Automated Deployments: Reduced Errors, Downtime and Complexity with Streamlined Workflows and Real-Time Feedback
Design Brief
OVN Deployments is a solution developed for internal operations teams (L1, L2) who manage and execute deployment cycles across different environments. As the company transitions from 25-year-old mainframe systems to modern architectures, we undertook a complete redesign and conducted thorough research to understand user pain points and requirements. The existing mainframe systems caused bottlenecks, resulting in convoluted, error-prone deployment processes, extended downtime, and high operational costs. The goal of this UX project is to create a seamless, intuitive, and efficient interface that streamlines deployment workflows, reduces errors, and minimizes downtime for the operations teams.
Company
Visa
Role
Lead UX Designer

Research
Conceptualization

Collaboration
Design
User testing 

Dev handoff
Team
01 Architect
01 Product Manager
02 UX Designers
05 Engineers
Duration
9 months
Problem
The deployment rollouts are currently handled through mainframe systems, relying on script-based processes that are highly error-prone. Although multiple layers of validation are in place, manual input of script lines still results in frequent mistakes. Moreover, the deployment process is time-consuming and cumbersome, especially when extracting or triaging issues. The lack of automation and user-friendly interfaces exacerbates these challenges. Additionally, there is significant wait time for approvals, as communication is done through emails, further delaying the deployment cycle and causing bottlenecks.
Challenges Caused by Legacy Mainframe Systems:

-
Complex, Manual Processes: Deployment tasks were mostly manual, increasing the likelihood of errors and slowing down workflows.
-
Lack of Real-Time Feedback: Operators had limited visibility into the deployment process, making it difficult to track progress or identify issues. This uncertainty heightened stress during critical operations.
-
Extended Downtime: Inefficient processes and deployment errors often resulted in prolonged downtime, impacting the availability of essential financial systems.
-
High Operational Costs: The reliance on manual interventions and inefficient workflows led to increased labor costs and the need for additional resources to troubleshoot and recover from failures.
-
Siloed Operations: Teams worked in isolation, with limited communication between L1, L2, and engineering teams. This fragmentation caused inefficiencies, slower troubleshooting, and delayed deployments.
HYPOTHESIS
We hypothesize that automating and simplifying the code deployment process with real-time feedback, built-in validations, and proactive error handling will enable Operators (L1) and Product Reliability Engineers (L2) to:
-
Reduce deployment errors by 40-50% through automation and improved accuracy.
-
Accelerate deployment cycles by 30-40%, minimizing manual steps and approval delays.
-
Enhance confidence and control during deployments with greater visibility into real-time progress and issues.
-
Improve operational efficiency and reduce downtime, ensuring more reliable and seamless deployments across environments.
Design Process
CONTEXTUAL INQUIERY (DISCOVERY INTERVIEW)
I utilized Contextual Inquiry, a qualitative research method, to observe and understand how users interact with the product in their natural environment. This approach was particularly valuable for gaining insights into users' workflows, behaviors, and the context in which they use the product.
During the interviews, I asked users to walk me through their experience, step-by-step. This allowed me to explore:
-
How users engage with the product in real-time.
-
The tools or systems they rely on throughout their workflow.
-
The decisions they make while using the product and the reasoning behind them.
-
The challenges they face and any workarounds they employ to address issues.
This hands-on, observational approach provided me with deep insights into how users currently interact with the system and highlighted areas for improvement in both functionality and user experience.
FOCUS GROUPS
After conducting Contextual Inquiry to observe user interactions with the legacy system, I organized focus groups to validate research findings and gather broader feedback from operations team members.
Key Outcomes:
-
User Needs: Participants highlighted critical features like automation, real-time feedback, and improved communication across teams.
-
Design Validation: Early wireframes were reviewed, and feedback helped identify areas of confusion and usability improvements.
-
Pain Points: Issues like lack of real-time feedback and system complexity were confirmed as key frustrations.
-
Collaboration: A need for better integration between L1, L2, and engineering teams emerged, with suggestions for messaging and notification features.
-
Feature Testing: Concepts like automated error detection and rollback features were tested, refining ideas based on user input.
INSIGHTS
Users experienced frustration due to limited visibility into deployment progress, making it difficult to monitor the process or catch issues early.
There’s a lack of communication between L1, L2, and engineering teams, which leads to slower troubleshooting and approval delays.
The approval process, handled via email, caused delays and bottlenecks in the deployment cycle.
Users expressed frustration with the reliance on manual processes, which often result in errors and delays during deployment.
GOALS
Improve Deployment Visibility
Provide real-time tracking and visibility of deployment progress, enabling users to monitor processes and identify issues early.
Enhance Communication Across Teams
Implement seamless communication tools and integrations between L1, L2, and engineering teams to reduce delays in troubleshooting and approvals.
Streamline Approval Process
Replace the email-based approval process with an integrated, centralized approval system to eliminate delays and bottlenecks in the deployment cycle.
Automate Processes to Reduce Errors
Automate manual tasks in the deployment process to reduce human error, minimize delays, and enhance overall efficiency.
USER PERSONA
Based on interviews with several team members, I created a single persona that reflects their common roles, responsibilities, and challenges in managing deployment processes.
.png)
USER RESEARCH
The interview process took approximately three weeks to engage with end users and understand their problems and pain points.
INSIGHTS
MAINFRAME JOURNEY MAPPING
I have mapped the existing deployment journey to highlight friction points and identify areas for improvement, particularly in error handling and manual interventions.
Identifying & Prioritizing User Jobs
A comprehensive list of user jobs across personas were framed and stack ranked. This exercise laid the foundation for our MVP.
An Example of a Job-To-Be-Done for deployment to capture user jobs and needs.
Actor
As an operator
Situation
When I deploy changes using the new automated system with pre-validation checks.
Motivation
I want to the system to reduce errors, minimize manual work, and gain better control over the deployment.
Outcome
So that I can deploy faster, with higher confidence, and reduce troubleshooting time.
BUCKETING USER JOURNEY
Bucketing user jobs into the stages of the user journey
User jobs from the Jobs-To-Be-Done exercises helped us see patterns and map out the user journey into nine broad buckets, For the MVP we ideated concepts under each of these buckets.
Schedule
User schedules deployment via the application scheduler
Full Activation
Once satisfied with the partial deployment, the user triggers the full deployment.
Approve & Trigger
The operator approves and initiates the deployment.
Monitoring
After full deployment, the operator monitors and verifies the services.
Precheck
All required health checks are completed
Post Processing
The user checks all post-deployment steps.
Partial Activation
Initially, some services are deployed
Monitoring
After partial deployment, the operator manually monitors the deployed services.
Next Data Center
A success signal is sent to the next data center for rollout.
Structure and streamlining the concepts
We developed the user flows to know the apps screens see the future flows, and organize the ideas that we had for redesign.
Solution Discovery
USER FLOWS
USABILITY STUDY
After completing the low fidelity prototype, we approached the participants that were initially interviewed during my primary research phase to conduct a moderate usability test to analyze how users will interact with the application. After iterations we have come up with the final fidelity designs.
FINAL VERSION OF FEDILITY DESIGNS
Final Solution
Automated Deployment Solution:
To reduce manual errors and increase deployment confidence, we introduced an automated solution:
-
Eliminating Manual Scripts: Replaced manual scripts with a system-readable deployment manifest, allowing operators to easily spot discrepancies.
-
Pre-Deployment Confidence: Automated pre-validation checks ensure the environment meets all criteria before deployment.
-
Automated Version Comparison: The system automatically compares the current version with the release version, confirming alignment and preventing errors.
.jpg)
Confirmation Popup with Pre-Validation Overview:
The confirmation popup provides full visibility of critical pre-validation checks, allowing users to confirm that all conditions are met before deployment, preventing issues and boosting confidence.
Upon deployment initiation, an automatic record is created in the audit log, capturing key deployment details. This ensures transparency, accountability, and provides a traceable history of all actions.
Execution Page: Automated Deployment with Transparency and Control
The Execution Page centralizes and automates the deployment process, offering full visibility and control at every stage:
-
Streamlined Approval: Approvals, previously handled via email, are now centralized, reducing bottlenecks and enabling smooth operator-triggered deployments.
-
Deployment Overview: Operators can quickly access key details like start time, duration, and progress status, with a unique Execution ID for traceability.
-
Progress Transparency: Real-time tracking through categorized workflow tabs ensures complete visibility of each deployment step.
-
Error Handling: Retry options, clear distinction between mandatory and optional checks, and rollback/cancel features enhance confidence and control.
-
Continuous Tracking: Operators can track deployment progress in real-time and address issues immediately.
