohesion.
- Pioneered a novel hypothesis-driven design approach to foster a data-centric mindset.
- Established a systematic and strategic development framework to enable ongoing iterative progress.
Alarm centers are struggling to achieve faster response times
To ensure client safety and property protection, alarm centers strive for faster response times to events. Our previous research revealed that they have developed intricate workflows and employed a significant workforce to address the shortcomings of tools used in event monitoring over the past decades.
.png)
AI only tackles a part of the problem
Alarm centers trialing our camera and platform were impressed by our AI accuracy, filtering out 97% of false alerts. However, the current Umbo platform experience, designed primarily for Integrators, did not integrate with their existing professional event monitoring workflow.
A Console for Professional Event Monitoring
Since 2017, Umbo had been sending instant AI-filtered alerts through our platform and app by the push notifications, which is great for end users with a few cameras. However, for alarm center, our objective was to reimagine how to deliver these "accurate" alerts to operators in a manner that suits their event management and monitoring workflow.
A MVP designed to onboard users for trial purposes
Alarm centers were a completely new user group for us at that time. To define the task of "professional event monitoring," we conducted on-site contextual studies at various alarm centers to understand how they operate.
Given the constraint of a 3-month timeline for delivery and the goal of innovation, the design process was based on the double diamond design thinking framework, aligned with our Agile product development workflow. This approach facilitated the rapid convergence of ideas with flexibility and validated solutions.
User Tasks → Design Possibilities
From the study, we gained insights into the alarm center's business and their workflow. The data covered various perspectives, including their challenges, existing solutions, difficulties, and desires. (See details in Projects: Contextual User Study)
a. Brainstorm the possible new solutions
We conducted a task analysis workshop to:
- Categorize and break down the real user stories behind the tools and solutions
- Brainstorm the possible solutions with the user stories
Participants include sales, the CEO, designers, and developers. The workshop facilitated a creative space for people with diverse backgrounds to participate in the design process.
.png)
The current workflow and tools could potentially constrain the team's ability to generate functional requirements. Together, we use user stories to uncover the underlying needs and explore various design possibilities.
b. Evaluate impacts on the current experiences
Due to specific business and system design concerns, the MVP was initially planned for integration into Umbo's existing platform. However, it soon became apparent that the alarm center represented a new user role focused on different tasks compared to the integrators.
While we lacked the resources to create an entirely new user group for alarm centers, we conducted an analysis aligned with the user segments of both existing users (integrators) and the new users (alarm center) to assess the experiences and potential impacts.
.png)
Fortunately, despite differences in the primary tasks of both user groups, there are similarities in how these organizations manage and distribute their tasks.
We decided to leverage our existing user roles for the new MVP feature before gaining a deeper understanding of the industry. This insight also lays the foundation for creating a comprehensive platform for both user groups in security industry.
A MVP that is not only feasible but also delightful
Monitoring platforms weren't a new concept, as there were robust products with comprehensive features already available in the market. Our aim was to set ourselves apart from these prominent monitoring platforms by providing a superior user experience through the use of AI technology. We hope operators not only to fulfill their tasks but also to do so in a more delightful manner.
This involved making trade-offs with some functional features but provide our customers with a more delightful experience that aligned with Umbo's values.
a. Align Umbo design principles
Requirements were developed to create an even more "reliable, straightforward, empowering, and nimble" experience based on Umbo principles.
b. Define the user experience goals
The 3 core UX keywords are established on the basis of:
- The “goals” and "ideal scenarios” mentioned in the interviews
- The values that Alarm Centers bring to their customers
- Alignment with Umbo's product principles
“Event” queue: stack alerts by the actual incident
The Problem
The system sent an alert whenever a "human" was detected, potentially triggering multiple "alerts" for an actual single “incident/event”. This often required operators to spend extra time distinguishing the events before taking action.
.png)
We realized that alerts serve as one of the 'details' for an incident, similar to playback. They provide information for operators to 'diagnose' what actually happened.
The Design
- “Incident queue” rather than “alert queue” :
Aggregate alerts by continuous timing and display them as "details" for incidents - Flexibility for human edit:
Enable bulk operations on a batch of "alerts”
Retain the option for individual event editing
The Challenge
We soon encounter many questions while “stacking” the alerts. Here is one of the questions:
🧐 🧐 🧐 A 'camera' is typically set up under a 'location' and a 'customer.' At which level would users need grouping?
How we tackle it - The hypothesis-driven approach
Given limited resources for additional exploratory studies, I developed the following structure to strategically design prototypes for exploring the most effective solutions.
The structure not only provides designers but the entire team with a clear understanding of why we make certain decisions, along with the benefits and potential risks we might encounter. Each verification from the interviews helped us enhance the knowledge base of the overall user experience.
For more controversial assumptions, we provide a detailed listing of the knowledge background we have collected. This proved valuable during subsequent iterations, ensuring that changes to the interaction were not arbitrary but rather logical tests of our hypotheses.
After some follow-up calls with potential customers, we decided to “group by cameras” as it represents the scenario of our target customer who is most likely to immediately onboard the MVP.
Manageable event queue
The Problem
Operators need to continuously talk to each other to avoid handling the same alerts repeatedly, or lose track after trying to contact the customer or looking for other information for event handling.
The Design
The behavior is similar to email management or a call center system. After some quick investigations and discussions with designers, the alert queuing logic was designed as follows:
- Alerts are automatically entered into the console, but operators can manually select alerts if they consider them more critical.
- An alert will be "claimed" once it is assigned to or selected by an operator, and this will display a different view for other operators.
The Challenge
From the interviews, various factors influence how operators select events, including priorities, customer demands, and situations reported by the dispatch team, among others.
🧐 🧐 🧐 How do we push the event to operators’ face?
How we tackle it - Design backend alert processing logic together
Following the hypothesis-drive process and the verifications from the follow-up calls with the customers. We realized both “auto-assign” and “manual-select” were needed from flexibility concern.
I worked with developers to generate a concept of “claiming” system so the alert sticks to the operator who “claim” to handle it. This enable the ability for other behavior like “disclaim” and “park” the events.
.png)
This feature is akin to systems found in call centers. By leveraging the research background from similar products such as task management platforms and understanding the user task flow, we designed an interaction flow that streamlined the basic event monitoring process.
All-in-One, Onboard Seamlessly, and more
The MVP's main goal is to onboard alarm centers, serving as the “central console” for event monitoring. The platform not only covers the basic task flow of event monitoring but also integrates services into a unified platform.
Connecting Umbo services, and more
- Leverage Umbo's seamless camera streaming and event timeline experiences.
- Plan to integrate "hardware actions" into the platform, such as camera torchlight, siren, or two-way audio.
Label System for Advanced Data Analysis
Similar to other AI services, a feedback system serves as a valuable connection between users and the AI. This not only enhances user engagement and a sense of control but also empowers professional users to collaborate effectively with AI and data.
With the collaborative effort of our sales team, the MVP has been successfully adopted by our valued customers, including a Security Center in a Fortune 500 company in the US.
The data in 2022 shows users who adopted the console churned at a lower rate than the company average.
Making decisions under limited resources
There were lots of uncertainties during the design process. Over 10 versions of interaction prototypes and numerous visual materials were created and shared with customers and teams through regular reviews and calls over a month.
During the intense process, we applied the following approaches to get closer to the real user needs, verify our ideas and get the team more synchronized.
a. Follow-up calls & co-design with potential customers
Hi-fi prototypes with interactive interfaces are crafted to allow enterprise users to clearly visualize how they can utilize the product. This is particularly crucial as they have limited time available for interviews.
.png)
b. Implement hypothesis-driven approach to our design process
Hypothesis-driven thinking provides us with a data-driven mindset not only before the design but also during and after the design. This makes the project an integral part of our product planning, allowing the entire team to have a broader perspective and greater confidence in the iterative process.
c. Design with engineers to evaluate the performance and efforts
As the MVP was built within the existing Umbo Platform, the discussions were proactively initiated with engineers from the prototyping stage to identify potential issues and understand how backend and system design may impact the overall performance of the product.
Interaction & UI Spec
A comprehensive interaction spec was presented and reviewed with developers during a handoff meeting. The details, issues, non-goals, limitations, and concerns were highlighted in the spec to minimize back-and-forth communication. Comments from every team member were documented and addressed within the specification to capture the context of the discussions.
Fortunately, I had the opportunity to mentor our graphic designer in UI design. She played a significant role in developing the majority of the UI specifications, which allowed me to focus on the interaction specifications (sending much love here 🫶🏻 )
.png)
Design System Updates
Designing with Umbo's first design system in mind, our team has seamlessly incorporated its principles into the UI of this project. Additionally, we have had the opportunity to leverage our resources to design the dark mode for our design system.
Future Work
This research has established the initial structure of our user database for the Event Monitoring Industry, forming the foundation for the majority of our platform's key features.
I led to organized the functional requirements list on JIRA, with the contextual descriptions and links to the study findings, serving as a framework for the ongoing development of this product by the product team.
Also, the studies were ongoing. As the data framework was established, more data was continuously added to the database to form a more complete and comprehensive understanding of how this product should be developed.
In 2020, Umbo underwent another product pivot, diverting a significant portion of resources away from this project. While many early adopters continue to use the product for monitoring tasks, it's unfortunate that we lacked the resources to thoroughly track and evaluate detailed product usage.