Springboard
Project - Data-driven cloud-based insight platformProject Type - Design the product, add features and create add-onsDuration - 13 months (2020-2021)Role - UI/UX Designer
Springboard was the flagship product of Trabeya; an advanced data analytics and artificial intelligence (AI) products and services start-up that served a global customer base. It integrated with client’s internal and external data for predictive analytics and augmented intelligence. It functioned with continuous learning through AI for real-time business insights.
Springboard was a suite of apps that operates on top of user’s existing data sources and external data sources as a semantic layer. With in-built AI apps, businesses could leverage automated machine learning to generate insights to enhance performance and drive collaboration.
Problem
A lot of businesses were accumulating large amounts of data without processing. This data often remained in isolated silos, preventing meaningful interaction. As the world was moving towards data-driven decision making, most of these businesses lacked the internal resources to process this information for decision making. At the same time, AI was a relatively novel concept that was not widely understood.
Redesign
The MVP was already used by a few selected clients. However, the product was built without a designer in the team hence it showed the lack of designer expertise. The goal was to redesign the product before the first release. The product was unique of its own, but we wanted it to follow modern design standards at the time.
Research and insights
When designing we wanted to understand the industry needs and workflows to create a solution that works and adapt into different businesses. The main focus industries were finance, healthcare, real estate, retail and agriculture. With the expertise of the business analyst team and several feedback sessions, I was able to gather these key insights.
-
We found that users required seamless data unification across siloed data systems and access to real-time, actionable insights that would reduce reliance on multiple tools and manual processes.
-
Each industry, while unique, needed robust data visualisation tools for intuitive forecasting and performance tracking.
-
Since Springboard would be used by a wide variety of professionals and given its complicated functionality, we identified common design patterns in other platforms to reduce learning curves and ensure ease of adoption.
-
Research and the feedback of the MVP suggested that users valued insights that go beyond traditional analytics.
Design process
Due to the scale of this product, I decided to invest time in building a design system. We searched for an existing design system we could use; however, at the time, there weren't as many ready-made options as there are today, and we couldn't find one that suited our needs. This led me to create a design system from the ground up for Springboard. Some core components and patterns included:
Navigation elements: To accommodate the products complex workflows.
Data visualisation elements: Charts, graphs and trend indicators.
Interactive components: Buttons, forms, tooltips etc.
ISO 9001 compliance: Typography, colour palette, iconography.
User flow and interactions
Data visualisation
Users needed to interpret complex data quickly and effectively. I designed flows that guided users from data integration to visualisation in a step-by-step process. For example, users could upload their datasets, configure filters, and generate real-time, interactive dashboards within minutes.
Progressive disclosure
To not overwhelm users, complex processes like setting up AI-driven predictive models were broken into manageable steps. Additional options were revealed progressively as users navigated deeper into workflows.
Inline assistance
Tooltips, help icons, and contextual guidance were integrated directly into the interface to assist users without interrupting their flow.
Interactive feedback
Real-time feedback mechanisms, such as loading indicators and confirmation messages, ensured that users always knew the status of their actions, enhancing trust in the platform.
Customisable dashboards
Flows allowed users to personalise their dashboards with widgets tailored to their industry, ensuring relevance and engagement.
Integrations
Integration with other systems was a crucial aspect of Springboard. The platform needed to connect to various data sources to access data and provide meaningful insights.
Since this integration process was often a user's first interaction with the product, I focused on making it easy and straightforward to create a strong first impression. For certain connectors, we simplified the process to just four steps.
Managing data
Many Springboard customers had massive amounts of unprocessed data that needed to be managed within the platform. To address this, I developed Data Ocean, a unified space in Springboard where users can view and manage all their fetched data in a familiar way, similar to a traditional file manage or Google Drive with a purpose built tagging system and filters.
With great amounts of data comes great responsibility. Privacy and security were top concerns from the beginning of the design process. To address these concerns, I created an "Organisation Admin" panel tailored for in-house IT teams. This panel enabled granular control over permissions and data source management.
AI powered file processing
A little help goes a long way. We built "Sia," an AI-powered assistant that uses natural language queries to help users navigate through huge amounts of files. Sia can easily find files matching keywords or act as an AI-powered OCR to read documents and provide answers.
Solution
We built a product that simplifies complex processes, enabling users to navigate sophisticated data-driven tasks with ease. The design bridges the gap between advanced technology and user-friendliness by offering plenty of seamless integrations to connect with existing data sources, an AI assistant (Sia) for intuitive queries and navigation, and organisation-level management tools tailored for IT teams. Additionally, granular access settings ensure robust data privacy, empowering users to maintain control over sensitive information.
Implementation
The project included many technical aspects that were completely new to me. Therefore, close collaboration with the business analysis team in the early stages of the project and working closely with ML engineers, data scientists, and developers was important to ensure my designs aligned with the product specifications.
Regular feedback sessions, prototype testing, and workshops with various stakeholders helped me stay in line with expectations.
We adopted an iterative development approach for this project to minimise errors and manage the workload effectively. This method allowed us to break the project into manageable pieces, focusing on one aspect at a time rather than attempting to build the entire system at once.
Collaboration
Challenges
Managing complex data visualisation
Designing data visualisations that could effectively communicate insights from massive, diverse datasets while remaining visually appealing was another challenge.
Ensuring scalability and consistency
With over 80 native connectors and a growing suite of AI apps, maintaining consistency while differentiating apps across all elements of the platform was critical to avoid user confusion.
Testing and feedback
The redesigned beta version was first released to the same set of customers who had been using the MVP. This gave us the opportunity to work closely with customers we already had a good relationship with to test our product.
The testing was primarily unmoderated due to the nature of our clientele, though we did conduct some moderated sessions as well. Testing methods included user interviews, mockup testing, session recordings as well as heat maps and A/B testing.
Iterations
A key finding from our user testing revealed that our clientele wasn't as tech-savvy as we initially assumed. To address this, we refined our workflows, making them less technical and more user-friendly. A standout feature was the visual query builder, which empowered users to create SQL queries without writing any code.
Results and impact
Data-driven discounts
A retail company used Springboard to establish a structured and accurate process for dynamic pricing. The company required an insights-driven system that combined internal and external data to automatically analyse these and implement discounts in a more structured way. Post-implementation, the company has an efficient, streamlined data gathering and analysing process that provides ready insights into demand patterns and sales of products, particularly those of the low-performing segments. This information is then used alongside external datasets to determine the optimum discounts, adding more than 12% to its bottom line.
Reduced reporting times
A financial services company, which had already developed a new cloud data warehouse, was looking for ways to translate this data into insights quickly to support strategic and tactical decision making as well as provide a seamless user experience to their customer. The requirement was multiple dashboards and reports that provided timely and accurate visibility into key metrics defined by them. Post implementation, the team was able to handle ad hoc requests and incorporate these to planned dashboards easily and deliver in multiple areas, such as onboarding trends, comparative sales personnel performance, product-centric views across the portfolio, client relationship management and KPI tracking, cash flow, and fund movement. Result was 15 % reduced reporting time.
Automating supply chain
An organisation that owns and operates a large cattle herd across multiple properties in two cities required a system that offered timely and accurate forecasts for future livestock supply based on multiple value-indicative dimensions. We built a modular system that complemented the existing ERP process to provide a dynamic view of available resources and production via an enterprise business intelligence tool. The solution involved connecting directly to dedicated views and data optimisation to lessen the burden on production systems. The developed system automated the entire pipeline, completely replacing the time-consuming manual system used previously. It now generates production forecasts on demand compared with 10-12 hours (sometimes up to a day) taken to perform the same task earlier.
All materials and content related to the Springboard platform presented in this case study are the exclusive property of Trabeya Ltd. As the designer, I am showcasing this work solely to illustrate my design contributions. All rights, including intellectual property, remain fully reserved by Trabeya.
Trabeya was later acquired by Tavistock Group and merged with Surge Global.