NXTDRIVE ADA
What is NXTDRIVE
The Future of Data-Driven Marketing
NXTDRIVE is a Customer Data Platform (CDP) that helps businesses unify messy customer data, enrich it with machine learning intelligence, and activate it across marketing channels like email, print, and digital ads. By integrating proprietary Valassis data and predictive analytics, it enables smarter audience targeting and campaign optimization.
With a self-service portal, users can analyze customer data, track marketing performance, and make data-driven decisions—all in one place. NXTDRIVE also connects legacy QuickPivot and Valassis systems, modernizing data infrastructure to drive more effective and automated marketing strategies.
Quickpivot Neptune: The Foundation of NXTDRIVE’s Evolution
WHAT IS ADA
The Intelligence Engine Powering NXTDRIVE
Ada is a scalable, serverless machine learning engine that enhances NXTDRIVE by delivering predictive insights for smarter marketing decisions. While NXTDRIVE unifies and activates customer data, Ada transforms it into actionable intelligence, enabling businesses to predict customer behavior, optimize campaigns, and automate data enrichment.It allows businesses to train, deploy, and monitor predictive models seamlessly, ensuring data-driven marketing strategies are more precise and impactful.
By simplifying model execution, training, and monitoring, Ada makes advanced analytics accessible without requiring deep technical expertise. Its auto-scaling, serverless architecture ensures efficiency, flexibility, and seamless integration—bridging the gap between raw data and real-time decision-making.
Our Users
Ada’s users include internal teams and external clients who range from non-technical to data-savvy users. The UX challenge was to create an experience that catered to both technical users needing control and non-technical users requiring simplicity and automation.
My Role in Ada
Making Machine Learning Accessible
As the design lead of this project, I played a key role in shaping the UX of Ada within NXTDRIVE, ensuring that its powerful machine learning capabilities were accessible and actionable for users. While Ada handled model execution, data transformation, and predictive analytics, my focus was on designing an intuitive experience that allowed users to interact with Ada’s intelligence without needing deep technical expertise.
My key contributions included:
• Designed the Ada Dashboard & Model Management UI, enabling users to monitor database health, marketing performance, and ML model execution seamlessly.
• Developed interactive controls for scheduling model updates, triggering executions, and integrating model results into marketing workflows—reducing backend reliance.
• Bridged UX & Data Science, collaborating with data scientists to simplify ML workflows and make predictive insights accessible to users of all technical levels.
• Refined the Product Vision, partnering with product managers to shape Ada’s features and align them with user needs and business goals.
Navigating Complexity
Bridging UX & Data Science in Ada
Integrating Ada into NXTDRIVE posed unique challenges from a UX standpoint, the primary hurdles included：
• Bridging Technical Complexity & Usability – Ada was initially designed for data scientists, making it highly technical. The challenge was to translate complex machine learning workflows into intuitive UI elements that could be easily understood by marketers and business users without deep ML expertise.
• Communicating with Data Scientists – Collaborating with data scientists required learning their terminology, understanding different model types (e.g., Classification, Regression), and aligning their technical goals with user needs and business objectives.
• Understanding ML Models & Their Value – Ada supported multiple predictive models, each serving a different purpose. The UX challenge was to clarify what each model did, how it was trained, and what insights it generated, ensuring that users could confidently interpret and apply the results.
• Defining the Right User Experience – Unlike traditional SaaS tools, ML-driven platforms require a different approach to interaction design. Users needed clear guidance on scheduling, executing, and interpreting model results, which required thoughtful workflows and user education.
Workflow Breakdown
Design Approach
Creating an Intuitive ML Experience
Before diving into each model’s details, I first focused on understanding the model status throughout its lifecycle and how to present it in a clear, intuitive way. My goal was allowing users could quickly identify which stage their models were in and what actions they could take next.
Mapping Model Status Across Different Scenarios in wireframe
To reduce confusion, I simplified the model status into three clear categories: Active, Pending, and Need Support, allowing users to quickly assess their model’s state at a glance. For those needing more detail, an expandable view provides a breakdown of progress, showing the exact stage of the model lifecycle. Since the initial plan was to embed the alpha version of Ada into the early-stage NXTDRIVE (3C), I prioritized a streamlined, view-only experience, limiting customization features to ensure the interface balanced user needs, business goals, and technical bandwidth effectively.
↓ Alpha NXTDrive (3C)
↓ Alpha Ada in 3C
I collaborated with data scientists to understand various machine learning models, including descriptive models, predictive models, classification models, linear regression models, clustering models, and deep learning models. With this knowledge, I designed a dashboard experience that translates these complex models into clear, actionable insights, ensuring that users with varying technical backgrounds can easily interpret and leverage the data. Below are few examples.
Laying the Foundation for Future Innovation
Although this project was put on hold due to the company acquisition, it remains a high-impact initiative with the potential to transform how users interact with machine learning models. Our UX contributions extended far beyond crafting screens—through strategic questioning and challenging assumptions, I helped the team rethink the scope and purpose of Ada. As one data scientist mentioned, they wished I had joined earlier to contribute to roadmap planning and early-stage decision-making.
While the project is currently paused, it is not abandoned. Once it returns as a priority, the UX foundation we established—clear model status tracking, intuitive dashboards, and accessible insights for all users—will serve as a strong starting point for future development.
